| blob | 6382b14521141cd7808e3f6a63842b8649b8982c |
1 ;;; cc-engine.el --- core syntax guessing engine for CC mode -*- coding: utf-8 -*-
3 ;; Copyright (C) 1985, 1987, 1992-2015 Free Software Foundation, Inc.
5 ;; Authors: 2001- Alan Mackenzie
6 ;; 1998- Martin Stjernholm
7 ;; 1992-1999 Barry A. Warsaw
8 ;; 1987 Dave Detlefs
9 ;; 1987 Stewart Clamen
10 ;; 1985 Richard M. Stallman
11 ;; Maintainer: bug-cc-mode@gnu.org
12 ;; Created: 22-Apr-1997 (split from cc-mode.el)
13 ;; Keywords: c languages
14 ;; Package: cc-mode
16 ;; This file is part of GNU Emacs.
18 ;; GNU Emacs is free software: you can redistribute it and/or modify
19 ;; it under the terms of the GNU General Public License as published by
20 ;; the Free Software Foundation, either version 3 of the License, or
21 ;; (at your option) any later version.
23 ;; GNU Emacs is distributed in the hope that it will be useful,
24 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
25 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
26 ;; GNU General Public License for more details.
28 ;; You should have received a copy of the GNU General Public License
29 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
31 ;;; Commentary:
33 ;; The functions which have docstring documentation can be considered
34 ;; part of an API which other packages can use in CC Mode buffers.
35 ;; Otoh, undocumented functions and functions with the documentation
36 ;; in comments are considered purely internal and can change semantics
37 ;; or even disappear in the future.
38 ;;
39 ;; (This policy applies to CC Mode as a whole, not just this file. It
40 ;; probably also applies to many other Emacs packages, but here it's
41 ;; clearly spelled out.)
43 ;; Hidden buffer changes
44 ;;
45 ;; Various functions in CC Mode use text properties for caching and
46 ;; syntactic markup purposes, and those of them that might modify such
47 ;; properties but still don't modify the buffer in a visible way are
48 ;; said to do "hidden buffer changes". They should be used within
49 ;; `c-save-buffer-state' or a similar function that saves and restores
50 ;; buffer modifiedness, disables buffer change hooks, etc.
51 ;;
52 ;; Interactive functions are assumed to not do hidden buffer changes,
53 ;; except in the specific parts of them that do real changes.
54 ;;
55 ;; Lineup functions are assumed to do hidden buffer changes. They
56 ;; must not do real changes, though.
57 ;;
58 ;; All other functions that do hidden buffer changes have that noted
59 ;; in their doc string or comment.
60 ;;
61 ;; The intention with this system is to avoid wrapping every leaf
62 ;; function that do hidden buffer changes inside
63 ;; `c-save-buffer-state'. It should be used as near the top of the
64 ;; interactive functions as possible.
65 ;;
66 ;; Functions called during font locking are allowed to do hidden
67 ;; buffer changes since the font-lock package run them in a context
68 ;; similar to `c-save-buffer-state' (in fact, that function is heavily
69 ;; inspired by `save-buffer-state' in the font-lock package).
71 ;; Use of text properties
72 ;;
73 ;; CC Mode uses several text properties internally to mark up various
74 ;; positions, e.g. to improve speed and to eliminate glitches in
75 ;; interactive refontification.
76 ;;
77 ;; Note: This doc is for internal use only. Other packages should not
78 ;; assume that these text properties are used as described here.
79 ;;
80 ;; 'category
81 ;; Used for "indirection". With its help, some other property can
82 ;; be cheaply and easily switched on or off everywhere it occurs.
83 ;;
84 ;; 'syntax-table
85 ;; Used to modify the syntax of some characters. It is used to
86 ;; mark the "<" and ">" of angle bracket parens with paren syntax, and
87 ;; to "hide" obtrusive characters in preprocessor lines.
88 ;;
89 ;; This property is used on single characters and is therefore
90 ;; always treated as front and rear nonsticky (or start and end open
91 ;; in XEmacs vocabulary). It's therefore installed on
92 ;; `text-property-default-nonsticky' if that variable exists (Emacs
93 ;; >= 21).
94 ;;
95 ;; 'c-is-sws and 'c-in-sws
96 ;; Used by `c-forward-syntactic-ws' and `c-backward-syntactic-ws' to
97 ;; speed them up. See the comment blurb before `c-put-is-sws'
98 ;; below for further details.
99 ;;
100 ;; 'c-type
101 ;; This property is used on single characters to mark positions with
102 ;; special syntactic relevance of various sorts. Its primary use is
103 ;; to avoid glitches when multiline constructs are refontified
104 ;; interactively (on font lock decoration level 3). It's cleared in
105 ;; a region before it's fontified and is then put on relevant chars
106 ;; in that region as they are encountered during the fontification.
107 ;; The value specifies the kind of position:
108 ;;
109 ;; 'c-decl-arg-start
110 ;; Put on the last char of the token preceding each declaration
111 ;; inside a declaration style arglist (typically in a function
112 ;; prototype).
113 ;;
114 ;; 'c-decl-end
115 ;; Put on the last char of the token preceding a declaration.
116 ;; This is used in cases where declaration boundaries can't be
117 ;; recognized simply by looking for a token like ";" or "}".
118 ;; `c-type-decl-end-used' must be set if this is used (see also
119 ;; `c-find-decl-spots').
120 ;;
121 ;; 'c-<>-arg-sep
122 ;; Put on the commas that separate arguments in angle bracket
123 ;; arglists like C++ template arglists.
124 ;;
125 ;; 'c-decl-id-start and 'c-decl-type-start
126 ;; Put on the last char of the token preceding each declarator
127 ;; in the declarator list of a declaration. They are also used
128 ;; between the identifiers cases like enum declarations.
129 ;; 'c-decl-type-start is used when the declarators are types,
130 ;; 'c-decl-id-start otherwise.
131 ;;
132 ;; 'c-awk-NL-prop
133 ;; Used in AWK mode to mark the various kinds of newlines. See
134 ;; cc-awk.el.
136 ;;; Code:
143 load-path)))
152 \f
153 ;; Make declarations for all the `c-lang-defvar' variables in cc-langs.
165 \f
166 ;;; Internal state variables.
168 ;; Internal state of hungry delete key feature
172 ;; The electric flag (toggled by `c-toggle-electric-state').
173 ;; If t, electric actions (like automatic reindentation, and (if
174 ;; c-auto-newline is also set) auto newlining) will happen when an electric
175 ;; key like `{' is pressed (or an electric keyword like `else').
179 ;; Internal state of auto newline feature.
183 ;; Included in the mode line to indicate the active submodes.
184 ;; (defvar c-submode-indicators nil)
185 ;; (make-variable-buffer-local 'c-submode-indicators)
188 ;; Calculate the new state of PREVSTATE, t or nil, based on arg. If
189 ;; arg is nil or zero, toggle the state. If arg is negative, turn
190 ;; the state off, and if arg is positive, turn the state on
196 \f
197 ;; Basic handling of preprocessor directives.
199 ;; This is a dynamically bound cache used together with
200 ;; `c-query-macro-start' and `c-query-and-set-macro-start'. It only
201 ;; works as long as point doesn't cross a macro boundary.
210 c-macro-start))
218 c-macro-start))
220 ;; One element macro cache to cope with continual movement within very large
221 ;; CPP macros.
224 ;; Nil or cons of the bounds of the most recent CPP form probed by
225 ;; `c-beginning-of-macro', `c-end-of-macro' or `c-syntactic-end-of-macro'.
226 ;; The cdr will be nil if we know only the start of the CPP form.
229 ;; The starting position from where we determined `c-macro-cache'.
232 ;; non-nil iff `c-macro-cache' has both elements set AND the cdr is at a
233 ;; syntactic end of macro, not merely an apparent one.
236 ;; Called from a before-change function. If the change region is before or
237 ;; in the macro characterized by `c-macro-cache' etc., nullify it
238 ;; appropriately. BEG and END are the standard before-change-functions
239 ;; parameters. END isn't used.
244 c-macro-cache-start-pos nil
245 c-macro-cache-syntactic nil))
250 c-macro-cache-syntactic nil))))
253 ;; Check that the ostensible CPP construct at point is a real one. In
254 ;; particular, if point is on the first line of a narrowed buffer, make sure
255 ;; that the "#" isn't, say, the second character of a "##" operator. Return
256 ;; t when the macro is real, nil otherwise.
267 t)
271 "Go to the beginning of a preprocessor directive.
272 Leave point at the beginning of the directive and return t if in one,
273 otherwise return nil and leave point unchanged.
275 Note that this function might do hidden buffer changes. See the
276 comment at the start of cc-engine.el for more info."
288 t))
290 c-macro-cache-start-pos nil
291 c-macro-cache-syntactic nil)
304 c-macro-cache-start-pos here)
305 t)
307 nil))))))
310 "Go to the end of a preprocessor directive.
311 More accurately, move the point to the end of the closest following
312 line that doesn't end with a line continuation backslash - no check is
313 done that the point is inside a cpp directive to begin with.
315 Note that this function might do hidden buffer changes. See the
316 comment at the start of cc-engine.el for more info."
325 c-macro-cache-start-pos nil
326 c-macro-cache-syntactic nil))
332 t)))
337 ;; Go to the end of a CPP directive, or a "safe" pos just before.
338 ;;
339 ;; This is normally the end of the next non-escaped line. A "safe"
340 ;; position is one not within a string or comment. (The EOL on a line
341 ;; comment is NOT "safe").
342 ;;
343 ;; This function must only be called from the beginning of a CPP construct.
344 ;;
345 ;; Note that this function might do hidden buffer changes. See the comment
346 ;; at the start of cc-engine.el for more info.
349 s)
361 ;; Assuming point is at the "#" that introduces a preprocessor
362 ;; directive, it's moved forward to the end of the identifier which is
363 ;; "#define"d (or whatever c-opt-cpp-macro-define specifies). Non-nil
364 ;; is returned in this case, in all other cases nil is returned and
365 ;; point isn't moved.
366 ;;
367 ;; This function might do hidden buffer changes.
373 ;; Assuming point is at the "#" that introduces a preprocessor
374 ;; directive, it's moved forward to the start of the definition body
375 ;; if it's a "#define" (or whatever c-opt-cpp-macro-define
376 ;; specifies). Non-nil is returned in this case, in all other cases
377 ;; nil is returned and point isn't moved.
378 ;;
379 ;; This function might do hidden buffer changes.
385 \f
386 ;;; Basic utility functions.
389 ;; Return the given region as a string where all syntactic
390 ;; whitespace is removed or, where necessary, replaced with a single
391 ;; space. If PAREN-LEVEL is given then all parens in the region are
392 ;; collapsed to "()", "[]" etc.
393 ;;
394 ;; This function might do hidden buffer changes.
438 ;; Shift the indentation of the current line with the specified
439 ;; amount (positive inwards). The buffer is modified only if
440 ;; SHIFT-AMT isn't equal to zero.
443 tmp-char-inserted)
445 nil
446 ;; If we're on an empty line inside a macro, we take the point
447 ;; to be at the current indentation and shift it to the
448 ;; appropriate column. This way we don't treat the extra
449 ;; whitespace out to the line continuation as indentation.
465 ;; If initial point was within line's indentation and we're not on
466 ;; a line with a line continuation in a macro, position after the
467 ;; indentation. Else stay at same point in text.
475 ;; Return non-nil if the string KEYWORD is a known keyword. More
476 ;; precisely, the value is the symbol for the keyword in
477 ;; `c-keywords-obarray'.
481 ;; Return non-nil if the symbol KEYWORD-SYM, as returned by
482 ;; `c-keyword-sym', is a member of LANG-CONSTANT, which is the name
483 ;; of a language constant that ends with "-kwds". If KEYWORD-SYM is
484 ;; nil then the result is nil.
487 ;; String syntax chars, suitable for skip-syntax-(forward|backward).
492 ;; Regexp matching string limit syntax.
497 ;; Regexp matching WS followed by string limit syntax.
501 ;; Holds formatted error strings for the few cases where parse errors
502 ;; are reported.
509 c-parsing-error)
511 ;; Faces given to comments and string literals. This is used in some
512 ;; situations to speed up recognition; it isn't mandatory that font
513 ;; locking is in use. This variable is extended with the face in
514 ;; `c-doc-face-name' when fontification is activated in cc-fonts.el.
518 ;; New in Emacs 22.
522 ;; Put a c-type property with the given value at POS.
526 ;; Remove all occurrences of the c-type property that has the given
527 ;; value in the region between FROM and TO. VALUE is assumed to not
528 ;; be nil.
529 ;;
530 ;; Note: This assumes that c-type is put on single chars only; it's
531 ;; very inefficient if matching properties cover large regions.
540 \f
541 ;; Some debug tools to visualize various special positions. This
542 ;; debug code isn't as portable as the rest of CC Mode.
578 \f
579 ;; `c-beginning-of-statement-1' and accompanying stuff.
581 ;; KLUDGE ALERT: c-maybe-labelp is used to pass information between
582 ;; c-crosses-statement-barrier-p and c-beginning-of-statement-1. A
583 ;; better way should be implemented, but this will at least shut up
584 ;; the byte compiler.
587 ;; New awk-compatible version of c-beginning-of-statement-1, ACM 2002/6/22
589 ;; Macros used internally in c-beginning-of-statement-1 for the
590 ;; automaton actions.
593 stack)))
598 t
599 ,do-if-done
604 ;; Throw nil if stack is empty, else throw non-nil.
622 "No matching `%s' found for `%s' on line %d"
629 noerror comma-delim)
630 "Move to the start of the current statement or declaration, or to
631 the previous one if already at the beginning of one. Only
632 statements/declarations on the same level are considered, i.e. don't
633 move into or out of sexps (not even normal expression parentheses).
635 If point is already at the earliest statement within braces or parens,
636 this function doesn't move back into any whitespace preceding it; it
637 returns `same' in this case.
641 is within the continuation. If starting at such a token, move to the
642 corresponding statement start. If at the beginning of a statement,
643 move to the closest containing statement if there is any. This might
644 also stop at a continuation clause.
646 Labels are treated as part of the following statements if
647 IGNORE-LABELS is non-nil. (FIXME: Doesn't work if we stop at a known
648 statement start keyword.) Otherwise, each label is treated as a
649 separate statement.
651 Macros are ignored \(i.e. skipped over) unless point is within one, in
652 which case the content of the macro is treated as normal code. Aside
653 from any normal statement starts found in it, stop at the first token
656 macros before leaving them.
658 Return:
660 `same' if stopped at the beginning of the current statement;
661 `up' if stepped to a containing statement;
662 `previous' if stepped to a preceding statement;
663 `beginning' if stepped from a statement continuation clause to
664 its start clause; or
665 `macro' if stepped to a macro start.
666 Note that `same' and not `label' is returned if stopped at the same
667 label without crossing the colon character.
669 LIM may be given to limit the search. If the search hits the limit,
670 point will be left at the closest following token, or at the start
671 position if that is less (`same' is returned in this case).
673 NOERROR turns off error logging to `c-parsing-error'.
675 Normally only `;' and virtual semicolons are considered to delimit
676 statements, but if COMMA-DELIM is non-nil then `,' is treated
677 as a delimiter too.
679 Note that this function might do hidden buffer changes. See the
680 comment at the start of cc-engine.el for more info."
682 ;; The bulk of this function is a pushdown automaton that looks at statement
683 ;; boundaries and the tokens (such as "while") in c-opt-block-stmt-key. Its
684 ;; purpose is to keep track of nested statements, ensuring that such
685 ;; statements are skipped over in their entirety (somewhat akin to what C-M-p
686 ;; does with nested braces/brackets/parentheses).
687 ;;
688 ;; Note: The position of a boundary is the following token.
689 ;;
690 ;; Beginning with the current token (the one following point), move back one
691 ;; sexp at a time (where a sexp is, more or less, either a token or the
692 ;; entire contents of a brace/bracket/paren pair). Each time a statement
693 ;; boundary is crossed or a "while"-like token is found, update the state of
694 ;; the PDA. Stop at the beginning of a statement when the stack (holding
695 ;; nested statement info) is empty and the position has been moved.
696 ;;
697 ;; The following variables constitute the PDA:
698 ;;
699 ;; sym: This is either the "while"-like token (e.g. 'for) we've just
700 ;; scanned back over, 'boundary if we've just gone back over a
701 ;; statement boundary, or nil otherwise.
702 ;; state: takes one of the values (nil else else-boundary while
703 ;; while-boundary catch catch-boundary).
704 ;; nil means "no "while"-like token yet scanned".
705 ;; 'else, for example, means "just gone back over an else".
706 ;; 'else-boundary means "just gone back over a statement boundary
707 ;; immediately after having gone back over an else".
708 ;; saved-pos: A vector of either saved positions (tok ptok pptok, etc.) or
709 ;; of error reporting information.
710 ;; stack: The stack onto which the PDA pushes its state. Each entry
711 ;; consists of a saved value of state and saved-pos. An entry is
712 ;; pushed when we move back over a "continuation" token (e.g. else)
713 ;; and popped when we encounter the corresponding opening token
714 ;; (e.g. if).
715 ;;
716 ;;
717 ;; The following diagram briefly outlines the PDA.
718 ;;
719 ;; Common state:
720 ;; "else": Push state, goto state `else'.
721 ;; "while": Push state, goto state `while'.
722 ;; "catch" or "finally": Push state, goto state `catch'.
723 ;; boundary: Pop state.
724 ;; other: Do nothing special.
725 ;;
726 ;; State `else':
727 ;; boundary: Goto state `else-boundary'.
728 ;; other: Error, pop state, retry token.
729 ;;
730 ;; State `else-boundary':
731 ;; "if": Pop state.
732 ;; boundary: Error, pop state.
733 ;; other: See common state.
734 ;;
735 ;; State `while':
736 ;; boundary: Save position, goto state `while-boundary'.
737 ;; other: Pop state, retry token.
738 ;;
739 ;; State `while-boundary':
740 ;; "do": Pop state.
741 ;; boundary: Restore position if it's not at start, pop state. [*see below]
742 ;; other: See common state.
743 ;;
744 ;; State `catch':
745 ;; boundary: Goto state `catch-boundary'.
746 ;; other: Error, pop state, retry token.
747 ;;
748 ;; State `catch-boundary':
749 ;; "try": Pop state.
750 ;; "catch": Goto state `catch'.
751 ;; boundary: Error, pop state.
752 ;; other: See common state.
753 ;;
754 ;; [*] In the `while-boundary' state, we had pushed a 'while state, and were
755 ;; searching for a "do" which would have opened a do-while. If we didn't
756 ;; find it, we discard the analysis done since the "while", go back to this
757 ;; token in the buffer and restart the scanning there, this time WITHOUT
758 ;; pushing the 'while state onto the stack.
759 ;;
760 ;; In addition to the above there is some special handling of labels
761 ;; and macros.
765 macro-start
768 c-stmt-delim-chars-with-comma
769 c-stmt-delim-chars))
770 c-in-literal-cache c-maybe-labelp after-case:-pos saved
771 ;; Current position.
772 pos
773 ;; Position of last stmt boundary character (e.g. ;).
774 boundary-pos
775 ;; The position of the last sexp or bound that follows the
776 ;; first found colon, i.e. the start of the nonlabel part of
777 ;; the statement. It's `start' if a colon is found just after
778 ;; the start.
779 after-labels-pos
780 ;; Like `after-labels-pos', but the first such position inside
781 ;; a label, i.e. the start of the last label before the start
782 ;; of the nonlabel part of the statement.
783 last-label-pos
784 ;; The last position where a label is possible provided the
785 ;; statement started there. It's nil as long as no invalid
786 ;; label content has been found (according to
787 ;; `c-nonlabel-token-key'). It's `start' if no valid label
788 ;; content was found in the label. Note that we might still
789 ;; regard it a label if it starts with `c-label-kwds'.
790 label-good-pos
791 ;; Putative positions of the components of a bitfield declaration,
792 ;; e.g. "int foo : NUM_FOO_BITS ;"
793 bitfield-type-pos bitfield-id-pos bitfield-size-pos
794 ;; Symbol just scanned back over (e.g. 'while or 'boundary).
795 ;; See above.
796 sym
797 ;; Current state in the automaton. See above.
798 state
799 ;; Current saved positions. See above.
800 saved-pos
801 ;; Stack of conses (state . saved-pos).
802 stack
803 ;; Regexp which matches "for", "if", etc.
806 ;; Return value.
808 ;; Positions of the last three sexps or bounds we've stopped at.
809 tok ptok pptok)
819 ;; Try to skip back over unary operator characters, to register
820 ;; that we've moved.
824 ;; Protect post-++/-- operators just before a virtual semicolon.
828 ;; Skip back over any semicolon here. If it was a bare semicolon, we're
829 ;; done. Later on we ignore the boundaries for statements that don't
830 ;; contain any sexp. The only thing that is affected is that the error
831 ;; checking is a little less strict, and we really don't bother.
841 pos saved)
843 ;; Begin at start and not pos to detect macros if we stand
844 ;; directly after the #.
847 ;; Record this as the first token if not starting inside it.
850 ;; The following while loop goes back one sexp (balanced parens,
851 ;; etc. with contents, or symbol or suchlike) each iteration. This
852 ;; movement is accomplished with a call to c-backward-sexp approx 170
853 ;; lines below.
854 ;;
855 ;; The loop is exited only by throwing nil to the (catch 'loop ...):
856 ;; 1. On reaching the start of a macro;
857 ;; 2. On having passed a stmt boundary with the PDA stack empty;
858 ;; 3. On reaching the start of an Objective C method def;
859 ;; 4. From macro `c-bos-pop-state'; when the stack is empty;
860 ;; 5. From macro `c-bos-pop-state-and-retry' when the stack is empty.
864 ;; Are we in a macro, just after the opening #?
879 ;; Stop at the first token in the content of the macro.
883 ret 'macro
884 ignore-labels t))
887 ;; Do a round through the automaton if we've just passed a
888 ;; statement boundary or passed a "while"-like token.
896 ;; The PDA state handling.
897 ;;
898 ;; Refer to the description of the PDA in the opening
899 ;; comments. In the following OR form, the first leaf
900 ;; attempts to handles one of the specific actions detailed
901 ;; (e.g., finding token "if" whilst in state `else-boundary').
902 ;; We drop through to the second leaf (which handles common
903 ;; state) if no specific handler is found in the first cond.
904 ;; If a parsing error is detected (e.g. an "else" with no
905 ;; preceding "if"), we throw to the enclosing catch.
906 ;;
907 ;; Note that the (eq state 'else) means
908 ;; "we've just passed an else", NOT "we're looking for an
909 ;; else".
926 ;; Since this can cause backtracking we do a
927 ;; little more careful analysis to avoid it:
928 ;; If there's a label in front of the while
929 ;; it can't be part of a do-while.
958 ;; This is state common. We get here when the previous
959 ;; cond statement found no particular state handler.
961 ;; If we have a boundary at the start
962 ;; position we push a frame to go to the
963 ;; previous statement.
972 ;; Is this a real while, or a do-while?
973 ;; The next `when' triggers unless we are SURE that
974 ;; the `while' is not the tail end of a `do-while'.
977 ;; The following kludge is to prevent
978 ;; infinite recursion when called from
979 ;; c-awk-after-if-for-while-condition-p,
980 ;; or the like.
984 ;; Since this can cause backtracking we do a
985 ;; little more careful analysis to avoid it: If
986 ;; the while isn't followed by a (possibly
987 ;; virtual) semicolon it can't be a do-while.
996 ;; We're either past a statement boundary or at the
997 ;; start of a statement, so throw away any label data
998 ;; for the previous one.
1000 last-label-pos nil
1001 c-maybe-labelp nil))))
1003 ;; Step to the previous sexp, but not if we crossed a
1004 ;; boundary, since that doesn't consume an sexp.
1008 ;; HERE IS THE SINGLE PLACE INSIDE THE PDA LOOP WHERE WE MOVE
1009 ;; BACKWARDS THROUGH THE SOURCE.
1013 ;; The end position of the area to search for statement
1014 ;; barriers in this round.
1017 ;; Go back over exactly one logical sexp, taking proper
1018 ;; account of macros and escaped EOLs.
1022 ;; Give up if we hit an unbalanced block. Since the
1023 ;; stack won't be empty the code below will report a
1024 ;; suitable error.
1027 ;; Have we moved into a macro?
1030 ;; Have we crossed a statement boundary? If not,
1031 ;; keep going back until we find one or a "real" sexp.
1038 ;; Have we just gone back over an escaped NL? This
1039 ;; doesn't count as a sexp.
1042 ;; Have we crossed a statement boundary?
1045 ;; Are we at a macro beginning?
1047 c-opt-cpp-prefix
1053 ;; Just gone back over a brace block?
1062 ;; Just gone back over some paren block?
1066 before-sws-pos)))
1069 ;; Just gone back over an ordinary symbol of some sort?
1075 ptok tok
1076 tok boundary-pos
1078 ;; Like a C "continue". Analyze the next sexp.
1081 ;; ObjC method def?
1088 ;; Might we have a bitfield declaration, "<type> <id> : <size>"?
1091 ;; The : <size> and <id> fields?
1101 ;; The <type> field?
1109 bitfield-id-pos nil)))))
1111 ;; Handle labels.
1114 ;; `c-crosses-statement-barrier-p' has found a colon, so we
1115 ;; might be in a label now. Have we got a real label
1116 ;; (including a case label) or something like C++'s "public:"?
1117 ;; A case label might use an expression rather than a token.
1120 ;; Catch C++'s inheritance construct "class foo : bar".
1134 ; been found.
1136 ;; We're in a potential label and it's the first
1137 ;; time we've found something that isn't allowed in
1138 ;; one.
1141 ;; We've moved back by a sexp, so update the token positions.
1143 pptok ptok
1144 ptok tok
1150 ;; If the stack isn't empty there might be errors to report.
1159 ;; Need to investigate closer whether we've crossed
1160 ;; between a substatement and its containing statement.
1164 pptok)
1166 ptok)
1174 after-labels-pos
1181 last-label-pos
1182 pos))
1184 ;; We're in a label. Maybe we should step to the statement
1185 ;; after it.
1190 ;; Might have jumped over several labels. Go to the last one.
1193 ;; Have we got "case <expression>:"?
1203 ;; Skip over the unary operators that can start the statement.
1206 ;; protect AWK post-inc/decrement operators, etc.
1211 ret)))
1214 "Return non-nil if there is a non-comment non-macro punctuation character
1215 between FROM and TO. FROM must not be in a string or comment. The returned
1216 value is the position of the first such character."
1227 "Return non-nil if buffer positions FROM to TO cross one or more
1228 statement or declaration boundaries. The returned value is actually
1229 the position of the earliest boundary char. FROM must not be within
1230 a string or comment.
1232 The variable `c-maybe-labelp' is set to the position of the first `:' that
1233 might start a label (i.e. not part of `::' and not preceded by `?'). If a
1234 single `?' is found, then `c-maybe-labelp' is cleared.
1236 For AWK, a statement which is terminated by an EOL (not a ; or a }) is
1238 the line. If this virtual semicolon is _at_ from, the function recognizes it.
1240 Note that this function might do hidden buffer changes. See the
1241 comment at the start of cc-engine.el for more info."
1243 ;; If the current language has CPP macros, insert # into skip-chars.
1246 c-opt-cpp-symbol ; usually "#"
1248 c-stmt-delim-chars))
1251 lit-range vsemi-pos)
1258 skip-chars
1262 ;; Virtual semicolon?
1272 ;; In a string/comment?
1279 ;; Ignore scope operators.
1283 ;; A question mark. Can't be a label, so stop
1284 ;; looking for more : and ?.
1287 ;; At a CPP construct or a "#" or "##" operator?
1298 ;; In trailing space after an as yet undetected virtual semicolon?
1301 ; unterminated string/regexp.
1306 nil))))))
1309 "Return non-nil if the point is at the first token in a statement
1310 or somewhere in the syntactic whitespace before it.
1313 Any kind of declaration-like construct that occur outside function
1316 Note that this function might do hidden buffer changes. See the
1317 comment at the start of cc-engine.el for more info."
1321 c-maybe-labelp)
1332 "Return non-nil if the point is at the first token in an expression or
1333 statement, or somewhere in the syntactic whitespace before it.
1336 grammar sense: It's any sequence of expression tokens except commas,
1337 unless they are enclosed inside parentheses of some kind. Also, an
1338 expression never continues past an enclosing parenthesis, but it might
1339 contain parenthesis pairs of any sort except braces.
1341 Since expressions never cross statement boundaries, this function also
1342 recognizes statement beginnings, just like `c-at-statement-start-p'.
1344 Note that this function might do hidden buffer changes. See the
1345 comment at the start of cc-engine.el for more info."
1350 c-maybe-labelp)
1358 \f
1359 ;; A set of functions that covers various idiosyncrasies in
1360 ;; implementations of `forward-comment'.
1362 ;; Note: Some emacsen considers incorrectly that any line comment
1363 ;; ending with a backslash continues to the next line. I can't think
1364 ;; of any way to work around that in a reliable way without changing
1365 ;; the buffer, though. Suggestions welcome. ;) (No, temporarily
1366 ;; changing the syntax for backslash doesn't work since we must treat
1367 ;; escapes in string literals correctly.)
1370 "Move forward past whitespace and the closest following comment, if any.
1371 Return t if a comment was found, nil otherwise. In either case, the
1372 point is moved past the following whitespace. Line continuations,
1373 i.e. a backslashes followed by line breaks, are treated as whitespace.
1374 The line breaks that end line comments are considered to be the
1375 comment enders, so the point will be put on the beginning of the next
1376 line if it moved past a line comment.
1378 This function does not do any hidden buffer changes."
1386 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1387 ;; forwards at eob.
1388 nil
1390 ;; Emacs includes the ending newline in a b-style (c++)
1391 ;; comment, but XEmacs doesn't. We depend on the Emacs
1392 ;; behavior (which also is symmetric).
1396 t))))
1399 "Move forward past all following whitespace and comments.
1400 Line continuations, i.e. a backslashes followed by line breaks, are
1401 treated as whitespace.
1403 Note that this function might do hidden buffer changes. See the
1404 comment at the start of cc-engine.el for more info."
1407 ;; If forward-comment in at least XEmacs 21 is given a large
1408 ;; positive value, it'll loop all the way through if it hits
1409 ;; eob.
1411 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1412 ;; forwards at eob.
1417 t))))
1420 "Move backward past whitespace and the closest preceding comment, if any.
1421 Return t if a comment was found, nil otherwise. In either case, the
1422 point is moved past the preceding whitespace. Line continuations,
1423 i.e. a backslashes followed by line breaks, are treated as whitespace.
1424 The line breaks that end line comments are considered to be the
1425 comment enders, so the point cannot be at the end of the same line to
1426 move over a line comment.
1428 This function does not do any hidden buffer changes."
1431 ;; When we got newline terminated comments, forward-comment in all
1432 ;; supported emacsen so far will stop at eol of each line not
1433 ;; ending with a comment when moving backwards. This corrects for
1434 ;; that, and at the same time handles line continuations.
1442 ;; Some emacsen (e.g. Emacs 19.34) return t when moving
1443 ;; backwards at bob.
1444 nil
1446 ;; Leave point after the closest following newline if we've
1447 ;; backed up over any above, since forward-comment won't move
1448 ;; backward over a line comment if point is at the end of the
1449 ;; same line.
1454 ;; If forward-comment above succeeded and we're at eol
1455 ;; then the newline we moved over above didn't end a
1456 ;; line comment, so we give it another go.
1459 t))
1461 ;; Emacs <= 20 and XEmacs move back over the closer of a
1462 ;; block comment that lacks an opener.
1465 t)))))
1468 "Move backward past all preceding whitespace and comments.
1469 Line continuations, i.e. a backslashes followed by line breaks, are
1470 treated as whitespace. The line breaks that end line comments are
1471 considered to be the comment enders, so the point cannot be at the end
1472 of the same line to move over a line comment. Unlike
1473 c-backward-syntactic-ws, this function doesn't move back over
1474 preprocessor directives.
1476 Note that this function might do hidden buffer changes. See the
1477 comment at the start of cc-engine.el for more info."
1481 ;; `forward-comment' in some emacsen (e.g. XEmacs 21.4)
1482 ;; return t when moving backwards at bob.
1488 ;; Cope specifically with ^M^J here -
1489 ;; forward-comment sometimes gets stuck after ^Ms,
1490 ;; sometimes after ^M^J.
1494 t)
1498 t))))
1499 moved-comment)
1501 ;; Emacs <= 20 and XEmacs move back over the
1502 ;; closer of a block comment that lacks an opener.
1504 t)
1506 ;; XEmacs treats line continuations as whitespace but
1507 ;; only in the backward direction, which seems a bit
1508 ;; odd. Anyway, this is necessary for Emacs.
1513 t))))))
1515 \f
1516 ;; Tools for skipping over syntactic whitespace.
1518 ;; The following functions use text properties to cache searches over
1519 ;; large regions of syntactic whitespace. It works as follows:
1520 ;;
1521 ;; o If a syntactic whitespace region contains anything but simple
1522 ;; whitespace (i.e. space, tab and line breaks), the text property
1523 ;; `c-in-sws' is put over it. At places where we have stopped
1524 ;; within that region there's also a `c-is-sws' text property.
1525 ;; That since there typically are nested whitespace inside that
1526 ;; must be handled separately, e.g. whitespace inside a comment or
1527 ;; cpp directive. Thus, from one point with `c-is-sws' it's safe
1528 ;; to jump to another point with that property within the same
1529 ;; `c-in-sws' region. It can be likened to a ladder where
1530 ;; `c-in-sws' marks the bars and `c-is-sws' the rungs.
1531 ;;
1532 ;; o The `c-is-sws' property is put on the simple whitespace chars at
1533 ;; a "rung position" and also maybe on the first following char.
1534 ;; As many characters as can be conveniently found in this range
1535 ;; are marked, but no assumption can be made that the whole range
1536 ;; is marked (it could be clobbered by later changes, for
1537 ;; instance).
1538 ;;
1539 ;; Note that some part of the beginning of a sequence of simple
1540 ;; whitespace might be part of the end of a preceding line comment
1541 ;; or cpp directive and must not be considered part of the "rung".
1542 ;; Such whitespace is some amount of horizontal whitespace followed
1543 ;; by a newline. In the case of cpp directives it could also be
1544 ;; two newlines with horizontal whitespace between them.
1545 ;;
1546 ;; The reason to include the first following char is to cope with
1547 ;; "rung positions" that doesn't have any ordinary whitespace. If
1548 ;; `c-is-sws' is put on a token character it does not have
1549 ;; `c-in-sws' set simultaneously. That's the only case when that
1550 ;; can occur, and the reason for not extending the `c-in-sws'
1551 ;; region to cover it is that the `c-in-sws' region could then be
1552 ;; accidentally merged with a following one if the token is only
1553 ;; one character long.
1554 ;;
1555 ;; o On buffer changes the `c-in-sws' and `c-is-sws' properties are
1556 ;; removed in the changed region. If the change was inside
1557 ;; syntactic whitespace that means that the "ladder" is broken, but
1558 ;; a later call to `c-forward-sws' or `c-backward-sws' will use the
1559 ;; parts on either side and use an ordinary search only to "repair"
1560 ;; the gap.
1561 ;;
1562 ;; Special care needs to be taken if a region is removed: If there
1563 ;; are `c-in-sws' on both sides of it which do not connect inside
1564 ;; the region then they can't be joined. If e.g. a marked macro is
1565 ;; broken, syntactic whitespace inside the new text might be
1566 ;; marked. If those marks would become connected with the old
1567 ;; `c-in-sws' range around the macro then we could get a ladder
1568 ;; with one end outside the macro and the other at some whitespace
1569 ;; within it.
1570 ;;
1571 ;; The main motivation for this system is to increase the speed in
1572 ;; skipping over the large whitespace regions that can occur at the
1573 ;; top level in e.g. header files that contain a lot of comments and
1574 ;; cpp directives. For small comments inside code it's probably
1575 ;; slower than using `forward-comment' straightforwardly, but speed is
1576 ;; not a significant factor there anyway.
1578 ; (defface c-debug-is-sws-face
1579 ; '((t (:background "GreenYellow")))
1580 ; "Debug face to mark the `c-is-sws' property.")
1581 ; (defface c-debug-in-sws-face
1582 ; '((t (:underline t)))
1583 ; "Debug face to mark the `c-in-sws' property.")
1585 ; (defun c-debug-put-sws-faces ()
1586 ; ;; Put the sws debug faces on all the `c-is-sws' and `c-in-sws'
1587 ; ;; properties in the buffer.
1588 ; (interactive)
1589 ; (save-excursion
1590 ; (c-save-buffer-state (in-face)
1591 ; (goto-char (point-min))
1592 ; (setq in-face (if (get-text-property (point) 'c-is-sws)
1593 ; (point)))
1594 ; (while (progn
1595 ; (goto-char (next-single-property-change
1596 ; (point) 'c-is-sws nil (point-max)))
1597 ; (if in-face
1598 ; (progn
1599 ; (c-debug-add-face in-face (point) 'c-debug-is-sws-face)
1600 ; (setq in-face nil))
1601 ; (setq in-face (point)))
1602 ; (not (eobp))))
1603 ; (goto-char (point-min))
1604 ; (setq in-face (if (get-text-property (point) 'c-in-sws)
1605 ; (point)))
1606 ; (while (progn
1607 ; (goto-char (next-single-property-change
1608 ; (point) 'c-in-sws nil (point-max)))
1609 ; (if in-face
1610 ; (progn
1611 ; (c-debug-add-face in-face (point) 'c-debug-in-sws-face)
1612 ; (setq in-face nil))
1613 ; (setq in-face (point)))
1614 ; (not (eobp)))))))
1617 ;;`(message ,@args)
1618 )
1621 ;; This macro does a hidden buffer change.
1628 ;; This macro does a hidden buffer change.
1635 ;; This macro does a hidden buffer change.
1642 ;; This macro does a hidden buffer change.
1649 ;; This macro does a hidden buffer change.
1657 ;; Called from `after-change-functions'. Note that if
1658 ;; `c-forward-sws' or `c-backward-sws' are used outside
1659 ;; `c-save-buffer-state' or similar then this will remove the cache
1660 ;; properties right after they're added.
1661 ;;
1662 ;; This function does hidden buffer changes.
1665 ;; Adjust the end to remove the properties in any following simple
1666 ;; ws up to and including the next line break, if there is any
1667 ;; after the changed region. This is necessary e.g. when a rung
1668 ;; marked empty line is converted to a line comment by inserting
1669 ;; "//" before the line break. In that case the line break would
1670 ;; keep the rung mark which could make a later `c-backward-sws'
1671 ;; move into the line comment instead of over it.
1681 ;; Ensure that an `c-in-sws' range gets broken. Note that it isn't
1682 ;; safe to keep a range that was continuous before the change. E.g:
1683 ;;
1684 ;; #define foo
1685 ;; \
1686 ;; bar
1687 ;;
1688 ;; There can be a "ladder" between "#" and "b". Now, if the newline
1689 ;; after "foo" is removed then "bar" will become part of the cpp
1690 ;; directive instead of a syntactically relevant token. In that
1691 ;; case there's no longer syntactic ws from "#" to "b".
1698 ;; Used by `c-forward-syntactic-ws' to implement the unbounded search.
1699 ;;
1700 ;; This function might do hidden buffer changes.
1703 ;; unmarked part of the simple ws region.
1705 rung-is-marked next-rung-is-marked simple-ws-end
1706 ;; `safe-start' is set when it's safe to cache the start position.
1707 ;; It's not set if we've initially skipped over comments and line
1708 ;; continuations since we might have gone out through the end of a
1709 ;; macro then. This provision makes `c-forward-sws' not populate the
1710 ;; cache in the majority of cases, but otoh is `c-backward-sws' by far
1711 ;; more common.
1712 safe-start)
1714 ;; Skip simple ws and do a quick check on the following character to see
1715 ;; if it's anything that can't start syntactic ws, so we can bail out
1716 ;; early in the majority of cases when there just are a few ws chars.
1723 ;; Find the last rung position to avoid setting properties in all
1724 ;; the cases when the marked rung is complete.
1725 ;; (`next-single-property-change' is certain to move at least one
1726 ;; step forward.)
1729 ;; Got no marked rung here. Since the simple ws might have started
1730 ;; inside a line comment or cpp directive we must set `rung-pos' as
1731 ;; high as possible.
1741 ;; The following search is the main reason that `c-in-sws'
1742 ;; and `c-is-sws' aren't combined to one property.
1746 ;; If the `c-in-sws' region extended past the last
1747 ;; `c-is-sws' char we have to go back a bit.
1754 "c-forward-sws cached move %s -> %s (max %s)"
1761 ;; We'll loop here if there is simple ws after the last rung.
1762 ;; That means that there's been some change in it and it's
1763 ;; possible that we've stepped into another ladder, so extend
1764 ;; the previous one to join with it if there is one, and try to
1765 ;; use the cache again.
1767 "c-forward-sws extending rung with [%s..%s] (max %s)"
1770 ;; Remove any `c-in-sws' property from the last char of
1771 ;; the rung before we mark it with `c-is-sws', so that we
1772 ;; won't connect with the remains of a broken "ladder".
1778 last-put-in-sws-pos rung-pos)))
1785 ;; Skipped over comments. Don't cache at eob in case the buffer
1786 ;; is narrowed.
1797 ;; Skip a preprocessor directive.
1804 ;; Don't cache at eob in case the buffer is narrowed.
1807 ;; We've searched over a piece of non-white syntactic ws. See if this
1808 ;; can be cached.
1814 ;; Cache if we haven't skipped comments only, and if we started
1815 ;; either from a marked rung or from a completely uncached
1816 ;; position.
1821 ;; See if there's a marked rung in the encountered simple ws. If
1822 ;; so then we can cache, unless `safe-start' is nil. Even then
1823 ;; we need to do this to check if the cache can be used for the
1824 ;; next step.
1828 safe-start))
1832 "c-forward-sws caching [%s..%s] - [%s..%s] (max %s)"
1836 ;; Remove the properties for any nested ws that might be cached.
1837 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
1838 ;; anyway.
1846 last-put-in-sws-pos rung-pos))
1848 ;; Remove any `c-in-sws' property from the last char of
1849 ;; the rung before we mark it with `c-is-sws', so that we
1850 ;; won't connect with the remains of a broken "ladder".
1853 rung-end-pos))
1856 "c-forward-sws not caching [%s..%s] - [%s..%s] (max %s)"
1860 ;; Set `rung-pos' for the next rung. It's the same thing here as
1861 ;; initially, except that the rung position is set as early as
1862 ;; possible since we can't be in the ending ws of a line comment or
1863 ;; cpp directive now.
1870 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
1871 ;; another one after the point (which might occur when editing inside a
1872 ;; comment or macro).
1876 "c-forward-sws clearing at %s for cache separation"
1877 last-put-in-sws-pos)
1881 ;; If at eob we have to clear the last character before the end
1882 ;; instead since the buffer might be narrowed and there might
1883 ;; be a `c-in-sws' after (point-max). In this case it's
1884 ;; necessary to clear both properties.
1886 "c-forward-sws clearing thoroughly at %s for cache separation"
1889 last-put-in-sws-pos))))
1890 ))))
1893 ;; Used by `c-backward-syntactic-ws' to implement the unbounded search.
1894 ;;
1895 ;; This function might do hidden buffer changes.
1898 ;; part of the simple ws region.
1900 rung-is-marked simple-ws-beg cmt-skip-pos)
1902 ;; Skip simple horizontal ws and do a quick check on the preceding
1903 ;; character to see if it's anything that can't end syntactic ws, so we can
1904 ;; bail out early in the majority of cases when there just are a few ws
1905 ;; chars. Newlines are complicated in the backward direction, so we can't
1906 ;; skip over them.
1913 ;; Try to find a rung position in the simple ws preceding point, so that
1914 ;; we can get a cache hit even if the last bit of the simple ws has
1915 ;; changed recently.
1921 ;; `rung-pos' will be the earliest marked position, which means that
1922 ;; there might be later unmarked parts in the simple ws region.
1923 ;; It's not worth the effort to fix that; the last part of the
1924 ;; simple ws is also typically edited often, so it could be wasted.
1936 ;; The following search is the main reason that `c-in-sws'
1937 ;; and `c-is-sws' aren't combined to one property.
1941 ;; If the `c-in-sws' region extended past the first
1942 ;; `c-is-sws' char we have to go forward a bit.
1947 "c-backward-sws cached move %s <- %s (min %s)"
1959 t
1961 nil))
1963 ;; We'll loop here if there is simple ws before the first rung.
1964 ;; That means that there's been some change in it and it's
1965 ;; possible that we've stepped into another ladder, so extend
1966 ;; the previous one to join with it if there is one, and try to
1967 ;; use the cache again.
1969 "c-backward-sws extending rung with [%s..%s] (min %s)"
1972 ;; Remove any `c-in-sws' property from the last char of
1973 ;; the rung before we mark it with `c-is-sws', so that we
1974 ;; won't connect with the remains of a broken "ladder".
1977 rung-pos)
1981 last-put-in-sws-pos rung-pos))
1990 ;; Inside a cpp directive. See if it should be skipped over.
1993 ;; Move back over all line continuations in the region skipped
1994 ;; over by `c-backward-comments'. If we go past it then we
1995 ;; started inside the cpp directive.
2004 ;; Don't move past the cpp directive if we began inside
2005 ;; it. Note that the position at the end of the last line
2006 ;; of the macro is also considered to be within it.
2008 nil)
2010 ;; It's worthwhile to spend a little bit of effort on finding
2011 ;; the end of the macro, to get a good `simple-ws-beg'
2012 ;; position for the cache. Note that `c-backward-comments'
2013 ;; could have stepped over some comments before going into
2014 ;; the macro, and then `simple-ws-beg' must be kept on the
2015 ;; same side of those comments.
2022 ;; Might happen if comments after the macro were skipped
2023 ;; over.
2027 t)))
2032 simple-ws-beg)
2033 ;; Skipped over comments. Must put point at the end of
2034 ;; the simple ws at point since we might be after a line
2035 ;; comment or cpp directive that's been partially
2036 ;; narrowed out, and we can't risk marking the simple ws
2037 ;; at the end of it.
2039 t)))
2041 ;; We've searched over a piece of non-white syntactic ws. See if this
2042 ;; can be cached.
2047 ;; Cache if we started either from a marked rung or from a
2048 ;; completely uncached position.
2049 rung-is-marked
2052 ;; Cache if there's a marked rung in the encountered simple ws.
2060 "c-backward-sws caching [%s..%s] - [%s..%s] (min %s)"
2065 ;; Remove the properties for any nested ws that might be cached.
2066 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
2067 ;; anyway.
2072 ;; Remove any `c-in-sws' property from the last char of
2073 ;; the rung before we mark it with `c-is-sws', so that we
2074 ;; won't connect with the remains of a broken "ladder".
2077 rung-end-pos)
2080 last-put-in-sws-pos simple-ws-beg)
2081 rung-pos)
2086 "c-backward-sws not caching [%s..%s] - [%s..%s] (min %s)"
2092 ))
2094 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
2095 ;; another one before the point (which might occur when editing inside a
2096 ;; comment or macro).
2100 "c-backward-sws clearing at %s for cache separation"
2103 last-put-in-sws-pos))
2105 ;; If at bob and the buffer is narrowed, we have to clear the
2106 ;; character we're standing on instead since there might be a
2107 ;; `c-in-sws' before (point-min). In this case it's necessary
2108 ;; to clear both properties.
2110 "c-backward-sws clearing thoroughly at %s for cache separation"
2111 last-put-in-sws-pos)
2114 ))))
2116 \f
2117 ;; Other whitespace tools
2119 ;; Is the region (beg end) WS, and is there WS (or BOB/EOB) next to the
2120 ;; region? This is a "heuristic" function. .....
2121 ;;
2122 ;; The motivation for the second bit is to check whether removing this
2123 ;; region would coalesce two symbols.
2124 ;;
2125 ;; FIXME!!! This function doesn't check virtual semicolons in any way. Be
2126 ;; careful about using this function for, e.g. AWK. (2007/3/7)
2135 \f
2136 ;; A system for finding noteworthy parens before the point.
2139 ;; A maximum comfortable scanning distance, e.g. between
2140 ;; `c-state-cache-good-pos' and "HERE" (where we call c-parse-state). When
2141 ;; this distance is exceeded, we take "emergency measures", e.g. by clearing
2142 ;; the cache and starting again from point-min or a beginning of defun. This
2143 ;; value can be tuned for efficiency or set to a lower value for testing.
2147 ;; The state cache used by `c-parse-state' to cut down the amount of
2148 ;; searching. It's the result from some earlier `c-parse-state' call. See
2149 ;; `c-parse-state''s doc string for details of its structure.
2150 ;;
2151 ;; The use of the cached info is more effective if the next
2152 ;; `c-parse-state' call is on a line close by the one the cached state
2153 ;; was made at; the cache can actually slow down a little if the
2154 ;; cached state was made very far back in the buffer. The cache is
2155 ;; most effective if `c-parse-state' is used on each line while moving
2156 ;; forward.
2160 ;; This is a position where `c-state-cache' is known to be correct, or
2161 ;; nil (see below). It's a position inside one of the recorded unclosed
2162 ;; parens or the top level, but not further nested inside any literal or
2163 ;; subparen that is closed before the last recorded position.
2164 ;;
2165 ;; The exact position is chosen to try to be close to yet earlier than
2166 ;; the position where `c-state-cache' will be called next. Right now
2167 ;; the heuristic is to set it to the position after the last found
2168 ;; closing paren (of any type) before the line on which
2169 ;; `c-parse-state' was called. That is chosen primarily to work well
2170 ;; with refontification of the current line.
2171 ;;
2172 ;; 2009-07-28: When `c-state-point-min' and the last position where
2173 ;; `c-parse-state' or for which `c-invalidate-state-cache' was called, are
2174 ;; both in the same literal, there is no such "good position", and
2175 ;; c-state-cache-good-pos is then nil. This is the ONLY circumstance in which
2176 ;; it can be nil. In this case, `c-state-point-min-literal' will be non-nil.
2177 ;;
2178 ;; 2009-06-12: In a brace desert, c-state-cache-good-pos may also be in
2179 ;; the middle of the desert, as long as it is not within a brace pair
2180 ;; recorded in `c-state-cache' or a paren/bracket pair.
2182 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2183 ;; We maintain a simple cache of positions which aren't in a literal, so as to
2184 ;; speed up testing for non-literality.
2186 ;; The approximate interval between entries in `c-state-nonlit-pos-cache'.
2190 ;; A list of buffer positions which are known not to be in a literal or a cpp
2191 ;; construct. This is ordered with higher positions at the front of the list.
2192 ;; Only those which are less than `c-state-nonlit-pos-cache-limit' are valid.
2196 ;; An upper limit on valid entries in `c-state-nonlit-pos-cache'. This is
2197 ;; reduced by buffer changes, and increased by invocations of
2198 ;; `c-state-literal-at'.
2202 ;; A list of buffer positions which are known not to be in a literal. This is
2203 ;; ordered with higher positions at the front of the list. Only those which
2204 ;; are less than `c-state-semi-nonlit-pos-cache-limit' are valid.
2208 ;; An upper limit on valid entries in `c-state-semi-nonlit-pos-cache'. This is
2209 ;; reduced by buffer changes, and increased by invocations of
2210 ;; `c-state-literal-at'. FIXME!!!
2213 ;; Do a parse-partial-sexp from FROM to TO, returning either
2214 ;; (STATE TYPE (BEG . END)) if TO is in a literal; or
2215 ;; (STATE) otherwise,
2216 ;; where STATE is the parsing state at TO, TYPE is the type of the literal
2217 ;; (one of 'c, 'c++, 'string) and (BEG . END) is the boundaries of the literal.
2218 ;;
2219 ;; Unless NOT-IN-DELIMITER is non-nil, when TO is inside a two-character
2220 ;; comment opener, this is recognized as being in a comment literal.
2221 ;;
2222 ;; Only elements 3 (in a string), 4 (in a comment), 5 (following a quote),
2223 ;; 7 (comment type) and 8 (start of comment/string) (and possibly 9) of
2224 ;; STATE are valid.
2227 ty co-st)
2235 nil ; TARGETDEPTH
2236 nil ; STOPBEFORE
2237 s ; OLDSTATE
2254 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2255 ;; string, comment, or macro.
2256 ;;
2257 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2258 ;; MAY NOT contain any positions within macros, since macros are frequently
2259 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2260 ;; We cannot rely on this mechanism whilst determining a cache pos since
2261 ;; this function is also called from outwith `c-parse-state'.
2266 pos npos high-pos lit macro-beg macro-end)
2267 ;; Trim the cache to take account of buffer changes.
2279 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2285 ;; Test for being in a literal. If so, go to after it.
2292 ;; Test for being in a macro. If so, go to after it.
2307 ;; Add one extra element above HERE so as to to avoid the previous
2308 ;; expensive calculation when the next call is close to the current
2309 ;; one. This is especially useful when inside a large macro.
2316 pos))))
2319 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2320 ;; string or comment. It may be in a macro.
2325 pos npos high-pos lit macro-beg macro-end)
2326 ;; Trim the cache to take account of buffer changes.
2338 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2342 ;; Test for being in a literal. If so, go to after it.
2355 pos))))
2358 ;; If position HERE is inside a literal, return (START . END), the
2359 ;; boundaries of the literal (which may be outside the accessible bit of the
2360 ;; buffer). Otherwise, return nil.
2361 ;;
2362 ;; This function is almost the same as `c-literal-limits'. Previously, it
2363 ;; differed in that it was a lower level function, and that it rigorously
2364 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2365 ;; virtually identical to this function.
2373 ;; Return the start of the literal containing POS, or POS itself.
2375 pos))
2378 ;; Return a position outside of a string/comment/macro at or before POS.
2379 ;; STATE is the parse-partial-sexp state at POS.
2383 pos)))
2388 res))))
2390 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2391 ;; Stuff to do with point-min, and coping with any literal there.
2394 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2395 ;; narrowing is likely to affect the parens that are visible before the point.
2401 ;; These two variables define the literal, if any, containing point-min.
2402 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2403 ;; literal. If there's no literal there, they're both nil.
2407 ;; This is the earliest buffer-pos from which scanning can be done. It is
2408 ;; either the end of the literal containing point-min, or point-min itself.
2409 ;; It becomes nil if the buffer is changed earlier than this point.
2411 ;; Return the lowest valid scanning pos. This will be the end of the
2412 ;; literal enclosing point-min, or point-min itself.
2424 ;; Determine the properties of any literal containing POINT-MIN, setting the
2425 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2426 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2428 lit)
2443 c-state-point-min-lit-start nil
2444 c-state-min-scan-pos p-min)))))
2447 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2448 ;; A variable which signals a brace dessert - helpful for reducing the number
2449 ;; of fruitless backward scans.
2452 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2453 ;; that defun has searched backwards for a brace pair and not found one. Its
2454 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2455 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2456 ;; nil when at top level) and FROM is where the backward search started. It
2457 ;; is reset to nil in `c-invalidate-state-cache'.
2460 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2461 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2462 ;; list of like structure.
2464 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2465 ;; (default `c-state-cache') (or nil).
2472 ;; Return the address of the latest brace/bracket/paren (whether left or
2473 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2480 ;; Return the position just after the latest brace/bracket/paren (whether
2481 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2489 ;; From the state-cache, determine the buffer position from which we might
2490 ;; scan forward to HERE to update this cache. This position will be just
2491 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2492 ;; return the earliest position in the accessible region which isn't within
2493 ;; a literal. If the visible portion of the buffer is entirely within a
2494 ;; literal, return NIL.
2496 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2512 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2513 ;; Variables which keep track of preprocessor constructs.
2522 ;; These are the limits of the macro containing point at the previous call of
2523 ;; `c-parse-state', or nil.
2525 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2526 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2528 ;; Return the position of the opening paren/brace/bracket before HERE- which
2529 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2530 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2531 ;;
2532 ;; ............................................
2533 ;; | |
2534 ;; ( [ ( .........#macro.. ) ( ) ] )
2535 ;; ^ ^ ^ ^
2536 ;; | | | |
2537 ;; return HERE- HERE+ TOP
2538 ;;
2539 ;; If there aren't enough opening paren/brace/brackets, return the position
2540 ;; of the outermost one found, or HERE- if there are none. If there are no
2541 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2542 ;; must not be inside literals. Only the accessible portion of the buffer
2543 ;; will be scanned.
2545 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2546 ;; `here'. Go round the next loop each time we pass over such a ")". These
2547 ;; probably match "("s before `here-'.
2559 ;; PART 2: Scan back before `here-' searching for the "("s
2560 ;; matching/mismatching the ")"s found above. We only need to direct the
2561 ;; caller to scan when we've encountered unmatched right parens.
2570 pos))
2573 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2574 ;; to minimize the amount of scanning. HERE is the pertinent position in
2575 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2576 ;; its head trimmed) is known to be good, or nil if there is no such
2577 ;; position.
2578 ;;
2579 ;; The return value is a list, one of the following:
2580 ;;
2581 ;; o - ('forward START-POINT) - scan forward from START-POINT,
2582 ;; which is not less than the highest position in `c-state-cache' below HERE,
2583 ;; which is after GOOD-POS.
2584 ;; o - ('backward nil) - scan backwards (from HERE).
2585 ;; o - ('back-and-forward START-POINT) - like 'forward, but when HERE is earlier
2586 ;; than GOOD-POS.
2587 ;; o - ('IN-LIT nil) - point is inside the literal containing point-min.
2589 strategy ; 'forward, 'backward, or 'IN-LIT.
2590 start-point)
2602 start-point cache-pos)))
2606 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2607 ;; Routines which change `c-state-cache' and associated values.
2609 ;; The region (more precisely, point-min) has changed since we
2610 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2612 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2613 ;; It would be possible to do a better job here and recalculate the top
2614 ;; only.
2618 c-state-cache-good-pos c-state-min-scan-pos
2619 c-state-brace-pair-desert nil))
2621 ;; point-min has MOVED FORWARD.
2623 ;; Is the new point-min inside a (different) literal?
2628 ;; Cut off a bit of the tail from `c-state-cache'.
2630 pa)
2638 c-state-cache-good-pos c-state-min-scan-pos)
2641 )))
2646 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2647 ;; of nesting (not necessarily immediately preceding), push a cons onto
2648 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2649 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2650 ;; UPPER-LIM.
2651 ;;
2652 ;; Return non-nil when this has been done.
2653 ;;
2654 ;; The situation it copes with is this transformation:
2655 ;;
2656 ;; OLD: { (.) {...........}
2657 ;; ^ ^
2658 ;; FROM HERE
2659 ;;
2660 ;; NEW: { {....} (.) {.........
2661 ;; ^ ^ ^
2662 ;; LOWER BRACE PAIR HERE or HERE
2663 ;;
2664 ;; This routine should be fast. Since it can get called a LOT, we maintain
2665 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2666 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2676 ;; Don't start scanning in the middle of a CPP construct unless
2677 ;; it contains HERE - these constructs, in Emacs, are "commented
2678 ;; out" with category properties.
2681 macro-start-or-from
2682 from))
2686 ;; If we're essentially repeating a fruitless search, just give up.
2692 ;; DESERT-LIM. Avoid repeated searching through the cached desert.
2698 ;; CACHE-LIM. This limit will be necessary when an opening
2699 ;; paren at `cache-pos' has just had its matching close paren
2700 ;; inserted into the buffer. `cache-pos' continues to be a
2701 ;; search bound, even though the algorithm below would skip
2702 ;; over the new paren pair.
2711 ;; The top limit is EOB to ensure that `bra' is inside the
2712 ;; accessible part of the buffer at the next scan operation.
2715 ;; In the next pair of nested loops, the inner one moves back past a
2716 ;; pair of (mis-)matching parens or brackets; the outer one moves
2717 ;; back over a sequence of unmatched close brace/paren/bracket each
2718 ;; time round.
2734 ; brace, ignore it.
2737 ;; We've found the desired brace-pair.
2749 ;; We haven't found a brace pair. Record this in the cache.
2752 bra
2757 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2758 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2759 ;; "push" "a" brace pair onto `c-state-cache'.
2760 ;;
2761 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2762 ;; otherwise push it normally.
2763 ;;
2764 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2765 ;; latter is inside a macro, not being a macro containing
2766 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2767 ;; base pair. This latter case is assumed to be rare.
2768 ;;
2769 ;; Note: POINT is not preserved in this routine.
2779 c-state-cache)))
2780 ;; N.B. This defsubst codes one method for the simple, normal case,
2781 ;; and a more sophisticated, slower way for the general case. Don't
2782 ;; eliminate this defsubst - it's a speed optimization.
2786 ;; Scan the buffer from FROM to HERE, adding elements into `c-state-cache'
2787 ;; for braces etc. Return a candidate for `c-state-cache-good-pos'.
2788 ;;
2789 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2790 ;; any. Typically, it is immediately after it. It must not be inside a
2791 ;; literal.
2797 here)))
2800 ; Is actually the pos. to scan for a (/{/[ from,
2801 ; which sometimes is after a silly )/}/].
2804 ; good-pos.
2812 ;; Each time round the following loop, we enter a successively deeper
2813 ;; level of brace/paren nesting. (Except sometimes we "continue at
2814 ;; the existing level".) `pa+1' is a pos inside an opening
2815 ;; brace/paren/bracket, usually just after it.
2818 ;; Each time round the next loop moves forward over an opening then
2819 ;; a closing brace/bracket/paren. This loop is white hot, so it
2820 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2821 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2822 ;; call of `scan-lists' signals an error, which happens when there
2823 ;; are no more b/b/p's to scan.
2834 ;; We've just entered a deeper nesting level.
2836 ;; Insert the brace pair (if present) and the single open
2837 ;; paren/brace/bracket into `c-state-cache' It cannot be
2838 ;; inside a macro, except one around point, because of what
2839 ;; `c-neutralize-syntax-in-CPP' has done.
2841 ;; Insert the opening brace/bracket/paren position.
2843 ;; Clear admin stuff for the next more nested part of the scan.
2847 ;; All open p/b/b's at this nesting level, if any, have probably
2848 ;; been closed by matching/mismatching ones. We're probably
2849 ;; finished - we just need to check for having found an
2850 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2851 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2854 ;; Record the final, innermost, brace-pair if there is one.
2857 ;; Determine a good pos
2865 here-bol))
2874 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2875 ;; not be in it when it is amended for position HERE. This may involve
2876 ;; replacing a CONS element for a brace pair containing HERE with its car.
2877 ;; Additionally, the "outermost" open-brace entry before HERE will be
2878 ;; converted to a cons if the matching close-brace is below HERE.
2879 ;;
2880 ;; START-POINT is a "maximal" "safe position" - there must be no open
2881 ;; parens/braces/brackets between START-POINT and HERE.
2882 ;;
2883 ;; As a second thing, calculate the result of parse-partial-sexp at
2884 ;; PPS-POINT, w.r.t. START-POINT. The motivation here is that
2885 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2886 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2887 ;; needs to be FAST).
2888 ;;
2889 ;; Return a list (GOOD-POS SCAN-BACK-POS CONS-SEPARATED PPS-STATE), where
2890 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2891 ;; to be good (we aim for this to be as high as possible);
2892 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2893 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2894 ;; position to scan backwards from. It is the position of the "{" of the
2895 ;; last element to be removed from `c-state-cache', when that elt is a
2896 ;; cons, otherwise nil.
2897 ;; o - CONS-SEPARATED is t when a cons element in `c-state-cache' has been
2898 ;; replaced by its car because HERE lies inside the brace pair represented
2899 ;; by the cons.
2900 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2910 ; start-point or nil
2917 ; HERE), or nil.
2923 pps-state ; Will be 9 or 10 elements long.
2924 pos
2925 upper-lim ; ,beyond which `c-state-cache' entries are removed
2926 scan-back-pos
2927 cons-separated
2928 pair-beg pps-point-state target-depth)
2930 ;; Remove entries beyond HERE. Also remove any entries inside
2931 ;; a macro, unless HERE is in the same macro.
2936 here
2942 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2943 ;; RBrace and indicate we'll need to search backwards for a previous
2944 ;; brace pair.
2950 cons-separated t))
2952 ;; The next loop jumps forward out of a nested level of parens each
2953 ;; time round; the corresponding elements in `c-state-cache' are
2954 ;; removed. `pos' is just after the brace-pair or the open paren at
2955 ;; (car c-state-cache). There can be no open parens/braces/brackets
2956 ;; between `start-point'/`start-point-actual-macro-start' and HERE,
2957 ;; due to the interface spec to this function.
2960 in-macro-start)))
2962 ; marked by a `category' text property.
2963 start-point))
2974 ;; Do nothing when we've merely reached pps-point.
2975 )
2977 ;; Scan!
2981 target-depth
2982 nil pps-state))
2992 ;; We've just moved out of the paren pair containing the brace-pair
2993 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2994 ;; and is potentially where the open brace of a cons in
2995 ;; c-state-cache will be.
3001 ; containing Lparen
3005 ; together with enclosed brace pair.
3006 ;; (t nil) ; Ignore an unmated Rparen.
3007 )))
3012 nil nil ; TARGETDEPTH, STOPBEFORE
3013 pps-state)))
3015 ;; If the last paren pair we moved out of was actually a brace pair,
3016 ;; insert it into `c-state-cache'.
3021 c-state-cache)))
3026 ;; Strip stale elements of `c-state-cache' by moving backwards through the
3027 ;; buffer, and inform the caller of the scenario detected.
3028 ;;
3029 ;; HERE is the position we're setting `c-state-cache' for.
3030 ;; CACHE-POS (a locally bound variable) is just after the latest recorded
3031 ;; position in `c-state-cache' before HERE, or a position at or near
3032 ;; point-min which isn't in a literal.
3033 ;;
3034 ;; This function must only be called only when (> `c-state-cache-good-pos'
3035 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
3036 ;; optimized to eliminate (or minimize) scanning between these two
3037 ;; positions.
3038 ;;
3039 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
3040 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
3041 ;; could become so after missing elements are inserted into
3042 ;; `c-state-cache'. This is JUST AFTER an opening or closing
3043 ;; brace/paren/bracket which is already in `c-state-cache' or just before
3044 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
3045 ;; before `here''s line, or the start of the literal containing it.
3046 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
3047 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
3048 ;; to scan backwards from.
3049 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
3050 ;; POS and HERE which aren't recorded in `c-state-cache'.
3051 ;;
3052 ;; The comments in this defun use "paren" to mean parenthesis or square
3053 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
3054 ;;
3055 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
3056 ;; | | | | | |
3057 ;; CP E here D C good
3060 pa ren ; positions of "(" and ")"
3061 dropped-cons ; whether the last element dropped from `c-state-cache'
3062 ; was a cons (representing a brace-pair)
3063 good-pos ; see above.
3064 lit ; (START . END) of a literal containing some point.
3065 here-lit-start here-lit-end ; bounds of literal containing `here'
3066 ; or `here' itself.
3071 ;; Remove completely irrelevant entries from `c-state-cache'.
3077 ;; At this stage, (>= pos here);
3078 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3083 ;; CASE 1: The top of the cache is a brace pair which now encloses
3084 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3085 ;; knowledge of what's inside these braces, we have no alternative but
3086 ;; to direct the caller to scan the buffer from the opening brace.
3090 ; entry into a { entry, so the caller needs to
3091 ; search for a brace pair before the {.
3093 ;; `here' might be inside a literal. Check for this.
3098 ;; Has `here' just "newly entered" a macro?
3111 ;; `here' might be nested inside any depth of parens (or brackets but
3112 ;; not braces). Scan backwards to find the outermost such opening
3113 ;; paren, if there is one. This will be the scan position to return.
3120 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3121 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3122 ;; a brace pair preceding this, it will already be in `c-state-cache',
3123 ;; unless there was a brace pair after it, i.e. there'll only be one to
3124 ;; scan for if we've just deleted one.
3127 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3128 ;; Further forward scanning isn't needed, but we still need to find a
3129 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3141 ;; CASE 3: After a }/)/] before `here''s BOL.
3146 ;; CASE 3.5: Just after an existing entry in `c-state-cache' on `here''s
3147 ;; line or the previous line.
3151 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3152 ;; literal containing it.
3156 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3157 ;; Externally visible routines.
3161 c-state-cache-good-pos 1
3162 c-state-nonlit-pos-cache nil
3163 c-state-nonlit-pos-cache-limit 1
3164 c-state-semi-nonlit-pos-cache nil
3165 c-state-semi-nonlit-pos-cache-limit 1
3166 c-state-brace-pair-desert nil
3167 c-state-point-min 1
3168 c-state-point-min-lit-type nil
3169 c-state-point-min-lit-start nil
3170 c-state-min-scan-pos 1
3171 c-state-old-cpp-beg nil
3172 c-state-old-cpp-end nil)
3175 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3176 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3177 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3178 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3179 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3180 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3181 ;; (defun c-state-dump ()
3182 ;; ;; For debugging.
3183 ;; ;(message
3184 ;; (concat
3185 ;; (c-sc-qde c-state-cache)
3186 ;; (c-sc-de c-state-cache-good-pos)
3187 ;; (c-sc-qde c-state-nonlit-pos-cache)
3188 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3189 ;; (c-sc-qde c-state-brace-pair-desert)
3190 ;; (c-sc-de c-state-point-min)
3191 ;; (c-sc-de c-state-point-min-lit-type)
3192 ;; (c-sc-de c-state-point-min-lit-start)
3193 ;; (c-sc-de c-state-min-scan-pos)
3194 ;; (c-sc-de c-state-old-cpp-beg)
3195 ;; (c-sc-de c-state-old-cpp-end)))
3196 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3199 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3200 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3201 ;; left in a consistent state.
3202 ;;
3203 ;; This is much like `c-whack-state-after', but it never changes a paren
3204 ;; pair element into an open paren element. Doing that would mean that the
3205 ;; new open paren wouldn't have the required preceding paren pair element.
3206 ;;
3207 ;; This function is called from c-after-change.
3209 ;; The caches of non-literals:
3210 ;; Note that we use "<=" for the possibility of the second char of a two-char
3211 ;; comment opener being typed; this would invalidate any cache position at
3212 ;; HERE.
3218 ;; `c-state-cache':
3219 ;; Case 1: if `here' is in a literal containing point-min, everything
3220 ;; becomes (or is already) nil.
3224 c-state-cache-good-pos nil
3225 c-state-min-scan-pos nil)
3227 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3228 ;; below `here'. To maintain its consistency, we may need to insert a new
3229 ;; brace pair.
3232 too-high-pa ; recorded {/(/[ next above here, or nil.
3233 dropped-cons ; was the last removed element a brace pair?
3234 pa)
3235 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3242 ;; Do we need to add in an earlier brace pair, having lopped one off?
3249 ;; The brace-pair desert marker:
3257 ;; Find and record all noteworthy parens between some good point earlier in
3258 ;; the file and point. That good point is at least the beginning of the
3259 ;; top-level construct we are in, or the beginning of the preceding
3260 ;; top-level construct if we aren't in one.
3261 ;;
3262 ;; The returned value is a list of the noteworthy parens with the last one
3263 ;; first. If an element in the list is an integer, it's the position of an
3264 ;; open paren (of any type) which has not been closed before the point. If
3265 ;; an element is a cons, it gives the position of a closed BRACE paren
3266 ;; pair[*]; the car is the start brace position and the cdr is the position
3267 ;; following the closing brace. Only the last closed brace paren pair
3268 ;; before each open paren and before the point is recorded, and thus the
3269 ;; state never contains two cons elements in succession. When a close brace
3270 ;; has no matching open brace (e.g., the matching brace is outside the
3271 ;; visible region), it is not represented in the returned value.
3272 ;;
3273 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3274 ;; This defun explicitly treats mismatching parens/braces/brackets as
3275 ;; matching. It is the open brace which makes it a "brace" pair.
3276 ;;
3277 ;; If POINT is within a macro, open parens and brace pairs within
3278 ;; THIS macro MIGHT be recorded. This depends on whether their
3279 ;; syntactic properties have been suppressed by
3280 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3281 ;;
3282 ;; Currently no characters which are given paren syntax with the
3283 ;; syntax-table property are recorded, i.e. angle bracket arglist
3284 ;; parens are never present here. Note that this might change.
3285 ;;
3286 ;; BUG: This function doesn't cope entirely well with unbalanced
3287 ;; parens in macros. (2008-12-11: this has probably been resolved
3288 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3289 ;; following case the brace before the macro isn't balanced with the
3290 ;; one after it:
3291 ;;
3292 ;; {
3293 ;; #define X {
3294 ;; }
3295 ;;
3296 ;; Note to maintainers: this function DOES get called with point
3297 ;; within comments and strings, so don't assume it doesn't!
3298 ;;
3299 ;; This function might do hidden buffer changes.
3302 open-paren-in-column-0-is-defun-start
3303 strategy ; 'forward, 'backward etc..
3304 ;; Candidate positions to start scanning from:
3305 cache-pos ; highest position below HERE already existing in
3306 ; cache (or 1).
3307 good-pos
3308 start-point ; (when scanning forward) a place below HERE where there
3309 ; are no open parens/braces between it and HERE.
3310 bopl-state
3311 res
3312 cons-separated
3314 ;; If POINT-MIN has changed, adjust the cache
3318 ;; Strategy?
3323 ;; SCAN!
3331 ; start-point)
3341 good-pos)))
3353 good-pos)))
3357 c-state-cache-good-pos nil))))
3359 c-state-cache)
3362 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3363 ;;
3364 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3365 ;; of all parens in preprocessor constructs, except for any such construct
3366 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3367 ;; worrying further about macros and template delimiters.
3369 ;; Emacs
3374 c-state-old-cpp-beg
3379 ;; XEmacs
3383 ;; If PLACE is non-nil, return a marker marking it, otherwise nil.
3384 ;; We (re)use MARKER.
3390 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3391 ;; description of the functionality and return value.
3392 ;;
3393 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3394 ;; of all parens in preprocessor constructs, except for any such construct
3395 ;; containing point. We can then call `c-parse-state-1' without worrying
3396 ;; further about macros and template delimiters.
3403 here-cpp-beg)
3405 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3406 ;; subsystem.
3409 ;; Emacs
3413 here-cpp-beg here-cpp-end
3417 ;; XEmacs
3421 c-state-old-cpp-end
3424 ;; Debug tool to catch cache inconsistencies. This is called from
3425 ;; 000tests.el.
3445 c-state-cache-good-pos
3446 c-state-nonlit-pos-cache
3447 c-state-nonlit-pos-cache-limit
3448 c-state-semi-nonlit-pos-cache
3449 c-state-semi-nonlit-pos-cache-limit
3450 c-state-brace-pair-desert
3451 c-state-point-min
3452 c-state-point-min-lit-type
3453 c-state-point-min-lit-start
3454 c-state-min-scan-pos
3455 c-state-old-cpp-beg
3456 c-state-old-cpp-end
3457 c-parse-state-point))))
3479 conses-not-ok))
3496 ;; The cache can actually go further back due to the ad-hoc way
3497 ;; the first paren is found, so try to whack off a bit of its
3498 ;; start before complaining.
3499 ;; (save-excursion
3500 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3501 ;; (c-beginning-of-defun-1)
3502 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3503 ;; (c-beginning-of-defun-1))
3504 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3505 ;; (message (concat "c-parse-state inconsistency at %s: "
3506 ;; "using cache: %s, from scratch: %s")
3507 ;; here res1 res2)))
3510 here res1 res2)
3516 here res1)
3521 res2 ; res1 correct a cascading series of errors ASAP
3522 ))
3528 'c-debug-parse-state
3536 \f
3538 ;; Whack off any state information from PAREN-STATE which lies
3539 ;; before BUFPOS. Not destructive on PAREN-STATE.
3542 car)
3553 ;; Whack off any state information from PAREN-STATE which lies at or
3554 ;; after BUFPOS. Not destructive on PAREN-STATE.
3559 ;; just check the car, because in a balanced brace
3560 ;; expression, it must be impossible for the corresponding
3561 ;; close brace to be before point, but the open brace to
3562 ;; be after.
3564 nil ; whack it off
3566 ;; its possible that the open brace is before
3567 ;; bufpos, but the close brace is after. In that
3568 ;; case, convert this to a non-cons element. The
3569 ;; rest of the state is before bufpos, so we're
3570 ;; done.
3572 ;; we know that both the open and close braces are
3573 ;; before bufpos, so we also know that everything else
3574 ;; on state is before bufpos.
3577 nil ; whack it off
3578 ;; it's before bufpos, so everything else should too.
3581 nil)))
3584 ;; Return the bufpos of the innermost enclosing open paren before
3585 ;; bufpos, or nil if none was found.
3595 enclosingp))
3598 ;; Return the bufpos of the outermost enclosing open paren, or nil
3599 ;; if none was found.
3606 pos))
3609 ;; Return the closest "safe" position recorded on PAREN-STATE that
3610 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3611 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3612 ;; find the closest limit before a given limit that might be nil.
3613 ;;
3614 ;; A "safe" position is a position at or after a recorded open
3615 ;; paren, or after a recorded close paren. The returned position is
3616 ;; thus either the first position after a close brace, or the first
3617 ;; position after an enclosing paren, or at the enclosing paren in
3618 ;; case BUFPOS is immediately after it.
3628 ;; See below.
3631 ;; elem is the position at and not after the opening paren, so
3632 ;; we can go forward one more step unless it's equal to
3633 ;; bufpos. This is useful in some cases avoid an extra paren
3634 ;; level between the safe position and bufpos.
3639 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3640 ;; goes to the closest previous point that is known to be outside
3641 ;; any string literal or comment. `c-state-cache' is used if it has
3642 ;; a position in the vicinity.
3644 elem
3647 ;; Note: Similar code in `c-safe-position'. The
3648 ;; difference is that we accept a safe position at
3649 ;; the point and don't bother to go forward past open
3650 ;; parens.
3665 ;; The position is far back. Try `c-beginning-of-defun-1'
3666 ;; (although we can't be entirely sure it will go to a position
3667 ;; outside a comment or string in current emacsen). FIXME:
3668 ;; Consult `syntax-ppss' here.
3673 \f
3674 ;; Tools for scanning identifiers and other tokens.
3677 "Return non-nil if the point is on or directly after an identifier.
3678 Keywords are recognized and not considered identifiers. If an
3679 identifier is detected, the returned value is its starting position.
3680 If an identifier ends at the point and another begins at it \(can only
3681 happen in Pike) then the point for the preceding one is returned.
3683 Note that this function might do hidden buffer changes. See the
3684 comment at the start of cc-engine.el for more info."
3686 ;; FIXME: Shouldn't this function handle "operator" in C++?
3693 ;; Check for a normal (non-keyword) identifier.
3699 ;; Handle the `<operator> syntax in Pike.
3703 t
3710 ;; Handle the "operator +" syntax in C++.
3727 )))
3730 ;; If the point is at the end of a symbol then skip backward to the
3731 ;; beginning of it. Don't move otherwise. Return non-nil if point
3732 ;; moved.
3733 ;;
3734 ;; This function might do hidden buffer changes.
3737 ;; Handle the `<operator> syntax in Pike.
3743 t
3745 nil)))))
3748 ;; Move to the beginning of the current token. Do not move if not
3749 ;; in the middle of one. BACK-LIMIT may be used to bound the
3750 ;; backward search; if given it's assumed to be at the boundary
3751 ;; between two tokens. Return non-nil if the point is moved, nil
3752 ;; otherwise.
3753 ;;
3754 ;; This function might do hidden buffer changes.
3761 ;; `c-nonsymbol-token-regexp' should always match
3762 ;; since we've skipped backward over punctuation
3763 ;; or paren syntax, but consume one char in case
3764 ;; it doesn't so that we don't leave point before
3765 ;; some earlier incorrect token.
3772 ;; Move to the end of the current token. Do not move if not in the
3773 ;; middle of one. BACK-LIMIT may be used to bound the backward
3774 ;; search; if given it's assumed to be at the boundary between two
3775 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3776 ;;
3777 ;; This function might do hidden buffer changes.
3785 ;; `c-nonsymbol-token-regexp' should always match since
3786 ;; we've skipped backward over punctuation or paren
3787 ;; syntax, but move forward in case it doesn't so that
3788 ;; we don't leave point earlier than we started with.
3804 "Move forward by tokens.
3805 A token is defined as all symbols and identifiers which aren't
3807 treated properly). Point is always either left at the beginning of a
3808 token or not moved at all. COUNT specifies the number of tokens to
3809 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3810 moves to the next token beginning only if not already at one. If
3811 BALANCED is true, move over balanced parens, otherwise move into them.
3812 Also, if BALANCED is true, never move out of an enclosing paren.
3814 LIMIT sets the limit for the movement and defaults to the point limit.
3815 The case when LIMIT is set in the middle of a token, comment or macro
3816 is handled correctly, i.e. the point won't be left there.
3818 Return the number of tokens left to move \(positive or negative). If
3819 BALANCED is true, a move over a balanced paren counts as one. Note
3820 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3821 be returned. Thus, a return value of 0 guarantees that point is at
3822 the requested position and a return value less \(without signs) than
3823 COUNT guarantees that point is at the beginning of some token.
3825 Note that this function might do hidden buffer changes. See the
3826 comment at the start of cc-engine.el for more info."
3833 c-jump-syntax-balanced
3834 c-jump-syntax-unbalanced))
3839 ;; If count is zero we should jump if in the middle of a token.
3846 ;; Skip whitespace. Count this as a move if we did in
3847 ;; fact move.
3851 ;; Moved out of bounds. Make sure the returned count isn't zero.
3856 ;; Use `condition-case' to avoid having the limit tests
3857 ;; inside the loop.
3865 t)
3868 t)
3869 ;; `c-nonsymbol-token-regexp' above should always
3870 ;; match if there are correct tokens. Try to
3871 ;; widen to see if the limit was set in the
3872 ;; middle of one, else fall back to treating
3873 ;; the offending thing as a one character token.
3878 nil)
3881 t))))
3891 count)))
3894 "Move backward by tokens.
3895 See `c-forward-token-2' for details."
3903 c-jump-syntax-balanced
3904 c-jump-syntax-unbalanced))
3908 ;; The count is zero so try to skip to the beginning of the
3909 ;; current token.
3913 ;; The limit is inside the same token, so return 1.
3916 ;; We're not in the middle of a token. If there's
3917 ;; whitespace after the point then we must move backward,
3918 ;; so set count to 1 in that case.
3920 ;; If we're looking at a '#' that might start a cpp
3921 ;; directive then we have to do a more elaborate check.
3934 ;; Use `condition-case' to avoid having to check for buffer
3935 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3944 ;; This can be very inefficient if there's a long
3945 ;; sequence of operator tokens without any separation.
3946 ;; That doesn't happen in practice, anyway.
3956 count)))
3959 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3961 characters are jumped over character by character. This function is
3962 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3967 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3969 characters are jumped over character by character. This function is
3970 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3974 \f
3975 ;; Tools for doing searches restricted to syntactically relevant text.
3978 paren-level not-inside-token
3979 lookbehind-submatch)
3980 "Like `re-search-forward', but only report matches that are found
3981 in syntactically significant text. I.e. matches in comments, macros
3982 or string literals are ignored. The start point is assumed to be
3983 outside any comment, macro or string literal, or else the content of
3984 that region is taken as syntactically significant text.
3986 If PAREN-LEVEL is non-nil, an additional restriction is added to
3987 ignore matches in nested paren sexps. The search will also not go
3988 outside the current list sexp, which has the effect that if the point
3989 should be moved to BOUND when no match is found \(i.e. NOERROR is
3990 neither nil nor t), then it will be at the closing paren if the end of
3991 the current list sexp is encountered first.
3993 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3994 ignored. Things like multicharacter operators and special symbols
3996 constants.
3998 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3999 subexpression in REGEXP. The end of that submatch is used as the
4000 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
4001 isn't used or if that subexpression didn't match then the start
4003 subexpression is never tested before the starting position, so it
4007 subexpression can match the end of nonwhite syntactic whitespace,
4008 i.e. the end of comments or cpp directives. This since the function
4009 skips over such things before resuming the search. It's on the other
4011 matches syntactic whitespace.
4013 Bug: Unbalanced parens inside cpp directives are currently not handled
4014 correctly \(i.e. they don't get ignored as they should) when
4015 PAREN-LEVEL is set.
4017 Note that this function might do hidden buffer changes. See the
4018 comment at the start of cc-engine.el for more info."
4023 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
4026 tmp
4027 ;; Start position for the last search.
4028 search-pos
4029 ;; The `parse-partial-sexp' state between the start position
4030 ;; and the point.
4031 state
4032 ;; The current position after the last state update. The next
4033 ;; `parse-partial-sexp' continues from here.
4035 ;; The position at which to check the state and the state
4036 ;; there. This is separate from `state-pos' since we might
4037 ;; need to back up before doing the next search round.
4038 check-pos check-state
4039 ;; Last position known to end a token.
4041 ;; Set when a valid match is found.
4042 found)
4060 state-pos check-pos paren-level nil state))
4062 check-state state))
4064 ;; NOTE: If we got a look behind subexpression and get
4065 ;; an insignificant match in something that isn't
4066 ;; syntactic whitespace (i.e. strings or in nested
4067 ;; parentheses), then we can never skip more than a
4068 ;; single character from the match start position
4069 ;; (i.e. `state-pos' here) before continuing the
4070 ;; search. That since the look behind subexpression
4071 ;; might match the end of the insignificant region in
4072 ;; the next search.
4076 ;; Match inside a line comment. Skip to eol. Use
4077 ;; `re-search-forward' instead of `skip-chars-forward' to get
4078 ;; the right bound behavior.
4082 ;; Match inside a block comment. Skip to the '*/'.
4089 ;; Match in the middle of the opener of a block or line
4090 ;; comment.
4095 ;; The last `parse-partial-sexp' above might have
4096 ;; stopped short of the real check position if the end
4097 ;; of the current sexp was encountered in paren-level
4098 ;; mode. The checks above are always false in that
4099 ;; case, and since they can do better skipping in
4100 ;; lookbehind-submatch mode, we do them before
4101 ;; checking the paren level.
4105 ;; Check the paren level first since we're short of the
4106 ;; syntactic checking position if the end of the
4107 ;; current sexp was encountered by `parse-partial-sexp'.
4110 ;; Inside a nested paren sexp.
4112 ;; See the NOTE above.
4114 ;; Skip out of the paren quickly.
4118 ;; Have exited the current paren sexp.
4121 ;; The last `parse-partial-sexp' call above
4122 ;; has left us just after the closing paren
4123 ;; in this case, so we can modify the bound
4124 ;; to leave the point at the right position
4125 ;; upon return.
4127 nil)
4131 ;; Match inside a string.
4134 ;; See the NOTE above.
4136 ;; Skip to the end of the string before continuing.
4145 continue)))
4150 ;; Match inside a macro. Skip to the end of it.
4158 (< check-pos
4164 ;; Inside a token.
4166 ;; See the NOTE above.
4171 ;; A real match.
4173 nil)))
4175 ;; Should loop to search again, but take care to avoid
4176 ;; looping on the same spot.
4180 nil
4183 t))))
4189 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4196 ;; Search failed. Set point as appropriate.
4200 nil)))
4205 ;; Return the start of the literal point is in, or nil.
4206 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4207 ;; bound in the caller.
4209 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4210 ;; if it's outside comments and strings.
4213 ;; Pick a safe position as close to the point as possible.
4214 ;;
4215 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4216 ;; position.
4229 ;; Cache positions along the way to use if we have to back up more. We
4230 ;; cache every closing paren on the same level. If the paren cache is
4231 ;; relevant in this region then we're typically already on the same
4232 ;; level as the target position. Note that we might cache positions
4233 ;; after opening parens in case safe-pos is in a nested list. That's
4234 ;; both uncommon and harmless.
4242 ;; If the state contains the start of the containing sexp we cache that
4243 ;; position too, so that parse-partial-sexp in the next run has a bigger
4244 ;; chance of starting at the same level as the target position and thus
4245 ;; will get more good safe positions into the list.
4251 ;; Inside string or comment. Continue search at the
4252 ;; beginning of it.
4256 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4257 i.e. don't stop at positions inside syntactic whitespace or string
4258 literals. Preprocessor directives are also ignored, with the exception
4259 of the one that the point starts within, if any. If LIMIT is given,
4260 it's assumed to be at a syntactically relevant position.
4262 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4263 sexps, and the search will also not go outside the current paren sexp.
4264 However, if LIMIT or the buffer limit is reached inside a nested paren
4265 then the point will be left at the limit.
4267 Non-nil is returned if the point moved, nil otherwise.
4269 Note that this function might do hidden buffer changes. See the
4270 comment at the start of cc-engine.el for more info."
4273 state-2
4274 ;; A list of syntactically relevant positions in descending
4275 ;; order. It's used to avoid scanning repeatedly over
4276 ;; potentially large regions with `parse-partial-sexp' to verify
4277 ;; each position. Used in `c-ssb-lit-begin'
4278 safe-pos-list
4279 ;; The result from `c-beginning-of-macro' at the start position or the
4280 ;; start position itself if it isn't within a macro. Evaluated on
4281 ;; demand.
4282 start-macro-beg
4283 ;; The earliest position after the current one with the same paren
4284 ;; level. Used only when `paren-level' is set.
4285 lit-beg
4290 ;; The next loop "tries" to find the end point each time round,
4291 ;; loops when it hasn't succeeded.
4297 ;; Don't stop inside a literal.
4300 t)))
4312 ;; Not at the right level.
4315 ;; We stop above if we go out of a paren.
4316 ;; Now check whether it precedes or is
4317 ;; nested in the starting sexp.
4321 pps-end-pos paren-level-pos
4322 nil nil state-2))
4325 ;; We've stopped short of the starting position
4326 ;; so the hit was inside a nested list. Go up
4327 ;; until we are at the right level.
4336 nil)
4337 t))
4340 nil))
4342 ;; The hit was outside the list at the start
4343 ;; position. Go to the start of the list and exit.
4345 nil))
4348 ;; Inside a macro.
4356 t
4358 ;; It's inside the same macro we started in so it's
4359 ;; a relevant match.
4361 nil))))))
4365 ;; Skip syntactic ws afterwards so that we don't stop at the
4366 ;; end of a comment if `skip-chars' is something like "^/".
4370 ;; We might want to extend this with more useful return values in
4371 ;; the future.
4374 ;; The following is an alternative implementation of
4375 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4376 ;; track of the syntactic context. It turned out to be generally
4377 ;; slower than the one above which uses forward checks from earlier
4378 ;; safe positions.
4379 ;;
4380 ;;(defconst c-ssb-stop-re
4381 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4382 ;; ;; stop at to avoid going into comments and literals.
4383 ;; (concat
4384 ;; ;; Match comment end syntax and string literal syntax. Also match
4385 ;; ;; '/' for block comment endings (not covered by comment end
4386 ;; ;; syntax).
4387 ;; "\\s>\\|/\\|\\s\""
4388 ;; (if (memq 'gen-string-delim c-emacs-features)
4389 ;; "\\|\\s|"
4390 ;; "")
4391 ;; (if (memq 'gen-comment-delim c-emacs-features)
4392 ;; "\\|\\s!"
4393 ;; "")))
4394 ;;
4395 ;;(defconst c-ssb-stop-paren-re
4396 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4397 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4398 ;;
4399 ;;(defconst c-ssb-sexp-end-re
4400 ;; ;; Regexp matching the ending syntax of a complex sexp.
4401 ;; (concat c-string-limit-regexp "\\|\\s)"))
4402 ;;
4403 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4404 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4405 ;;i.e. don't stop at positions inside syntactic whitespace or string
4406 ;;literals. Preprocessor directives are also ignored. However, if the
4407 ;;point is within a comment, string literal or preprocessor directory to
4408 ;;begin with, its contents is treated as syntactically relevant chars.
4409 ;;If LIMIT is given, it limits the backward search and the point will be
4410 ;;left there if no earlier position is found.
4411 ;;
4412 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4413 ;;sexps, and the search will also not go outside the current paren sexp.
4414 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4415 ;;then the point will be left at the limit.
4416 ;;
4417 ;;Non-nil is returned if the point moved, nil otherwise.
4418 ;;
4419 ;;Note that this function might do hidden buffer changes. See the
4420 ;;comment at the start of cc-engine.el for more info."
4421 ;;
4422 ;; (save-restriction
4423 ;; (when limit
4424 ;; (narrow-to-region limit (point-max)))
4425 ;;
4426 ;; (let ((start (point)))
4427 ;; (catch 'done
4428 ;; (while (let ((last-pos (point))
4429 ;; (stop-pos (progn
4430 ;; (skip-chars-backward skip-chars)
4431 ;; (point))))
4432 ;;
4433 ;; ;; Skip back over the same region as
4434 ;; ;; `skip-chars-backward' above, but keep to
4435 ;; ;; syntactically relevant positions.
4436 ;; (goto-char last-pos)
4437 ;; (while (and
4438 ;; ;; `re-search-backward' with a single char regexp
4439 ;; ;; should be fast.
4440 ;; (re-search-backward
4441 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4442 ;; stop-pos 'move)
4443 ;;
4444 ;; (progn
4445 ;; (cond
4446 ;; ((looking-at "\\s(")
4447 ;; ;; `paren-level' is set and we've found the
4448 ;; ;; start of the containing paren.
4449 ;; (forward-char)
4450 ;; (throw 'done t))
4451 ;;
4452 ;; ((looking-at c-ssb-sexp-end-re)
4453 ;; ;; We're at the end of a string literal or paren
4454 ;; ;; sexp (if `paren-level' is set).
4455 ;; (forward-char)
4456 ;; (condition-case nil
4457 ;; (c-backward-sexp)
4458 ;; (error
4459 ;; (goto-char limit)
4460 ;; (throw 'done t))))
4461 ;;
4462 ;; (t
4463 ;; (forward-char)
4464 ;; ;; At the end of some syntactic ws or possibly
4465 ;; ;; after a plain '/' operator.
4466 ;; (let ((pos (point)))
4467 ;; (c-backward-syntactic-ws)
4468 ;; (if (= pos (point))
4469 ;; ;; Was a plain '/' operator. Go past it.
4470 ;; (backward-char)))))
4471 ;;
4472 ;; (> (point) stop-pos))))
4473 ;;
4474 ;; ;; Now the point is either at `stop-pos' or at some
4475 ;; ;; position further back if `stop-pos' was at a
4476 ;; ;; syntactically irrelevant place.
4477 ;;
4478 ;; ;; Skip additional syntactic ws so that we don't stop
4479 ;; ;; at the end of a comment if `skip-chars' is
4480 ;; ;; something like "^/".
4481 ;; (c-backward-syntactic-ws)
4482 ;;
4483 ;; (< (point) stop-pos))))
4484 ;;
4485 ;; ;; We might want to extend this with more useful return values
4486 ;; ;; in the future.
4487 ;; (/= (point) start))))
4489 \f
4490 ;; Tools for handling comments and string literals.
4493 "Return the type of literal point is in, if any.
4494 The return value is `c' if in a C-style comment, `c++' if in a C++
4495 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4496 is non-nil and in a preprocessor line, or nil if somewhere else.
4497 Optional LIM is used as the backward limit of the search. If omitted,
4498 or nil, `c-beginning-of-defun' is used.
4500 The last point calculated is cached if the cache is enabled, i.e. if
4501 `c-in-literal-cache' is bound to a two element vector.
4503 Note that this function might do hidden buffer changes. See the
4504 comment at the start of cc-engine.el for more info."
4515 "Return a cons of the beginning and end positions of the comment or
4516 string surrounding point (including both delimiters), or nil if point
4518 to start parsing from. If NEAR is non-nil, then the limits of any
4520 spaces and tabs between point and the literal. The search for such a
4521 literal is done first in forward direction. If NOT-IN-DELIMITER is
4522 non-nil, the case when point is inside a starting delimiter won't be
4523 recognized. This only has effect for comments which have starting
4524 delimiters with more than one character.
4526 Note that this function might do hidden buffer changes. See the
4527 comment at the start of cc-engine.el for more info."
4543 ;; Search forward for a literal.
4554 ;; Search backward.
4565 ;; Block comment. Due to the nature of line
4566 ;; comments, they will always be covered by the
4567 ;; normal case above.
4570 ;; If LIM is bogus, beg will be bogus.
4574 ))))
4576 ;; In case external callers use this; it did have a docstring.
4580 "If the argument is a cons of two buffer positions (such as returned by
4581 `c-literal-limits'), and that range contains a C++ style line comment,
4582 then an extended range is returned that contains all adjacent line
4583 comments (i.e. all comments that starts in the same column with no
4584 empty lines or non-whitespace characters between them). Otherwise the
4585 argument is returned.
4587 Note that this function might do hidden buffer changes. See the
4588 comment at the start of cc-engine.el for more info."
4599 ;; Got to take care in the backward direction to handle
4600 ;; comments which are preceded by code.
4614 range)
4618 "Convenience function that given the result of `c-literal-limits',
4619 returns nil or the type of literal that the range surrounds, one
4620 of the symbols `c', `c++' or `string'. It's much faster than using
4621 `c-in-literal' and is intended to be used when you need both the
4622 type of a literal and its limits.
4624 Note that this function might do hidden buffer changes. See the
4625 comment at the start of cc-engine.el for more info."
4636 range))
4639 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4640 ;; This doesn't preserve point.
4649 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4650 ;; (default point). This is done by going back further in the buffer then
4651 ;; searching forward for literals. The position found won't be in a
4652 ;; literal. We start searching for the sought position TRY-SIZE (default
4653 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4654 ;; :-)
4662 stack elt size
4665 ;; Move forward one literal each time round this loop.
4666 ;; Move forward to the start of a comment or string.
4668 pos
4669 start
4670 nil ; target-depth
4671 nil ; stop-before
4672 s ; state
4675 ;; Gather details of the non-literal-bit - starting pos and size.
4679 pos))
4683 ;; Move forward to the end of the comment/string.
4687 start
4688 nil ; target-depth
4689 nil ; stop-before
4690 s ; state
4694 ;; Now try and find enough non-literal characters recorded on the stack.
4695 ;; Go back one recorded literal each time round this loop.
4697 stack)
4702 ;; Have we found enough yet?
4712 ;; Return a buffer position about HOW-FAR non-literal characters forward
4713 ;; from START-POS (default point), which must not be inside a literal.
4720 ;; Scan over counted characters.
4722 pos
4724 nil ; target-depth
4725 nil ; stop-before
4726 s ; state
4730 ;; Scan over literal characters.
4733 pos
4735 nil ; target-depth
4736 nil ; stop-before
4737 s ; state
4742 \f
4743 ;; `c-find-decl-spots' and accompanying stuff.
4745 ;; Variables used in `c-find-decl-spots' to cache the search done for
4746 ;; the first declaration in the last call. When that function starts,
4747 ;; it needs to back up over syntactic whitespace to look at the last
4748 ;; token before the region being searched. That can sometimes cause
4749 ;; moves back and forth over a quite large region of comments and
4750 ;; macros, which would be repeated for each changed character when
4751 ;; we're called during fontification, since font-lock refontifies the
4752 ;; current line for each change. Thus it's worthwhile to cache the
4753 ;; first match.
4754 ;;
4755 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4756 ;; the syntactic whitespace less or equal to some start position.
4757 ;; There's no cached value if it's nil.
4758 ;;
4759 ;; `c-find-decl-match-pos' is the match position if
4760 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4761 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4772 ; (defface c-debug-decl-spot-face
4773 ; '((t (:background "Turquoise")))
4774 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4775 ; (defface c-debug-decl-sws-face
4776 ; '((t (:background "Khaki")))
4777 ; "Debug face to mark the syntactic whitespace between the declaration
4778 ; spots and the preceding token end.")
4794 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4795 ;; but it contains lots of free variables that refer to things
4796 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4797 ;; if there is a match, otherwise at `cfd-limit'.
4798 ;;
4799 ;; The macro moves point forward to the next putative start of a declaration
4800 ;; or cfd-limit. This decl start is the next token after a "declaration
4801 ;; prefix". The declaration prefix is the earlier of `cfd-prop-match' and
4802 ;; `cfd-re-match'. `cfd-match-pos' is set to the decl prefix.
4803 ;;
4804 ;; This macro might do hidden buffer changes.
4807 ;; Find the next property match position if we haven't got one already.
4818 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4819 ;; got one already.
4825 ;; Each time round, the next `while' moves forward over a pseudo match
4826 ;; of `c-decl-prefix-or-start-re' which is either inside a literal, or
4827 ;; is a ":" not preceded by "public", etc.. `cfd-re-match' and
4828 ;; `cfd-re-match-end' get set.
4835 ;; No match. Finish up and exit the loop.
4837 nil)
4840 ;; Matched the end of a token preceding a decl spot.
4844 ;; Matched a token that start a decl spot.
4847 c-literal-faces)
4848 ;; Pseudo match inside a comment or string literal. Skip out
4849 ;; of comments and string literals.
4861 ;; Found a ":" which isn't part of "public:", etc.
4862 t)
4865 ;; If our match was at the decl start, we have to back up over the
4866 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4867 ;; any decl spots in the syntactic ws.
4872 ;; Choose whichever match is closer to the start.
4875 cfd-re-match nil)
4877 cfd-prop-match nil))
4882 ;; Skip forward past comments only so we don't skip macros.
4884 ;; Set the position to continue at. We can avoid going over
4885 ;; the comments skipped above a second time, but it's possible
4886 ;; that the comment skipping has taken us past `cfd-prop-match'
4887 ;; since the property might be used inside comments.
4893 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4894 ;; label from the point to CFD-LIMIT.
4895 ;;
4896 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4897 ;; arguments: The first is the end position of the token preceding the spot,
4898 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4899 ;; the match is inside a macro. Point should be moved forward by at least
4900 ;; one token.
4901 ;;
4902 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4903 ;; it should return non-nil to ensure that the next search will find them.
4904 ;;
4905 ;; Such a spot is:
4906 ;; o The first token after bob.
4907 ;; o The first token after the end of submatch 1 in
4908 ;; `c-decl-prefix-or-start-re' when that submatch matches. This
4909 ;; submatch is typically a (L or R) brace or paren, a ;, or a ,.
4910 ;; o The start of each `c-decl-prefix-or-start-re' match when
4911 ;; submatch 1 doesn't match. This is, for example, the keyword
4912 ;; "class" in Pike.
4913 ;; o The start of a previously recognized declaration; "recognized"
4914 ;; means that the last char of the previous token has a `c-type'
4915 ;; text property with the value `c-decl-end'; this only holds
4916 ;; when `c-type-decl-end-used' is set.
4917 ;;
4918 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4919 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4920 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4921 ;;
4922 ;; If the match is inside a macro then the buffer is narrowed to the
4923 ;; end of it, so that CFD-FUN can investigate the following tokens
4924 ;; without matching something that begins inside a macro and ends
4925 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4926 ;; CFD-FACE-CHECKLIST checks exist.
4927 ;;
4928 ;; The spots are visited approximately in order from top to bottom.
4929 ;; It's however the positions where `c-decl-prefix-or-start-re'
4930 ;; matches and where `c-decl-end' properties are found that are in
4931 ;; order. Since the spots often are at the following token, they
4932 ;; might be visited out of order insofar as more spots are reported
4933 ;; later on within the syntactic whitespace between the match
4934 ;; positions and their spots.
4935 ;;
4936 ;; It's assumed that comments and strings are fontified in the
4937 ;; searched range.
4938 ;;
4939 ;; This is mainly used in fontification, and so has an elaborate
4940 ;; cache to handle repeated calls from the same start position; see
4941 ;; the variables above.
4942 ;;
4943 ;; All variables in this function begin with `cfd-' to avoid name
4944 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4945 ;;
4946 ;; This function might do hidden buffer changes.
4950 ;; The end of the token preceding the decl spot last found
4951 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4952 ;; no match.
4953 cfd-re-match
4954 ;; The end position of the last `c-decl-prefix-or-start-re'
4955 ;; match. If this is greater than `cfd-continue-pos', the
4956 ;; next regexp search is started here instead.
4958 ;; The end of the last `c-decl-end' found by
4959 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4960 ;; match. If searching for the property isn't needed then we
4961 ;; disable it by setting it to `cfd-limit' directly.
4963 ;; The end of the token preceding the decl spot last found by
4964 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4965 ;; bob. `cfd-limit' if there's no match. In other words,
4966 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4968 ;; The position to continue searching at.
4969 cfd-continue-pos
4970 ;; The position of the last "real" token we've stopped at.
4971 ;; This can be greater than `cfd-continue-pos' when we get
4972 ;; hits inside macros or at `c-decl-end' positions inside
4973 ;; comments.
4975 ;; The end position of the last entered macro.
4978 ;; Initialize by finding a syntactically relevant start position
4979 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4980 ;; search unless we're at bob.
4983 ;; Must back up a bit since we look for the end of the previous
4984 ;; statement or declaration, which is earlier than the first
4985 ;; returned match.
4987 ;; This `cond' moves back over any literals or macros. It has special
4988 ;; handling for when the region being searched is entirely within a
4989 ;; macro. It sets `cfd-continue-pos' (unless we've reached
4990 ;; `cfd-limit').
4992 ;; First we need to move to a syntactically relevant position.
4993 ;; Begin by backing out of comment or string literals.
4994 ;;
4995 ;; This arm of the cond actually triggers if we're in a literal,
4996 ;; and cfd-limit is at most at BONL.
4998 ;; This arm of the `and' moves backwards out of a literal when
4999 ;; the face at point is a literal face. In this case, its value
5000 ;; is always non-nil.
5002 ;; Try to use the faces to back up to the start of the
5003 ;; literal. FIXME: What if the point is on a declaration
5004 ;; inside a comment?
5010 ;; XEmacs doesn't fontify the quotes surrounding string
5011 ;; literals.
5020 ;; Don't trust the literal to contain only literal faces
5021 ;; (the font lock package might not have fontified the
5022 ;; start of it at all, for instance) so check that we have
5023 ;; arrived at something that looks like a start or else
5024 ;; resort to `c-literal-limits'.
5031 ;; The start is in a literal. If the limit is in the same
5032 ;; one we don't have to find a syntactic position etc. We
5033 ;; only check that if the limit is at or before bonl to save
5034 ;; time; it covers the by far most common case when font-lock
5035 ;; refontifies the current line only.
5046 ;; Completely inside a literal. Set up variables to trig the
5047 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
5048 ;; find a suitable start position.
5051 ;; Check if the region might be completely inside a macro, to
5052 ;; optimize that like the completely-inside-literal above.
5059 ;; (Maybe) completely inside a macro. Only need to trig the
5060 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
5061 ;; set things up.
5063 start-in-macro t))
5065 ;; The default arm of the `cond' moves back over any macro we're in
5066 ;; and over any syntactic WS. It sets `c-find-decl-syntactic-pos'.
5068 ;; Back out of any macro so we don't miss any declaration
5069 ;; that could follow after it.
5073 ;; Now we're at a proper syntactically relevant position so we
5074 ;; can use the cache. But first clear it if it applied
5075 ;; further down.
5080 ;; Don't have to do this if the cache is relevant here,
5081 ;; typically if the same line is refontified again. If
5082 ;; we're just some syntactic whitespace further down we can
5083 ;; still use the cache to limit the skipping.
5086 ;; If we hit `c-find-decl-syntactic-pos' and
5087 ;; `c-find-decl-match-pos' is set then we install the cached
5088 ;; values. If we hit `c-find-decl-syntactic-pos' and
5089 ;; `c-find-decl-match-pos' is nil then we know there's no decl
5090 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
5091 ;; and so we can continue the search from this point. If we
5092 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
5093 ;; the right spot to begin searching anyway.
5095 c-find-decl-match-pos)
5097 cfd-continue-pos syntactic-pos)
5102 ;; Always consider bob a match to get the first
5103 ;; declaration in the file. Do this separately instead of
5104 ;; letting `c-decl-prefix-or-start-re' match bob, so that
5105 ;; regexp always can consume at least one character to
5106 ;; ensure that we won't get stuck in an infinite loop.
5111 ;; Do an initial search now. In the bob case above it's
5112 ;; only done to search for a `c-decl-end' spot.
5118 ;; Advance `cfd-continue-pos' if it's before the start position.
5119 ;; The closest continue position that might have effect at or
5120 ;; after the start depends on what we started in. This also
5121 ;; finds a suitable start position in the special cases when the
5122 ;; region is completely within a literal or macro.
5127 ;; If we're in a macro then it's the closest preceding token
5128 ;; in the macro. Check this before `start-in-literal',
5129 ;; since if we're inside a literal in a macro, the preceding
5130 ;; token is earlier than any `c-decl-end' spot inside the
5131 ;; literal (comment).
5133 ;; The only syntactic ws in macros are comments.
5139 ;; If we're in a comment it can only be the closest
5140 ;; preceding `c-decl-end' position within that comment, if
5141 ;; any. Go back to the beginning of such a property so that
5142 ;; `c-find-decl-prefix-search' will find the end of it.
5143 ;; (Can't stop at the end and install it directly on
5144 ;; `cfd-prop-match' since that variable might be cleared
5145 ;; after `cfd-fun' below.)
5146 ;;
5147 ;; Note that if the literal is a string then the property
5148 ;; search will simply skip to the beginning of it right
5149 ;; away.
5161 ;; Didn't find any property inside the comment, so we can
5162 ;; skip it entirely. (This won't skip past a string, but
5163 ;; that'll be handled quickly by the next
5164 ;; `c-find-decl-prefix-search' anyway.)
5170 ;; If we started in normal code, the only match that might
5171 ;; apply before the start is what we already got in
5172 ;; `cfd-match-pos' so we can continue at the start position.
5173 ;; (Note that we don't get here if the first match is below
5174 ;; it.)
5177 ;; Delete found matches if they are before our new continue
5178 ;; position, so that `c-find-decl-prefix-search' won't back up
5179 ;; to them later on.
5187 ;; This is the normal case and we got a proper syntactic
5188 ;; position. If there's a match then it's always outside
5189 ;; macros and comments, so advance to the next token and set
5190 ;; `cfd-token-pos'. The loop below will later go back using
5191 ;; `cfd-continue-pos' to fix declarations inside the
5192 ;; syntactic ws.
5200 ;; Have one of the special cases when the region is completely
5201 ;; within a literal or macro. `cfd-continue-pos' is set to a
5202 ;; good start position for the search, so do it.
5205 ;; Now loop, one decl spot per iteration. We already have the first
5206 ;; match in `cfd-match-pos'.
5208 ;; Go forward over "false matches", one per iteration.
5213 ;; Kludge to filter out matches on the "<" that
5214 ;; aren't open parens, for the sake of languages
5215 ;; that got `c-recognize-<>-arglists' set.
5220 ;; If `cfd-continue-pos' is less or equal to
5221 ;; `cfd-token-pos', we've got a hit inside a macro
5222 ;; that's in the syntactic whitespace before the last
5223 ;; "real" declaration we've checked. If they're equal
5224 ;; we've arrived at the declaration a second time, so
5225 ;; there's nothing to do.
5229 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5230 ;; we're still searching for declarations embedded in
5231 ;; the syntactic whitespace. In that case we need
5232 ;; only to skip comments and not macros, since they
5233 ;; can't be nested, and that's already been done in
5234 ;; `c-find-decl-prefix-search'.
5239 ;; Continue if the following token fails the
5240 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5247 t)))
5258 ;; Narrow to the end of the macro if we got a hit inside
5259 ;; one, to avoid recognizing things that start inside the
5260 ;; macro and end outside it.
5262 ;; Not in the same macro as in the previous round.
5273 t
5276 t)
5277 ;; The matched token was the last thing in the macro,
5278 ;; so the whole match is bogus.
5287 ;; Restore limits if we did macro narrowing above.
5294 ; cfd-match-pos, etc.
5296 \f
5297 ;; A cache for found types.
5299 ;; Buffer local variable that contains an obarray with the types we've
5300 ;; found. If a declaration is recognized somewhere we record the
5301 ;; fully qualified identifier in it to recognize it as a type
5302 ;; elsewhere in the file too. This is not accurate since we do not
5303 ;; bother with the scoping rules of the languages, but in practice the
5304 ;; same name is seldom used as both a type and something else in a
5305 ;; file, and we only use this as a last resort in ambiguous cases (see
5306 ;; `c-forward-decl-or-cast-1').
5307 ;;
5308 ;; Not every type need be in this cache. However, things which have
5309 ;; ceased to be types must be removed from it.
5310 ;;
5311 ;; Template types in C++ are added here too but with the template
5312 ;; arglist replaced with "<>" in references or "<" for the one in the
5313 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5314 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5315 ;; template specs can be fairly sized programs in themselves) and
5316 ;; improves the hit ratio (it's a type regardless of the template
5317 ;; args; it's just not the same type, but we're only interested in
5318 ;; recognizing types, not telling distinct types apart). Note that
5319 ;; template types in references are added here too; from the example
5320 ;; above there will also be an entry "Foo<".
5325 ;; Clears `c-found-types'.
5329 ;; Add the given region as a type in `c-found-types'. If the region
5330 ;; doesn't match an existing type but there is a type which is equal
5331 ;; to the given one except that the last character is missing, then
5332 ;; the shorter type is removed. That's done to avoid adding all
5333 ;; prefixes of a type as it's being entered and font locked. This
5334 ;; doesn't cover cases like when characters are removed from a type
5335 ;; or added in the middle. We'd need the position of point when the
5336 ;; font locking is invoked to solve this well.
5337 ;;
5338 ;; This function might do hidden buffer changes.
5345 ;; Remove the "NAME" from c-found-types, if present.
5349 ;; Return non-nil if the given region contains a type in
5350 ;; `c-found-types'.
5351 ;;
5352 ;; This function might do hidden buffer changes.
5354 c-found-types))
5357 ;; Return all the types in `c-found-types' as a sorted list of
5358 ;; strings.
5362 type-list)))
5363 c-found-types)
5366 ;; Shut up the byte compiler.
5370 ;; An after change function which, in conjunction with the info in
5371 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5372 ;; from `c-found-types', should this type have become stale. For
5373 ;; example, this happens to "foo" when "foo \n bar();" becomes
5374 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5375 ;; the fontification.
5376 ;;
5377 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5378 ;; type?
5398 ;; Changing the amount of (already existing) whitespace - don't do anything.
5403 ;; The syntactic relationship which defined a "found type" has been
5404 ;; destroyed.
5407 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5408 )))
5410 \f
5411 ;; Setting and removing syntax properties on < and > in languages (C++
5412 ;; and Java) where they can be template/generic delimiters as well as
5413 ;; their normal meaning of "less/greater than".
5415 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5416 ;; be delimiters, they are marked as such with the category properties
5417 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5419 ;; STRATEGY:
5420 ;;
5421 ;; It is impossible to determine with certainty whether a <..> pair in
5422 ;; C++ is two comparison operators or is template delimiters, unless
5423 ;; one duplicates a lot of a C++ compiler. For example, the following
5424 ;; code fragment:
5425 ;;
5426 ;; foo (a < b, c > d) ;
5427 ;;
5428 ;; could be a function call with two integer parameters (each a
5429 ;; relational expression), or it could be a constructor for class foo
5430 ;; taking one parameter d of templated type "a < b, c >". They are
5431 ;; somewhat easier to distinguish in Java.
5432 ;;
5433 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5434 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5435 ;; individually when their context so indicated. This gave rise to
5436 ;; intractable problems when one of a matching pair was deleted, or
5437 ;; pulled into a literal.]
5438 ;;
5439 ;; At each buffer change, the syntax-table properties are removed in a
5440 ;; before-change function and reapplied, when needed, in an
5441 ;; after-change function. It is far more important that the
5442 ;; properties get removed when they they are spurious than that they
5443 ;; be present when wanted.
5444 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5446 ;; POS (default point) is at a < character. If it is marked with
5447 ;; open paren syntax-table text property, remove the property,
5448 ;; together with the close paren property on the matching > (if
5449 ;; any).
5464 ;; POS (default point) is at a > character. If it is marked with
5465 ;; close paren syntax-table property, remove the property, together
5466 ;; with the open paren property on the matching < (if any).
5481 ;; POS (default point) is at a < or > character. If it has an
5482 ;; open/close paren syntax-table property, remove this property both
5483 ;; from the current character and its partner (which will also be
5484 ;; thusly marked).
5494 ;; POS (default point) is at a < character. If it is both marked
5495 ;; with open/close paren syntax-table property, and has a matching >
5496 ;; (also marked) which is after LIM, remove the property both from
5497 ;; the current > and its partner. Return t when this happens, nil
5498 ;; when it doesn't.
5512 t)))
5515 ;; POS (default point) is at a > character. If it is both marked
5516 ;; with open/close paren syntax-table property, and has a matching <
5517 ;; (also marked) which is before LIM, remove the property both from
5518 ;; the current < and its partner. Return t when this happens, nil
5519 ;; when it doesn't.
5533 t)))
5535 ;; Set by c-common-init in cc-mode.el.
5540 ;; Unmark certain pairs of "< .... >" which are currently marked as
5541 ;; template/generic delimiters. (This marking is via syntax-table
5542 ;; text properties).
5543 ;;
5544 ;; These pairs are those which are in the current "statement" (i.e.,
5545 ;; the region between the {, }, or ; before BEG and the one after
5546 ;; END), and which enclose any part of the interval (BEG END).
5547 ;;
5548 ;; Note that in C++ (?and Java), template/generic parens cannot
5549 ;; enclose a brace or semicolon, so we use these as bounds on the
5550 ;; region we must work on.
5551 ;;
5552 ;; This function is called from before-change-functions (via
5553 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5554 ;; and point is undefined, both at entry and exit.
5555 ;;
5556 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5557 ;; 2010-01-29.
5561 new-beg new-end need-new-beg need-new-end)
5562 ;; Locate the barrier before the changed region
5567 ;; Remove the syntax-table/category properties from each pertinent <...>
5568 ;; pair. Firsly, the ones with the < before beg and > after beg.
5574 ;; Locate the barrier after END.
5579 ;; Remove syntax-table properties from the remaining pertinent <...>
5580 ;; pairs, those with a > after end and < before end.
5585 ;; Extend the fontification region, if needed.
5595 ;; This is called from `after-change-functions' when
5596 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5597 ;; chars with paren syntax become part of another operator like "<<"
5598 ;; or ">=".
5599 ;;
5600 ;; This function might do hidden buffer changes.
5633 \f
5634 ;; Handling of small scale constructs like types and names.
5636 ;; Dynamically bound variable that instructs `c-forward-type' to also
5637 ;; treat possible types (i.e. those that it normally returns 'maybe or
5638 ;; 'found for) as actual types (and always return 'found for them).
5639 ;; This means that it records them in `c-record-type-identifiers' if
5640 ;; that is set, and that it adds them to `c-found-types'.
5643 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5644 ;; mark up successfully parsed arglists with paren syntax properties on
5645 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5646 ;; `c-type' property of each argument separating comma.
5647 ;;
5648 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5649 ;; all arglists for side effects (i.e. recording types), otherwise it
5650 ;; exploits any existing paren syntax properties to quickly jump to the
5651 ;; end of already parsed arglists.
5652 ;;
5653 ;; Marking up the arglists is not the default since doing that correctly
5654 ;; depends on a proper value for `c-restricted-<>-arglists'.
5657 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5658 ;; not accept arglists that contain binary operators.
5659 ;;
5660 ;; This is primarily used to handle C++ template arglists. C++
5661 ;; disambiguates them by checking whether the preceding name is a
5662 ;; template or not. We can't do that, so we assume it is a template
5663 ;; if it can be parsed as one. That usually works well since
5664 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5665 ;; in almost all cases would be pointless.
5666 ;;
5667 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5668 ;; should let the comma separate the function arguments instead. And
5669 ;; in a context where the value of the expression is taken, e.g. in
5670 ;; "if (a < b || c > d)", it's probably not a template.
5673 ;; Dynamically bound variables that instructs
5674 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5675 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5676 ;; `c-forward-label' to record the ranges of all the type and
5677 ;; reference identifiers they encounter. They will build lists on
5678 ;; these variables where each element is a cons of the buffer
5679 ;; positions surrounding each identifier. This recording is only
5680 ;; activated when `c-record-type-identifiers' is non-nil.
5681 ;;
5682 ;; All known types that can't be identifiers are recorded, and also
5683 ;; other possible types if `c-promote-possible-types' is set.
5684 ;; Recording is however disabled inside angle bracket arglists that
5685 ;; are encountered inside names and other angle bracket arglists.
5686 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5687 ;; instead.
5688 ;;
5689 ;; Only the names in C++ template style references (e.g. "tmpl" in
5690 ;; "tmpl<a,b>::foo") are recorded as references, other references
5691 ;; aren't handled here.
5692 ;;
5693 ;; `c-forward-label' records the label identifier(s) on
5694 ;; `c-record-ref-identifiers'.
5698 ;; This variable will receive a cons cell of the range of the last
5699 ;; single identifier symbol stepped over by `c-forward-name' if it's
5700 ;; successful. This is the range that should be put on one of the
5701 ;; record lists above by the caller. It's assigned nil if there's no
5702 ;; such symbol in the name.
5707 ;; Always true.
5717 ;; Always true.
5725 ;; Dynamically bound variable that instructs `c-forward-type' to
5726 ;; record the ranges of types that only are found. Behaves otherwise
5727 ;; like `c-record-type-identifiers'.
5731 ;; Used internally in `c-forward-keyword-clause' to move forward
5732 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5733 ;; possibly is prefixed by keywords and their associated clauses.
5734 ;; Try with a type/name first to not trip up on those that begin
5735 ;; with a keyword. Return t if a known or found type is moved
5736 ;; over. The point is clobbered if nil is returned. If range
5737 ;; recording is enabled, the identifier is recorded on as a type
5738 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5739 ;;
5740 ;; This macro might do hidden buffer changes.
5745 nil
5752 t)))
5755 ;; Used internally in `c-forward-keyword-clause' to move forward
5756 ;; over a comma separated list of types or names using
5757 ;; `c-forward-keyword-prefixed-id'.
5758 ;;
5759 ;; This macro might do hidden buffer changes.
5770 ;; Submatch MATCH in the current match data is assumed to surround a
5771 ;; token. If it's a keyword, move over it and any immediately
5772 ;; following clauses associated with it, stopping at the start of
5773 ;; the next token. t is returned in that case, otherwise the point
5774 ;; stays and nil is returned. The kind of clauses that are
5775 ;; recognized are those specified by `c-type-list-kwds',
5776 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5777 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5778 ;; and `c-<>-arglist-kwds'.
5779 ;;
5780 ;; This function records identifier ranges on
5781 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5782 ;; `c-record-type-identifiers' is non-nil.
5783 ;;
5784 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5785 ;; apply directly after the keyword, the type list is moved over
5786 ;; only when there is no unaccounted token before it (i.e. a token
5787 ;; that isn't moved over due to some other keyword list). The
5788 ;; identifier ranges in the list are still recorded if that should
5789 ;; be done, though.
5790 ;;
5791 ;; This function might do hidden buffer changes.
5794 ;; The call to `c-forward-<>-arglist' below is made after
5795 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5796 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5797 ;; should therefore be nil.
5809 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5814 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5819 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5826 ;; Use `c-forward-type' on every identifier we can find
5827 ;; inside the paren, to record the types.
5852 ;; If we are at the colon already, we move over the type
5853 ;; list after it.
5859 ;; Not at the colon, so stop here. But the identifier
5860 ;; ranges in the type list later on should still be
5861 ;; recorded.
5864 ;; If a keyword matched both one of the types above and
5865 ;; this one, we match `c-colon-type-list-re' after the
5866 ;; clause matched above.
5873 ;; There's a type after the `c-colon-type-list-re' match
5874 ;; after a keyword in `c-colon-type-list-kwds'.
5878 t)))
5880 ;; cc-mode requires cc-fonts.
5884 ;; The point is assumed to be at a "<". Try to treat it as the open
5885 ;; paren of an angle bracket arglist and move forward to the
5886 ;; corresponding ">". If successful, the point is left after the
5887 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5888 ;; returned. If ALL-TYPES is t then all encountered arguments in
5889 ;; the arglist that might be types are treated as found types.
5890 ;;
5891 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5892 ;; function handles text properties on the angle brackets and argument
5893 ;; separating commas.
5894 ;;
5895 ;; `c-restricted-<>-arglists' controls how lenient the template
5896 ;; arglist recognition should be.
5897 ;;
5898 ;; This function records identifier ranges on
5899 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5900 ;; `c-record-type-identifiers' is non-nil.
5901 ;;
5902 ;; This function might do hidden buffer changes.
5905 ;; If `c-record-type-identifiers' is set then activate
5906 ;; recording of any found types that constitute an argument in
5907 ;; the arglist.
5915 ;; `nconc' doesn't mind that the tail of
5916 ;; `c-record-found-types' is t.
5919 t)
5922 nil)))
5925 ;; Recursive part of `c-forward-<>-arglist'.
5926 ;;
5927 ;; This function might do hidden buffer changes.
5929 ;; Cover this so that any recorded found type ranges are
5930 ;; automatically lost if it turns out to not be an angle
5931 ;; bracket arglist. It's propagated through the return value
5932 ;; on successful completion.
5934 ;; List that collects the positions after the argument
5935 ;; separating ',' in the arglist.
5936 arg-start-pos)
5937 ;; If the '<' has paren open syntax then we've marked it as an angle
5938 ;; bracket arglist before, so skip to the end.
5946 t
5947 ;; Got unmatched paren angle brackets. We don't clear the paren
5948 ;; syntax properties and retry, on the basis that it's very
5949 ;; unlikely that paren angle brackets become operators by code
5950 ;; manipulation. It's far more likely that it doesn't match due
5951 ;; to narrowing or some temporary change.
5953 nil))
5958 ;; go forward one non-alphanumeric character (group) per iteration of
5959 ;; this loop.
5965 ;; All encountered identifiers are types, so set the
5966 ;; promote flag and parse the type.
5989 ;; Note: These regexps exploit the match order in \| so
5990 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5992 ;; Stop on ',', '|', '&', '+' and '-' to catch
5993 ;; common binary operators that could be between
5994 ;; two comparison expressions "a<b" and "c>d".
5996 nil t t))
6000 ;; Either an operator starting with '>' or the end of
6001 ;; the angle bracket arglist.
6008 ;; The angle bracket arglist is finished.
6020 ;; Either an operator starting with '<' or a nested arglist.
6024 ;; The '<' begins a multi-char operator.
6028 ;; We're at a nested <.....>
6034 ;; There's always an identifier before an angle
6035 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
6036 ;; or `c-<>-arglist-kwds'.
6052 ;; It was an angle bracket arglist.
6055 ;; Record the identifier before the template as a type
6056 ;; or reference depending on whether the arglist is last
6057 ;; in a qualified identifier.
6067 ;; At a "less than" operator.
6070 )))
6077 ;; Just another argument. Record the position. The
6078 ;; type check stuff that made us stop at it is at
6079 ;; the top of the loop.
6083 ;; Got a character that can't be in an angle bracket
6084 ;; arglist argument. Abort using `throw', since
6085 ;; it's useless to try to find a surrounding arglist
6086 ;; if we're nested.
6092 ;; The point is assumed to be directly after a ">". Try to treat it
6093 ;; as the close paren of an angle bracket arglist and move back to
6094 ;; the corresponding "<". If successful, the point is left at
6095 ;; the "<" and t is returned, otherwise the point isn't moved and
6096 ;; nil is returned. ALL-TYPES is passed on to
6097 ;; `c-forward-<>-arglist'.
6098 ;;
6099 ;; If the optional LIMIT is given, it bounds the backward search.
6100 ;; It's then assumed to be at a syntactically relevant position.
6101 ;;
6102 ;; This is a wrapper around `c-forward-<>-arglist'. See that
6103 ;; function for more details.
6112 t
6113 ;; See corresponding note in `c-forward-<>-arglist'.
6115 nil)
6122 t
6123 ;; Stopped at bob or a char that isn't allowed in an
6124 ;; arglist, so we've failed.
6126 nil)
6131 ;; If we moved then the "<" was part of some
6132 ;; multicharacter token.
6133 t
6139 ;; Matched the arglist. Break the while.
6141 nil)
6143 ;; We started from a non-paren ">" inside an
6144 ;; arglist.
6146 nil)
6148 ;; Matched a shorter arglist. Can be a nested
6149 ;; one so continue looking.
6151 t))
6152 t))))))
6157 ;; Move forward over a complete name if at the beginning of one,
6158 ;; stopping at the next following token. A keyword, as such,
6159 ;; doesn't count as a name. If the point is not at something that
6160 ;; is recognized as a name then it stays put.
6161 ;;
6162 ;; A name could be something as simple as "foo" in C or something as
6163 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6164 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6165 ;; int>::*volatile const" in C++ (this function is actually little
6166 ;; more than a `looking-at' call in all modes except those that,
6167 ;; like C++, have `c-recognize-<>-arglists' set).
6168 ;;
6169 ;; Return
6170 ;; o - nil if no name is found;
6171 ;; o - 'template if it's an identifier ending with an angle bracket
6172 ;; arglist;
6173 ;; o - 'operator of it's an operator identifier;
6174 ;; o - t if it's some other kind of name.
6175 ;;
6176 ;; This function records identifier ranges on
6177 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6178 ;; `c-record-type-identifiers' is non-nil.
6179 ;;
6180 ;; This function might do hidden buffer changes.
6183 ;; Turn off `c-promote-possible-types' here since we might
6184 ;; call `c-forward-<>-arglist' and we don't want it to promote
6185 ;; every suspect thing in the arglist to a type. We're
6186 ;; typically called from `c-forward-type' in this case, and
6187 ;; the caller only wants the top level type that it finds to
6188 ;; be promoted.
6189 c-promote-possible-types)
6195 ;; Check for keyword. We go to the last symbol in
6196 ;; `c-identifier-key' first.
6209 ;; "template" is only valid inside an
6210 ;; identifier if preceded by "::".
6215 t))
6217 ;; Handle a C++ operator or template identifier.
6221 ;; Got "... ::template".
6225 res subres))))
6228 ;; Got a cast operator.
6232 ;; Now we should match a sequence of either
6233 ;; '*', '&' or a name followed by ":: *",
6234 ;; where each can be followed by a sequence
6235 ;; of `c-opt-type-modifier-key'.
6238 t)
6248 t))))
6256 ;; Got some other operator.
6264 nil)
6266 ;; `id-start' is equal to `id-end' if we've jumped over
6267 ;; an identifier that doesn't end with a symbol token.
6268 ;; That can occur e.g. for Java import directives on the
6269 ;; form "foo.bar.*".
6276 res t)))
6286 ;; Got a concatenated identifier. This handles the
6287 ;; cases with tricky syntactic whitespace that aren't
6288 ;; covered in `c-identifier-key'.
6291 t)
6295 ;; Maybe an angle bracket arglist.
6303 c-last-identifier-range nil)
6308 ;; Continue if there's an identifier concatenation
6309 ;; operator after the template argument.
6315 t)
6320 nil)))
6321 )))))
6324 res))
6327 ;; Move forward over a type spec if at the beginning of one,
6328 ;; stopping at the next following token. The keyword "typedef"
6329 ;; isn't part of a type spec here.
6330 ;;
6331 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6332 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6333 ;; The current (2009-03-10) intention is to convert all uses of
6334 ;; `c-forward-type' to call with this parameter set, then to
6335 ;; eliminate it.
6336 ;;
6337 ;; Return
6338 ;; o - t if it's a known type that can't be a name or other
6339 ;; expression;
6340 ;; o - 'known if it's an otherwise known type (according to
6341 ;; `*-font-lock-extra-types');
6342 ;; o - 'prefix if it's a known prefix of a type;
6343 ;; o - 'found if it's a type that matches one in `c-found-types';
6344 ;; o - 'maybe if it's an identifier that might be a type;
6345 ;; o - 'decltype if it's a decltype(variable) declaration; - or
6346 ;; o - nil if it can't be a type (the point isn't moved then).
6347 ;;
6348 ;; The point is assumed to be at the beginning of a token.
6349 ;;
6350 ;; Note that this function doesn't skip past the brace definition
6351 ;; that might be considered part of the type, e.g.
6352 ;; "enum {a, b, c} foo".
6353 ;;
6354 ;; This function records identifier ranges on
6355 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6356 ;; `c-record-type-identifiers' is non-nil.
6357 ;;
6358 ;; This function might do hidden buffer changes.
6366 ;; Skip leading type modifiers. If any are found we know it's a
6367 ;; prefix of a type.
6386 ; "typedef".
6394 ;; In many languages the name can be used without the
6395 ;; prefix, so we add it to `c-found-types'.
6398 c-last-identifier-range)
6420 id-range c-last-identifier-range))))
6432 id-end)
6434 ;; Looking at a primitive or known type identifier. We've
6435 ;; checked for a name first so that we don't go here if the
6436 ;; known type match only is a prefix of another name.
6447 ;; There might be more keywords for the type.
6475 ;; A normal identifier.
6485 ;; It's an identifier that has been used as
6486 ;; a type somewhere else.
6487 'found
6488 ;; It's an identifier that might be a type.
6491 ;; A template is a type.
6495 ;; Otherwise it's an operator identifier, which is not a type.
6500 ;; Skip trailing type modifiers. If any are found we know it's
6501 ;; a type.
6508 ;; Step over any type suffix operator. Do not let the existence
6509 ;; of these alter the classification of the found type, since
6510 ;; these operators typically are allowed in normal expressions
6511 ;; too.
6518 ;; Look for a trailing operator that concatenates the type
6519 ;; with a following one, and if so step past that one through
6520 ;; a recursive call. Note that we don't record concatenated
6521 ;; types in `c-found-types' - it's the component types that
6522 ;; are recorded when appropriate.
6525 c-promote-possible-types))
6526 ;; If we can't promote then set `c-record-found-types' so that
6527 ;; we can merge in the types from the second part afterwards if
6528 ;; it turns out to be a known type there.
6531 subres)
6540 ;; If either operand certainly is a type then both are, but we
6541 ;; don't let the existence of the operator itself promote two
6542 ;; uncertain types to a certain one.
6561 ;; Merge in the ranges of any types found by the second
6562 ;; `c-forward-type'.
6564 ;; `nconc' doesn't mind that the tail of
6565 ;; `c-record-found-types' is t.
6567 c-record-type-identifiers))))
6575 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6577 res))
6580 ;; Used for Java code only at the moment. Assumes point is on the
6581 ;; @, moves forward an annotation. returns nil if there is no
6582 ;; annotation at point.
6589 t)))
6592 ;; Pull the next open brace from PS (which has the form of paren-state),
6593 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
6601 ;; Point is just after a closing brace/parenthesis. Try to parse this as a
6602 ;; C++ member initializer list, going back to just after the introducing ":"
6603 ;; and returning t. Otherwise return nil, leaving point unchanged.
6628 res))
6631 ;; Test whether we are in a C++ member initializer list, and if so, go back
6632 ;; to the introducing ":", returning the position of the opening paren of
6633 ;; the function's arglist. Otherwise return nil, leaving point unchanged.
6636 res)
6674 res))
6676 \f
6677 ;; Handling of large scale constructs like statements and declarations.
6679 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6680 ;; defsubst or perhaps even a defun, but it contains lots of free
6681 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6683 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6684 ;; of types when parsing a declaration, which means that it
6685 ;; sometimes consumes the identifier in the declaration as a type.
6686 ;; This is used to "backtrack" and make the last type be treated as
6687 ;; an identifier instead.
6690 ;; These identifiers are bound only in the inner let.
6692 identifier-start type-start
6693 got-parens nil
6694 got-identifier t
6695 got-suffix t
6696 got-suffix-after-parens id-start
6700 t
6701 backup-at-type))
6703 id-start backup-id-start)
6705 id-start start-pos))
6707 ;; When these flags already are set we've found specifiers that
6708 ;; unconditionally signal these attributes - backtracking doesn't
6709 ;; change that. So keep them set in that case.
6716 ;; This identifier is bound only in the inner let.
6720 ;; Move forward over a declaration or a cast if at the start of one.
6721 ;; The point is assumed to be at the start of some token. Nil is
6722 ;; returned if no declaration or cast is recognized, and the point
6723 ;; is clobbered in that case.
6724 ;;
6725 ;; If a declaration is parsed:
6726 ;;
6727 ;; The point is left at the first token after the first complete
6728 ;; declarator, if there is one. The return value is a cons where
6729 ;; the car is the position of the first token in the declarator. (See
6730 ;; below for the cdr.)
6731 ;; Some examples:
6732 ;;
6733 ;; void foo (int a, char *b) stuff ...
6734 ;; car ^ ^ point
6735 ;; float (*a)[], b;
6736 ;; car ^ ^ point
6737 ;; unsigned int a = c_style_initializer, b;
6738 ;; car ^ ^ point
6739 ;; unsigned int a (cplusplus_style_initializer), b;
6740 ;; car ^ ^ point (might change)
6741 ;; class Foo : public Bar {}
6742 ;; car ^ ^ point
6743 ;; class PikeClass (int a, string b) stuff ...
6744 ;; car ^ ^ point
6745 ;; enum bool;
6746 ;; car ^ ^ point
6747 ;; enum bool flag;
6748 ;; car ^ ^ point
6749 ;; void cplusplus_function (int x) throw (Bad);
6750 ;; car ^ ^ point
6751 ;; Foo::Foo (int b) : Base (b) {}
6752 ;; car ^ ^ point
6753 ;;
6754 ;; auto foo = 5;
6755 ;; car ^ ^ point
6756 ;; auto cplusplus_11 (int a, char *b) -> decltype (bar):
6757 ;; car ^ ^ point
6758 ;;
6759 ;;
6760 ;;
6761 ;; The cdr of the return value is non-nil when a
6762 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6763 ;; Specifically it is a dotted pair (A . B) where B is t when a
6764 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6765 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6766 ;; specifier is present. I.e., (some of) the declared
6767 ;; identifier(s) are types.
6768 ;;
6769 ;; If a cast is parsed:
6770 ;;
6771 ;; The point is left at the first token after the closing paren of
6772 ;; the cast. The return value is `cast'. Note that the start
6773 ;; position must be at the first token inside the cast parenthesis
6774 ;; to recognize it.
6775 ;;
6776 ;; PRECEDING-TOKEN-END is the first position after the preceding
6777 ;; token, i.e. on the other side of the syntactic ws from the point.
6778 ;; Use a value less than or equal to (point-min) if the point is at
6779 ;; the first token in (the visible part of) the buffer.
6780 ;;
6781 ;; CONTEXT is a symbol that describes the context at the point:
6782 ;; 'decl In a comma-separated declaration context (typically
6783 ;; inside a function declaration arglist).
6784 ;; '<> In an angle bracket arglist.
6785 ;; 'arglist Some other type of arglist.
6786 ;; nil Some other context or unknown context. Includes
6787 ;; within the parens of an if, for, ... construct.
6788 ;;
6789 ;; LAST-CAST-END is the first token after the closing paren of a
6790 ;; preceding cast, or nil if none is known. If
6791 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6792 ;; the position after the closest preceding call where a cast was
6793 ;; matched. In that case it's used to discover chains of casts like
6794 ;; "(a) (b) c".
6795 ;;
6796 ;; This function records identifier ranges on
6797 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6798 ;; `c-record-type-identifiers' is non-nil.
6799 ;;
6800 ;; This function might do hidden buffer changes.
6803 ;; first type, i.e. after any leading specifiers. It might
6804 ;; also point at the beginning of the preceding syntactic
6805 ;; whitespace.
6807 ;; Set to the result of `c-forward-type'.
6808 at-type
6809 ;; The position of the first token in what we currently
6810 ;; believe is the type in the declaration or cast, after any
6811 ;; specifiers and their associated clauses.
6812 type-start
6813 ;; The position of the first token in what we currently
6814 ;; believe is the declarator for the first identifier. Set
6815 ;; when the type is found, and moved forward over any
6816 ;; `c-decl-hangon-kwds' and their associated clauses that
6817 ;; occurs after the type.
6818 id-start
6819 ;; These store `at-type', `type-start' and `id-start' of the
6820 ;; identifier before the one in those variables. The previous
6821 ;; identifier might turn out to be the real type in a
6822 ;; declaration if the last one has to be the declarator in it.
6823 ;; If `backup-at-type' is nil then the other variables have
6824 ;; undefined values.
6825 backup-at-type backup-type-start backup-id-start
6826 ;; This stores `kwd-sym' of the symbol before the current one.
6827 ;; This is needed to distinguish the C++11 version of "auto" from
6828 ;; the pre C++11 meaning.
6829 backup-kwd-sym
6830 ;; Set if we've found a specifier (apart from "typedef") that makes
6831 ;; the defined identifier(s) types.
6832 at-type-decl
6833 ;; Set if we've a "typedef" keyword.
6834 at-typedef
6835 ;; Set if we've found a specifier that can start a declaration
6836 ;; where there's no type.
6837 maybe-typeless
6838 ;; Save the value of kwd-sym between loops of the "Check for a
6839 ;; type" loop. Needed to distinguish a C++11 "auto" from a pre
6840 ;; C++11 one.
6841 prev-kwd-sym
6842 ;; If a specifier is found that also can be a type prefix,
6843 ;; these flags are set instead of those above. If we need to
6844 ;; back up an identifier, they are copied to the real flag
6845 ;; variables. Thus they only take effect if we fail to
6846 ;; interpret it as a type.
6847 backup-at-type-decl backup-maybe-typeless
6848 ;; Whether we've found a declaration or a cast. We might know
6849 ;; this before we've found the type in it. It's 'ids if we've
6850 ;; found two consecutive identifiers (usually a sure sign, but
6851 ;; we should allow that in labels too), and t if we've found a
6852 ;; specifier keyword (a 100% sure sign).
6853 at-decl-or-cast
6854 ;; Set when we need to back up to parse this as a declaration
6855 ;; but not as a cast.
6856 backup-if-not-cast
6857 ;; For casts, the return position.
6858 cast-end
6859 ;; Have we got a new-style C++11 "auto"?
6860 new-style-auto
6861 ;; Save `c-record-type-identifiers' and
6862 ;; `c-record-ref-identifiers' since ranges are recorded
6863 ;; speculatively and should be thrown away if it turns out
6864 ;; that it isn't a declaration or cast.
6871 ;; Check for a type. Unknown symbols are treated as possible
6872 ;; types, but they could also be specifiers disguised through
6873 ;; macros like __INLINE__, so we recognize both types and known
6874 ;; specifiers after them too.
6878 ;; Look for a specifier keyword clause.
6891 ;; Found a known or possible type or a prefix of a known type.
6900 ;; Got two identifiers with nothing but whitespace
6901 ;; between them. That can only happen in declarations.
6905 ;; If the previous identifier is a found type we
6906 ;; record it as a real one; it might be some sort of
6907 ;; alias for a prefix like "unsigned".
6914 backup-type-start type-start
6915 backup-id-start id-start
6916 backup-kwd-sym kwd-sym
6917 at-type found-type
6918 type-start start
6920 ;; The previous ambiguous specifier/type turned out
6921 ;; to be a type since we've parsed another one after
6922 ;; it, so clear these backup flags.
6923 backup-at-type-decl nil
6924 backup-maybe-typeless nil))
6928 ;; Handle known specifier keywords and
6929 ;; `c-decl-hangon-kwds' which can occur after known
6930 ;; types.
6933 ;; It's a hang-on keyword that can occur anywhere.
6937 ;; Move the identifier start position if
6938 ;; we've passed a type.
6940 ;; Otherwise treat this as a specifier and
6941 ;; move the fallback position.
6945 ;; It's an ordinary specifier so we know that
6946 ;; anything before this can't be the type.
6948 start-pos kwd-clause-end)
6951 ;; It's ambiguous whether this keyword is a
6952 ;; specifier or a type prefix, so set the backup
6953 ;; flags. (It's assumed that `c-forward-type'
6954 ;; moved further than `c-forward-keyword-clause'.)
6962 ;; This test only happens after we've scanned a type.
6963 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6968 ;; Haven't matched a type so it's an unambiguous
6969 ;; specifier keyword and we know we're in a
6970 ;; declaration.
6976 ;; If the type isn't known we continue so that we'll jump
6977 ;; over all specifiers and type identifiers. The reason
6978 ;; to do this for a known type prefix is to make things
6979 ;; like "unsigned INT16" work.
6984 ;; If a known type was found, we still need to skip over any
6985 ;; hangon keyword clauses after it. Otherwise it has already
6986 ;; been done in the loop above.
6992 ;; A prefix type is itself a primitive type when it's not
6993 ;; followed by another type.
6997 ;; Got no type but set things up to continue anyway to handle
6998 ;; the various cases when a declaration doesn't start with a
6999 ;; type.
7004 ;; If it's C++ then check if the last "type" ends on the form
7005 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
7006 ;; (con|de)structor.
7017 ;; Cheating in the handling of syntactic ws below.
7024 name))
7025 ;; It is a (con|de)structor name. In that case the
7026 ;; declaration is typeless so zap out any preceding
7027 ;; identifier(s) that we might have taken as types.
7030 backup-at-type nil
7031 id-start type-start))))))
7033 ;; Check for and step over a type decl expression after the thing
7034 ;; that is or might be a type. This can't be skipped since we
7035 ;; need the correct end position of the declarator for
7036 ;; `max-type-decl-end-*'.
7038 ;; True if there's a non-open-paren match of
7039 ;; `c-type-decl-prefix-key'.
7040 got-prefix
7041 ;; True if the declarator is surrounded by a parenthesis pair.
7042 got-parens
7043 ;; True if there is an identifier in the declarator.
7044 got-identifier
7045 ;; True if there's a non-close-paren match of
7046 ;; `c-type-decl-suffix-key'.
7047 got-suffix
7048 ;; True if there's a prefix match outside the outermost
7049 ;; paren pair that surrounds the declarator.
7050 got-prefix-before-parens
7051 ;; True if there's a suffix match outside the outermost
7052 ;; paren pair that surrounds the declarator. The value is
7053 ;; the position of the first suffix match.
7054 got-suffix-after-parens
7055 ;; True if we've parsed the type decl to a token that is
7056 ;; known to end declarations in this context.
7057 at-decl-end
7058 ;; The earlier values of `at-type' and `type-start' if we've
7059 ;; shifted the type backwards.
7060 identifier-type identifier-start
7061 ;; If `c-parse-and-markup-<>-arglists' is set we need to
7062 ;; turn it off during the name skipping below to avoid
7063 ;; getting `c-type' properties that might be bogus. That
7064 ;; can happen since we don't know if
7065 ;; `c-restricted-<>-arglists' will be correct inside the
7066 ;; arglist paren that gets entered.
7067 c-parse-and-markup-<>-arglists
7068 ;; Start of the identifier for which `got-identifier' was set.
7069 name-start)
7073 ;; Skip over type decl prefix operators. (Note similar code in
7074 ;; `c-font-lock-declarators'.)
7084 ;; If the third submatch matches in C++ then
7085 ;; we're looking at an identifier that's a
7086 ;; prefix only if it specifies a member pointer.
7091 ;; We only check for a trailing "::" and
7092 ;; let the "*" that should follow be
7093 ;; matched in the next round.
7095 ;; It turned out to be the real identifier,
7096 ;; so stop.
7097 nil))
7098 t))
7112 ;; Skip over an identifier.
7119 ;; Skip over type decl suffix operators.
7126 t)
7130 t)
7136 ;; No suffix matched. We might have matched the
7137 ;; identifier as a type and the open paren of a
7138 ;; function arglist as a type decl prefix. In that
7139 ;; case we should "backtrack": Reinterpret the last
7140 ;; type as the identifier, move out of the arglist and
7141 ;; continue searching for suffix operators.
7142 ;;
7143 ;; Do this even if there's no preceding type, to cope
7144 ;; with old style function declarations in K&R C,
7145 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
7146 ;; style declarations. That isn't applicable in an
7147 ;; arglist context, though.
7152 maybe-typeless
7153 backup-maybe-typeless
7160 t))
7169 at-type))
7170 ;; Have found no identifier but `c-typeless-decl-kwds' has
7171 ;; matched so we know we're inside a declaration. The
7172 ;; preceding type must be the identifier instead.
7175 ;; Prepare the "-> type;" for fontification later on.
7185 at-decl-or-cast
7188 ;; CASE 1
7190 ;; Encountered something inside parens that isn't matched by
7191 ;; the `c-type-decl-*' regexps, so it's not a type decl
7192 ;; expression. Try to skip out to the same paren depth to
7193 ;; not confuse the cast check below.
7195 ;; If we've found a specifier keyword then it's a
7196 ;; declaration regardless.
7204 ;; Now we've collected info about various characteristics of
7205 ;; the construct we're looking at. Below follows a decision
7206 ;; tree based on that. It's ordered to check more certain
7207 ;; signs before less certain ones.
7212 ;; CASE 2
7215 ;; Got another identifier directly after the type, so it's a
7216 ;; declaration.
7221 ;; (not got-suffix-after-parens)
7223 maybe-typeless
7224 backup-maybe-typeless
7229 ;; Got a declaration of the form "foo bar (gnu);" or "bar
7230 ;; (gnu);" where we've recognized "bar" as the type and "gnu"
7231 ;; as the declarator. In this case it's however more likely
7232 ;; that "bar" is the declarator and "gnu" a function argument
7233 ;; or initializer (if `c-recognize-paren-inits' is set),
7234 ;; since the parens around "gnu" would be superfluous if it's
7235 ;; a declarator. Shift the type one step backward.
7238 ;; Found no identifier.
7243 ;; CASE 3
7245 ;; Got a plain list of identifiers. If a colon follows it's
7246 ;; a valid label, or maybe a bitfield. Otherwise the last
7247 ;; one probably is the declared identifier and we should
7248 ;; back up to the previous type, providing it isn't a cast.
7252 ;; If we've found a specifier keyword then it's a
7253 ;; declaration regardless.
7264 ;; CASE 4
7268 ;; Got a plain list of identifiers followed by some suffix.
7269 ;; If this isn't a cast then the last identifier probably is
7270 ;; the declared one and we should back up to the previous
7271 ;; type.
7275 ;; CASE 5
7277 ;; If the type is known we know that there can't be any
7278 ;; identifier somewhere else, and it's only in declarations in
7279 ;; e.g. function prototypes and in casts that the identifier may
7280 ;; be left out.
7284 ;; Only got a single identifier (parsed as a type so far).
7285 ;; CASE 6
7287 ;; Check that the identifier isn't at the start of an
7288 ;; expression.
7289 at-decl-end
7292 ;; Inside an arglist that contains declarations. If K&R
7293 ;; style declarations and parenthesis style initializers
7294 ;; aren't allowed then the single identifier must be a
7295 ;; type, else we require that it's known or found
7296 ;; (primitive types are handled above).
7301 ;; Inside a template arglist. Accept known and found
7302 ;; types; other identifiers could just as well be
7303 ;; constants in C++.
7306 ;; CASE 7
7307 ;; Can't be a valid declaration or cast, but if we've found a
7308 ;; specifier it can't be anything else either, so treat it as
7309 ;; an invalid/unfinished declaration or cast.
7317 maybe-typeless
7318 backup-maybe-typeless
7322 c-after-suffixed-type-maybe-decl-key))))))
7323 ;; Got an empty paren pair and a preceding type that probably
7324 ;; really is the identifier. Shift the type backwards to make
7325 ;; the last one the identifier. This is analogous to the
7326 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7327 ;; above.
7328 ;;
7329 ;; Exception: In addition to the conditions in that
7330 ;; "backtracking" code, do not shift backward if we're not
7331 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7332 ;; Since there's no preceding type, the shift would mean that
7333 ;; the declaration is typeless. But if the regexp doesn't match
7334 ;; then we will simply fall through in the tests below and not
7335 ;; recognize it at all, so it's better to try it as an abstract
7336 ;; declarator instead.
7339 ;; Still no identifier.
7340 ;; CASE 8
7342 ;; Require `got-prefix' together with either `got-parens' or
7343 ;; `got-suffix' to recognize it as an abstract declarator:
7344 ;; `got-parens' only is probably an empty function call.
7345 ;; `got-suffix' only can build an ordinary expression together
7346 ;; with the preceding identifier which we've taken as a type.
7347 ;; We could actually accept on `got-prefix' only, but that can
7348 ;; easily occur temporarily while writing an expression so we
7349 ;; avoid that case anyway. We could do a better job if we knew
7350 ;; the point when the fontification was invoked.
7353 ;; CASE 9
7357 got-suffix-after-parens
7359 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7360 ;; normal function call after all (or perhaps a C++ style object
7361 ;; instantiation expression).
7364 ;; CASE 10
7366 ;; By now we've located the type in the declaration that we know
7367 ;; we're in.
7370 ;; CASE 11
7378 ;; Shift the type backward in the case that there's a
7379 ;; single identifier inside parens. That can only
7380 ;; occur in K&R style function declarations so it's
7381 ;; more likely that it really is a function call.
7382 ;; Therefore we only do this after
7383 ;; `c-after-suffixed-type-decl-key' has matched.
7385 got-suffix-after-parens))
7386 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7389 ;; CASE 12
7392 ;; It's a declaration if a known type precedes it and it can't be a
7393 ;; function call.
7396 ;; If we get here we can't tell if this is a type decl or a normal
7397 ;; expression by looking at it alone. (That's under the assumption
7398 ;; that normal expressions always can look like type decl expressions,
7399 ;; which isn't really true but the cases where it doesn't hold are so
7400 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7401 ;; the effort to look for them.)
7403 ;;; 2008-04-16: commented out the next form, to allow the function to recognize
7404 ;;; "foo (int bar)" in CC (an implicit type (in class foo) without a semicolon)
7405 ;;; as a(n almost complete) declaration, enabling it to be fontified.
7406 ;; CASE 13
7407 ;; (unless (or at-decl-end (looking-at "=[^=]"))
7408 ;; If this is a declaration it should end here or its initializer(*)
7409 ;; should start here, so check for allowed separation tokens. Note
7410 ;; that this rule doesn't work e.g. with a K&R arglist after a
7411 ;; function header.
7412 ;;
7413 ;; *) Don't check for C++ style initializers using parens
7414 ;; since those already have been matched as suffixes.
7415 ;;
7416 ;; If `at-decl-or-cast' is then we've found some other sign that
7417 ;; it's a declaration or cast, so then it's probably an
7418 ;; invalid/unfinished one.
7419 ;; (throw 'at-decl-or-cast at-decl-or-cast))
7421 ;; Below are tests that only should be applied when we're certain to
7422 ;; not have parsed halfway through an expression.
7424 ;; CASE 14
7426 ;; The expression starts with a known type so treat it as a
7427 ;; declaration.
7430 ;; CASE 15
7432 ;; In C++ we check if the identifier is a known type, since
7433 ;; (con|de)structors use the class name as identifier.
7434 ;; We've always shifted over the identifier as a type and
7435 ;; then backed up again in this case.
7436 identifier-type
7439 ;; `at-type' probably won't be 'found for
7440 ;; destructors since the "~" is then part of the
7441 ;; type name being checked against the list of
7442 ;; known types, so do a check without that
7443 ;; operator.
7448 c-identifier-syntax-table
7452 ;; We have already parsed the type earlier,
7453 ;; so it'd be possible to cache the end
7454 ;; position instead of redoing it here, but
7455 ;; then we'd need to keep track of another
7456 ;; position everywhere.
7464 ;; CASE 16
7466 at-type
7470 ;; Got something like "foo * bar;". Since we're not inside an
7471 ;; arglist it would be a meaningless expression because the
7472 ;; result isn't used. We therefore choose to recognize it as
7473 ;; a declaration. Do not allow a suffix since it could then
7474 ;; be a function call.
7477 ;; CASE 17
7482 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7483 ;; be an odd expression or it could be a declaration. Treat
7484 ;; it as a declaration if "a" has been used as a type
7485 ;; somewhere else (if it's a known type we won't get here).
7488 ;; CASE 18
7494 ;; Got a type followed by an abstract declarator. If `got-prefix'
7495 ;; is set it's something like "a *" without anything after it. If
7496 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7497 ;; or similar, which we accept only if the context rules out
7498 ;; expressions.
7501 ;; If we had a complete symbol table here (which rules out
7502 ;; `c-found-types') we should return t due to the disambiguation rule
7503 ;; (in at least C++) that anything that can be parsed as a declaration
7504 ;; is a declaration. Now we're being more defensive and prefer to
7505 ;; highlight things like "foo (bar);" as a declaration only if we're
7506 ;; inside an arglist that contains declarations.
7507 ;; CASE 19
7510 ;; The point is now after the type decl expression.
7513 ;; Check for a cast.
7516 c-cast-parens
7518 ;; Should be the first type/identifier in a cast paren.
7522 ;; The closing paren should follow.
7527 ;; There should be a primary expression after it.
7536 ;; Check if the expression begins with a prefix keyword.
7539 ;; Expression begins with an ambiguous operator. Treat
7540 ;; it as a cast if it's a type decl or if we've
7541 ;; recognized the type somewhere else.
7544 ;; Unless it's a keyword, it's the beginning of a primary
7545 ;; expression.
7547 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7548 ;; that it matched a whole one so that we don't e.g. confuse
7549 ;; the operator '-' with '->'. It's ok if it matches further,
7550 ;; though, since it e.g. can match the float '.5' while the
7551 ;; operator regexp only matches '.'.
7555 ;; There should either be a cast before it or something that isn't an
7556 ;; identifier or close paren.
7564 ;; It's a symbol. Accept it only if it's one of the
7565 ;; keywords that can precede an expression (without
7566 ;; surrounding parens).
7569 ;; Check that it isn't a close paren (block close is ok,
7570 ;; though).
7572 ;; Check that it isn't a nonsymbol identifier.
7575 ;; Handle the cast.
7585 ;; We're at a declaration. Highlight the type and the following
7586 ;; declarators.
7592 ;; Make sure to propagate the `c-decl-arg-start' property to
7593 ;; the next argument if it's set in this one, to cope with
7594 ;; interactive refontification.
7597 ;; Record the type's coordinates in `c-record-type-identifiers' for
7598 ;; later fontification.
7600 ;; There seems no reason to exclude a token from
7601 ;; fontification just because it's "a known type that can't
7602 ;; be a name or other expression". 2013-09-18.
7603 )
7614 ;; False alarm. Restore the recorded ranges.
7616 c-record-ref-identifiers save-rec-ref-ids)
7617 nil))))
7620 ;; Assuming that point is at the beginning of a token, check if it starts a
7621 ;; label and if so move over it and return non-nil (t in default situations,
7622 ;; specific symbols (see below) for interesting situations), otherwise don't
7623 ;; move and return nil. "Label" here means "most things with a colon".
7624 ;;
7625 ;; More precisely, a "label" is regarded as one of:
7626 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7627 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7628 ;; bare "case", should the colon be missing. We return t;
7629 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7630 ;; return t;
7631 ;; (iv) One of QT's "extended" C++ variants of
7632 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7633 ;; Returns the symbol `qt-2kwds-colon'.
7634 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7635 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7636 ;; colon). Currently (2006-03), this applies only to Objective C's
7637 ;; keywords "@private", "@protected", and "@public". Returns t.
7638 ;;
7639 ;; One of the things which will NOT be recognized as a label is a bit-field
7640 ;; element of a struct, something like "int foo:5".
7641 ;;
7642 ;; The end of the label is taken to be just after the colon, or the end of
7643 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7644 ;; after the end on return. The terminating char gets marked with
7645 ;; `c-decl-end' to improve recognition of the following declaration or
7646 ;; statement.
7647 ;;
7648 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7649 ;; label, if any, has already been marked up like that.
7650 ;;
7651 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7652 ;; after the preceding token, i.e. on the other side of the
7653 ;; syntactic ws from the point. Use a value less than or equal to
7654 ;; (point-min) if the point is at the first token in (the visible
7655 ;; part of) the buffer.
7656 ;;
7657 ;; The optional LIMIT limits the forward scan for the colon.
7658 ;;
7659 ;; This function records the ranges of the label symbols on
7660 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7661 ;; non-nil.
7662 ;;
7663 ;; This function might do hidden buffer changes.
7666 label-end
7667 qt-symbol-idx
7668 macro-start ; if we're in one.
7669 label-type
7670 kwd)
7672 ;; "case" or "default" (Doesn't apply to AWK).
7675 ;; Record only the keyword itself for fontification, since in
7676 ;; case labels the following is a constant expression and not
7677 ;; a label.
7681 ;; Find the label end.
7685 ;; Stop on chars that aren't allowed in expressions,
7686 ;; and on operator chars that would be meaningless
7687 ;; there. FIXME: This doesn't cope with ?: operators.
7695 t)
7697 ;; It's an unfinished label. We consider the keyword enough
7698 ;; to recognize it as a label, so that it gets fontified.
7699 ;; Leave the point at the end of it, but don't put any
7700 ;; `c-decl-end' marker.
7702 t))))
7704 ;; @private, @protected, @public, in Objective C, or similar.
7707 ;; For a `c-opt-extra-label-key' match, we record the whole
7708 ;; thing for fontification. That's to get the leading '@' in
7709 ;; Objective-C protection labels fontified.
7716 ;; All other cases of labels.
7719 ;; A colon label must have something before the colon.
7722 ;; Check that we're not after a token that can't precede a label.
7724 ;; Trivially succeeds when there's no preceding token.
7725 ;; Succeeds when we're at a virtual semicolon.
7734 ;; Check if we're after a label, if we're after a closing
7735 ;; paren that belong to statement, and with
7736 ;; `c-label-prefix-re'. It's done in different order
7737 ;; depending on `assume-markup' since the checks have
7738 ;; different expensiveness.
7765 ;; Might be after another label, so check it recursively.
7769 ;; Essentially the same as the
7770 ;; `c-syntactic-re-search-forward' regexp below.
7776 ;; Note: the following should work instead of the
7777 ;; narrow-to-region above. Investigate why not,
7778 ;; sometime. ACM, 2006-03-31.
7779 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7780 ;; macro-start t)
7782 ;; If the caller turned on recording for us,
7783 ;; it shouldn't apply when we check the
7784 ;; preceding label.
7785 c-record-type-identifiers)
7786 ;; A label can't start at a cpp directive. Check for
7787 ;; this, since c-forward-syntactic-ws would foul up on it.
7792 ;; Point is still at the beginning of the possible label construct.
7793 ;;
7794 ;; Check that the next nonsymbol token is ":", or that we're in one
7795 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7796 ;; arguments. FIXME: Should build this regexp from the language
7797 ;; constants.
7799 ;; public: protected: private:
7809 ;; QT double keyword like "protected slots:" or goto target.
7828 'qt-1kwd-colon
7840 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7850 t)
7853 ;; Record the identifiers in the label for fontification, unless
7854 ;; it begins with `c-label-kwds' in which case the following
7855 ;; identifiers are part of a (constant) expression that
7856 ;; shouldn't be fontified.
7868 ;; Not a label.
7870 label-type))
7873 ;; Assuming the point is at the beginning of a token, try to move
7874 ;; forward to the end of the Objective-C directive that starts
7875 ;; there. Return t if a directive was fully recognized, otherwise
7876 ;; the point is moved as far as one could be successfully parsed and
7877 ;; nil is returned.
7878 ;;
7879 ;; This function records identifier ranges on
7880 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7881 ;; `c-record-type-identifiers' is non-nil.
7882 ;;
7883 ;; This function might do hidden buffer changes.
7886 start-char
7888 lim
7889 ;; Turn off recognition of angle bracket arglists while parsing
7890 ;; types here since the protocol reference list might then be
7891 ;; considered part of the preceding name or superclass-name.
7902 t)
7911 ;; Handle the name of the class itself.
7913 ;; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7914 ;; at EOB.
7921 ;; Look for ": superclass-name" or "( category-name )".
7932 ;; Look for a protocol reference list.
7938 t))))
7944 t)
7947 nil)))
7950 ;; Go to the first non-whitespace after the colon that starts a
7951 ;; multiple inheritance introduction. Optional LIM is the farthest
7952 ;; back we should search.
7953 ;;
7954 ;; This function might do hidden buffer changes.
7962 ;; Return nil if we aren't in a method definition, otherwise the
7963 ;; position of the initial [+-].
7964 ;;
7965 ;; This function might do hidden buffer changes.
7971 ))
7973 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7975 ;; Return non-nil if point is within a gcc \"asm\" block.
7976 ;;
7977 ;; This should be called with point inside an argument list.
7978 ;;
7979 ;; Only one level of enclosing parentheses is considered, so for
7980 ;; instance nil is returned when in a function call within an asm
7981 ;; operand.
7982 ;;
7983 ;; This function might do hidden buffer changes.
7995 a function.
7997 More precisely, being at the top-level means that point is either
7998 outside any enclosing block (such as a function definition), or
7999 directly inside a class, namespace or other block that contains
8000 another declaration level.
8002 If point is not at the top-level (e.g. it is inside a method
8003 definition), then nil is returned. Otherwise, if point is at a
8004 top-level not enclosed within a class definition, t is returned.
8005 Otherwise, a 2-vector is returned where the zeroth element is the
8006 buffer position of the start of the class declaration, and the first
8007 element is the buffer position of the enclosing class's opening
8008 brace.
8010 Note that this function might do hidden buffer changes. See the
8011 comment at the start of cc-engine.el for more info."
8017 ;; Return non-nil if the point is in the region after the argument
8018 ;; list of a function and its opening brace (or semicolon in case it
8019 ;; got no body). If there are K&R style argument declarations in
8020 ;; that region, the point has to be inside the first one for this
8021 ;; function to recognize it.
8022 ;;
8023 ;; If successful, the point is moved to the first token after the
8024 ;; function header (see `c-forward-decl-or-cast-1' for details) and
8025 ;; the position of the opening paren of the function arglist is
8026 ;; returned.
8027 ;;
8028 ;; The point is clobbered if not successful.
8029 ;;
8030 ;; LIM is used as bound for backward buffer searches.
8031 ;;
8032 ;; This function might do hidden buffer changes.
8045 ;; There should not be a '=' or ',' between beg and the
8046 ;; start of the declaration since that means we were in the
8047 ;; "expression part" of the declaration.
8052 ;; Check that there's an arglist paren in the
8053 ;; declaration.
8056 ;; The declarator is a paren expression, so skip past it
8057 ;; so that we don't get stuck on that instead of the
8058 ;; function arglist.
8062 ;; Don't trip up on "operator ()".
8069 ;; Return the position of the first argument declaration if point is
8070 ;; inside a K&R style argument declaration list, nil otherwise.
8071 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
8072 ;; position that bounds the backward search for the argument list.
8073 ;;
8074 ;; Point must be within a possible K&R region, e.g. just before a top-level
8075 ;; "{". It must be outside of parens and brackets. The test can return
8076 ;; false positives otherwise.
8077 ;;
8078 ;; This function might do hidden buffer changes.
8082 ;; If we're in a macro, our search range is restricted to it. Narrow to
8083 ;; the searchable range.
8087 before-lparen after-rparen
8089 ; we give up
8092 ;; Search backwards for the defun's argument list. We give up if we
8093 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
8094 ;; a knr region) or BOB.
8095 ;;
8096 ;; The criterion for a paren structure being the arg list is:
8097 ;; o - there is non-WS stuff after it but before any "{"; AND
8098 ;; o - the token after it isn't a ";" AND
8099 ;; o - it is preceded by either an identifier (the function name) or
8100 ;; a macro expansion like "DEFUN (...)"; AND
8101 ;; o - its content is a non-empty comma-separated list of identifiers
8102 ;; (an empty arg list won't have a knr region).
8103 ;;
8104 ;; The following snippet illustrates these rules:
8105 ;; int foo (bar, baz, yuk)
8106 ;; int bar [] ;
8107 ;; int (*baz) (my_type) ;
8108 ;; int (*) (void) (*yuk) (void) ;
8109 ;; {
8120 ; parens/brackets.
8124 ;; We're inside a paren. Could it be our argument list....?
8130 ;; FIXME!!! What about macros between the parens? 2007/01/20
8133 ;; It can't be the arg list if next token is ; or {
8138 ;; Is the thing preceding the list an identifier (the
8139 ;; function name), or a macro expansion?
8149 ;; Have we got a non-empty list of comma-separated
8150 ;; identifiers?
8164 ;; ...Yes. We've identified the function's argument list.
8170 ;; ...No. The current parens aren't the function's arg list.
8177 ;; skip forward over conditional at point, including any predicate
8178 ;; statements in parentheses. No error checking is performed.
8179 ;;
8180 ;; This function might do hidden buffer changes.
8182 ;; else if()
8187 ;; do, else, try, finally
8192 ;; for, if, while, switch, catch, synchronized, foreach
8196 ;; If looking at the token after a conditional then return the
8197 ;; position of its start, otherwise return nil.
8198 ;;
8199 ;; This function might do hidden buffer changes.
8210 ;; If the point is after an operator identifier that isn't handled
8211 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
8212 ;; position of the start of that identifier is returned. nil is
8213 ;; returned otherwise. The point may be anywhere in the syntactic
8214 ;; whitespace after the last token of the operator identifier.
8215 ;;
8216 ;; This function might do hidden buffer changes.
8228 ;; Assuming point is at a brace that opens a statement block of some
8229 ;; kind, move to the proper anchor point for that block. It might
8230 ;; need to be adjusted further by c-add-stmt-syntax, but the
8231 ;; position at return is suitable as start position for that
8232 ;; function.
8233 ;;
8234 ;; This function might do hidden buffer changes.
8241 ;; Assuming point is at a brace that opens the block of a top level
8242 ;; declaration of some kind, move to the proper anchor point for
8243 ;; that block.
8244 ;;
8245 ;; This function might do hidden buffer changes.
8250 ;; Search forward for the end of the "header" of the current
8251 ;; declaration. That's the position where the definition body
8252 ;; starts, or the first variable initializer, or the ending
8253 ;; semicolon. I.e. search forward for the closest following
8254 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
8255 ;; _after_ the first found token, or at point-max if none is found.
8256 ;;
8257 ;; This function might do hidden buffer changes.
8262 ;; In C++ we need to take special care to handle operator
8263 ;; tokens and those pesky template brackets.
8268 ;; Handle operator identifiers, i.e. ignore any
8269 ;; operator token preceded by "operator".
8276 t
8278 nil)))))
8287 ;; Go to the beginning of the current declaration, or the beginning
8288 ;; of the previous one if already at the start of it. Point won't
8289 ;; be moved out of any surrounding paren. Return a cons cell of the
8290 ;; form (MOVE . KNR-POS). MOVE is like the return value from
8291 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
8292 ;; style argument declarations (and they are to be recognized) then
8293 ;; KNR-POS is set to the start of the first such argument
8294 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
8295 ;; position that bounds the backward search.
8296 ;;
8297 ;; NB: Cases where the declaration continues after the block, as in
8298 ;; "struct foo { ... } bar;", are currently recognized as two
8299 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
8300 ;;
8301 ;; This function might do hidden buffer changes.
8307 ;; `c-beginning-of-statement-1' stops at a block start, but we
8308 ;; want to continue if the block doesn't begin a top level
8309 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8310 ;; or an open paren.
8312 ;; Go back one "statement" each time round the loop until we're just
8313 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8314 ;; an ObjC method. This will move over a multiple declaration whose
8315 ;; components are comma separated.
8317 ;; Must check with c-opt-method-key in ObjC mode.
8327 ;; Check that we don't move from the first thing in a
8328 ;; macro to its header.
8334 move tentative-move))
8339 ;; Handle K&R argdecls. Back up after the "statement" jumped
8340 ;; over by `c-beginning-of-statement-1', unless it was the
8341 ;; function body, in which case we're sitting on the opening
8342 ;; brace now. Then test if we're in a K&R argdecl region and
8343 ;; that we started at the other side of the first argdecl in
8344 ;; it.
8354 'previous
8356 knr-argdecl-start))
8359 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8360 ;; statement, so the result will be 'previous if we've moved over any.
8361 ;; So change our result back to 'same if necessary.
8362 ;;
8363 ;; If they were brace list initializers we might not have moved over a
8364 ;; declaration boundary though, so change it to 'same if we've moved
8365 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8366 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8367 ;; potentially can search over a large amount of text.). Take special
8368 ;; pains not to get mislead by C++'s "operator=", and the like.
8371 c++-template-syntax-table
8377 ; no more, or get to one which isn't an "operator ="
8380 c-overloadable-operators-regexp
8381 c-opt-op-identifier-prefix
8396 ;; Assuming point is at the start of a declaration (as detected by
8397 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8398 ;; `c-beginning-of-decl-1', this function handles the case when a
8399 ;; block is followed by identifiers in e.g. struct declarations in C
8400 ;; or C++. If a proper end was found then t is returned, otherwise
8401 ;; point is moved as far as possible within the current sexp and nil
8402 ;; is returned. This function doesn't handle macros; use
8403 ;; `c-end-of-macro' instead in those cases.
8404 ;;
8405 ;; This function might do hidden buffer changes.
8408 c++-template-syntax-table
8416 ;; Stopped at the ';' in a K&R argdecl section which is
8417 ;; detected using the same criteria as in
8418 ;; `c-beginning-of-decl-1'. Move to the following block
8419 ;; start.
8423 ;; Encountered a block in the declaration. Jump over it.
8434 ;; Check for `c-opt-block-decls-with-vars-key'
8435 ;; before the first paren.
8438 c-opt-block-decls-with-vars-key
8440 lim t t t)
8443 ;; Check that the first following paren is
8444 ;; the block.
8446 lim t t t)
8448 ;; The declaration doesn't have any of the
8449 ;; `c-opt-block-decls-with-vars' keywords in the
8450 ;; beginning, so it ends here at the end of the block.
8458 nil)))
8461 ;; Assuming the point is at an open brace, check if it starts a
8462 ;; block that contains another declaration level, i.e. that isn't a
8463 ;; statement block or a brace list, and if so return non-nil.
8464 ;;
8465 ;; If the check is successful, the return value is the start of the
8466 ;; keyword that tells what kind of construct it is, i.e. typically
8467 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8468 ;; the point will be at the start of the construct, before any
8469 ;; leading specifiers, otherwise it's at the returned position.
8470 ;;
8471 ;; The point is clobbered if the check is unsuccessful.
8472 ;;
8473 ;; CONTAINING-SEXP is the position of the open of the surrounding
8474 ;; paren, or nil if none.
8475 ;;
8476 ;; The optional LIMIT limits the backward search for the start of
8477 ;; the construct. It's assumed to be at a syntactically relevant
8478 ;; position.
8479 ;;
8480 ;; If any template arglists are found in the searched region before
8481 ;; the open brace, they get marked with paren syntax.
8482 ;;
8483 ;; This function might do hidden buffer changes.
8490 ;; Could be at the end of a template arglist.
8498 ;; Note: Can't get bogus hits inside template arglists below since they
8499 ;; have gotten paren syntax above.
8501 ;; If `goto-start' is set we begin by searching for the
8502 ;; first possible position of a leading specifier list.
8503 ;; The `c-decl-block-key' search continues from there since
8504 ;; we know it can't match earlier.
8507 open-brace t t)
8509 t)
8510 t)
8516 ;; Exclude cases where we matched what would ordinarily
8517 ;; be a block declaration keyword, except where it's not
8518 ;; legal because it's part of a "compound keyword" like
8519 ;; "enum class". Of course, if c-after-brace-list-key
8520 ;; is nil, we can skip the test.
8530 ;; Found a keyword that can't be a type?
8533 ;; Can be a type too, in which case it's the return type of a
8534 ;; function (under the assumption that no declaration level
8535 ;; block construct starts with a type).
8538 ;; Jumped over a type, but it could be a declaration keyword
8539 ;; followed by the declared identifier that we've jumped over
8540 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8541 ;; then we should be at the declarator now, so check for a
8542 ;; valid declarator start.
8543 ;;
8544 ;; Note: This doesn't cope with the case when a declared
8545 ;; identifier is followed by e.g. '(' in a language where '('
8546 ;; also might be part of a declarator expression. Currently
8547 ;; there's no such language.
8551 ;; In Pike a list of modifiers may be followed by a brace
8552 ;; to make them apply to many identifiers. Note that the
8553 ;; match data will be empty on return in this case.
8559 ;; Use this variant to avoid yet another special regexp.
8563 t)))
8565 ;; Got a match.
8568 ;; Back up over any preceding specifiers and their clauses
8569 ;; by going forward from `first-specifier-pos', which is the
8570 ;; earliest possible position where the specifier list can
8571 ;; start.
8577 ;; Accept any plain symbol token on the ground that
8578 ;; it's a specifier masked through a macro (just
8579 ;; like `c-forward-decl-or-cast-1' skip forward over
8580 ;; such tokens).
8581 ;;
8582 ;; Could be more restrictive wrt invalid keywords,
8583 ;; but that'd only occur in invalid code so there's
8584 ;; no use spending effort on it.
8590 ;; Can't parse a declaration preamble and is still
8591 ;; before `kwd-start'. That means `first-specifier-pos'
8592 ;; was in some earlier construct. Search again.
8596 ;; Got no preamble before the block declaration keyword.
8602 kwd-start)))
8605 ;; Check if the closest containing paren sexp is a declaration
8606 ;; block, returning a 2 element vector in that case. Aref 0
8607 ;; contains the bufpos at boi of the class key line, and aref 1
8608 ;; contains the bufpos of the open brace. This function is an
8609 ;; obsolete wrapper for `c-looking-at-decl-block'.
8610 ;;
8611 ;; This function might do hidden buffer changes.
8619 nil))
8624 ;; Return the buffer position of the most enclosing decl-block brace (in the
8625 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8626 ;; none was found.
8635 open-brace))
8638 ;; Return the position of the L-brace if point is inside a brace list
8639 ;; initialization of an array, etc. This is an approximate function,
8640 ;; designed for speed over accuracy. It will not find every bracelist, but
8641 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8642 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8643 ;; is everywhere else.
8652 b-pos)))
8655 ;; We're at the token after what might be a type prefixed with a colon. Try
8656 ;; moving backward over this type and the colon. On success, return t and
8657 ;; leave point before colon; on failure, leave point unchanged. Will clobber
8658 ;; match data.
8675 nil)
8677 t)
8679 t)
8683 t)))
8686 ;; We're at a "{". Move back to the enum-like keyword that starts this
8687 ;; declaration and return t, otherwise don't move and return nil.
8689 up-sexp-pos before-identifier)
8705 t)
8711 t)
8717 ;; return the buffer position of the beginning of the brace list
8718 ;; statement if we're inside a brace list, otherwise return nil.
8719 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8720 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8721 ;; braces
8722 ;;
8723 ;; N.B.: This algorithm can potentially get confused by cpp macros
8724 ;; placed in inconvenient locations. It's a trade-off we make for
8725 ;; speed.
8726 ;;
8727 ;; This function might do hidden buffer changes.
8729 ;; This will pick up brace list declarations.
8733 ;; this will pick up array/aggregate init lists, even if they are nested.
8736 ;; Pike can have class definitions anywhere, so we must
8737 ;; check for the class key here.
8739 c-decl-block-key))
8740 bufpos braceassignp lim next-containing macro-start)
8742 containing-sexp)
8753 ;; We're in an in-expression block of some kind. Do not
8754 ;; check nesting. We deliberately set the limit to the
8755 ;; containing sexp, so that c-looking-at-inexpr-block
8756 ;; doesn't check for an identifier before it.
8758 ;; see if the open brace is preceded by = or [...] in
8759 ;; this statement, but watch out for operator=
8762 ;; Checks to do only on the first sexp before the brace.
8765 ;; In Java, an initialization brace list may follow
8766 ;; directly after "new Foo[]", so check for a "new"
8767 ;; earlier.
8773 ;; Carry on looking if this is an
8774 ;; identifier (may contain "." in Java)
8775 ;; or another "[]" sexp.
8778 ;; Checks to do on all sexps before the brace, up to the
8779 ;; beginning of the statement.
8787 ;; We've seen a =, but must check earlier tokens so
8788 ;; that it isn't something that should be ignored.
8794 ;; Check for operator =
8797 nil)
8798 ;; Check for `<opchar>= in Pike.
8801 ;; Special case for Pikes
8802 ;; `[]=, since '[' is not in
8803 ;; the punctuation class.
8806 nil)
8808 ;; make sure we're not in a C++ template
8809 ;; argument assignment
8821 ))))
8822 nil)
8829 ;; We've hit the beginning of the aggregate list.
8834 ;; Brace lists can't contain a semicolon, so we're done.
8839 ;; We've a macro whose expansion starts with the '{'.
8840 ;; Heuristically, if we have a ';' in it we've not got a
8841 ;; brace list, otherwise we have.
8848 containing-sexp nil)
8851 ;; Go up one level
8853 lim nil
8854 next-containing nil)))))
8856 bufpos))
8857 ))
8860 ;; If we're looking at the start of a pike-style list, i.e., `({Â })',
8861 ;; `([Â ])', `(<Â >)', etc., a cons of a cons of its starting and ending
8862 ;; positions and its entry in c-special-brace-lists is returned, nil
8863 ;; otherwise. The ending position is nil if the list is still open.
8864 ;; LIM is the limit for forward search. The point may either be at
8865 ;; the `(' or at the following paren character. Tries to check the
8866 ;; matching closer, but assumes it's correct if no balanced paren is
8867 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8868 ;; a special brace list).
8869 ;;
8870 ;; This function might do hidden buffer changes.
8875 inner-beg end type)
8890 nil)))))
8900 ;; Check balancing of the inner paren
8901 ;; below.
8904 t)
8905 ;; If the inner char isn't a paren then
8906 ;; we can't check balancing, so just
8907 ;; check the char before the outer
8908 ;; closing paren.
8923 ;; Return non-nil if between two statements or declarations, assuming
8924 ;; point is not inside a literal or comment.
8925 ;;
8926 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8927 ;; are recommended instead.
8928 ;;
8929 ;; This function might do hidden buffer changes.
8934 ;; Return non-nil if we're looking at the beginning of a block
8935 ;; inside an expression. The value returned is actually a cons of
8936 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8937 ;; position of the beginning of the construct.
8938 ;;
8939 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8940 ;; position of the closest containing list. If it's nil, the
8941 ;; containing paren isn't used to decide whether we're inside an
8942 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8943 ;; needs to be farther back.
8944 ;;
8945 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8946 ;; brace block might be done. It should only be used when the
8947 ;; construct can be assumed to be complete, i.e. when the original
8948 ;; starting position was further down than that.
8949 ;;
8950 ;; This function might do hidden buffer changes.
8955 ;; Look at the character after point only as a last resort
8956 ;; when we can't disambiguate.
8976 ;; If the class definition is at the start of
8977 ;; a statement, we don't consider it an
8978 ;; in-expression class.
8986 ;; Also, in Pike we treat it as an
8987 ;; in-expression class if it's used in an
8988 ;; object clone expression.
9004 nil)
9012 ;; Accept several square bracket sexps for
9013 ;; Java array initializations.
9021 block-follows
9022 containing-sexp
9031 nil
9034 res))))
9037 ;; Returns non-nil if we're looking at the end of an in-expression
9038 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
9039 ;; PAREN-STATE is the paren state relevant at the current position.
9040 ;;
9041 ;; This function might do hidden buffer changes.
9043 ;; We currently only recognize a block.
9046 containing-sexp)
9055 paren-state)
9056 containing-sexp)))))
9059 ;; Is there a "virtual semicolon" at POS or point?
9060 ;; (See cc-defs.el for full details of "virtual semicolons".)
9061 ;;
9062 ;; This is true when point is at the last non syntactic WS position on the
9063 ;; line, there is a macro call last on the line, and this particular macro's
9064 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
9065 ;; semicolon.
9073 c-macro-with-semi-re
9076 ;; Check we've got nothing after this except comments and empty lines
9077 ;; joined by escaped EOLs.
9090 nil)
9098 ;; Check for one of the listed macros being before point.
9102 t))
9110 \f
9111 ;; `c-guess-basic-syntax' and the functions that precedes it below
9112 ;; implements the main decision tree for determining the syntactic
9113 ;; analysis of the current line of code.
9115 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
9116 ;; auto newline analysis.
9120 ;; BRACEPOS is the position of a brace in a construct like "namespace
9121 ;; Bar {". Return the anchor point in this construct; this is the
9122 ;; earliest symbol on the brace's line which isn't earlier than
9123 ;; "namespace".
9124 ;;
9125 ;; Currently (2007-08-17), "like namespace" means "matches
9126 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
9127 ;; or anything like that.
9137 ;; A simple function to prepend a new syntax element to
9138 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
9139 ;; should always be dynamically bound but since we read it first
9140 ;; we'll fail properly anyway if this function is misused.
9142 c-syntactic-context)))
9145 ;; Like `c-add-syntax' but appends to the end of the syntax list.
9146 ;; (Normally not necessary.)
9151 syntax-extra-args
9152 stop-at-boi-only
9153 containing-sexp
9154 paren-state)
9155 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
9156 ;; needed with further syntax elements of the types `substatement',
9157 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
9158 ;; `defun-block-intro'.
9159 ;;
9160 ;; Do the generic processing to anchor the given syntax symbol on
9161 ;; the preceding statement: Skip over any labels and containing
9162 ;; statements on the same line, and then search backward until we
9163 ;; find a statement or block start that begins at boi without a
9164 ;; label or comment.
9165 ;;
9166 ;; Point is assumed to be at the prospective anchor point for the
9167 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
9168 ;; skip past open parens and containing statements. Most of the added
9169 ;; syntax elements will get the same anchor point - the exception is
9170 ;; for an anchor in a construct like "namespace"[*] - this is as early
9171 ;; as possible in the construct but on the same line as the {.
9172 ;;
9173 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
9174 ;;
9175 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
9176 ;; syntax symbol. They are appended after the anchor point.
9177 ;;
9178 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
9179 ;; if the current statement starts there.
9180 ;;
9181 ;; Note: It's not a problem if PAREN-STATE "overshoots"
9182 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
9183 ;;
9184 ;; This function might do hidden buffer changes.
9187 ;; This is by far the most common case, so let's give it special
9188 ;; treatment.
9193 ;; Set when we're on a label, so that we don't stop there.
9194 ;; FIXME: To be complete we should check if we're on a label
9195 ;; now at the start.
9196 on-label)
9198 ;; Use point as the anchor point for "namespace", "extern", etc.
9202 syntax-extra-args)
9204 ;; Loop while we have to back out of containing blocks.
9209 ;; Loop while we have to back up statements.
9211 on-label
9214 ;; Skip past any comments that stands between the
9215 ;; statement start and boi.
9223 ;; Skip to the beginning of this statement or backward
9224 ;; another one.
9232 ;; If we didn't move we're at the start of a block and
9233 ;; have to continue outside it.
9242 ;; Special case to avoid deeper and deeper indentation
9243 ;; of "else if" clauses.
9244 )
9249 ;; If stop-at-boi-only is nil, we shouldn't back up
9250 ;; over previous or containing statements to try to
9251 ;; reach boi, so go back to the last position and
9252 ;; exit.
9259 ;; If we've moved into another statement then we
9260 ;; should no longer try to stop in the middle of a
9261 ;; line.
9264 ;; Record this as a substatement if we skipped up one
9265 ;; level.
9268 )))
9270 containing-sexp)
9272 ;; Now we have to go out of this block.
9275 ;; Don't stop in the middle of a special brace list opener
9276 ;; like "({".
9288 ;; Analyze the construct in front of the block we've stepped out
9289 ;; from and add the right syntactic element for it.
9292 step-type)
9295 ;; Stepped out of a parenthesis block, so we're in an
9296 ;; expression now.
9308 ;; Stepped out of an in-expression statement. This
9309 ;; syntactic element won't get an anchor pos.
9312 ;; A parenthesis normally belongs to an arglist.
9320 on-label nil))
9322 ;; Stepped out of a brace block.
9334 containing-sexp)))
9336 'defun-block-intro
9338 nil))
9342 c-other-decl-block-key-in-symbols-alist))
9349 ;; Always done if the open brace was at boi. The
9350 ;; c-beginning-of-statement-1 call above is necessary
9351 ;; anyway, to decide the type of block-intro to add.
9354 ))
9356 ;; Fill in the current point as the anchor for all the symbols
9357 ;; added above.
9366 )))
9369 containing-decl-open
9370 containing-decl-start
9371 containing-decl-kwd
9372 paren-state)
9373 ;; The inclass and class-close syntactic symbols are added in
9374 ;; several places and some work is needed to fix everything.
9375 ;; Therefore it's collected here.
9376 ;;
9377 ;; This function might do hidden buffer changes.
9382 containing-decl-open)
9384 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9385 ;; here, but we have to do like this for compatibility.
9395 char-after-ip
9396 beg-of-same-or-containing-stmt
9397 containing-sexp
9398 paren-state)
9399 ;; This function contains the decision tree reached through both
9400 ;; cases 18 and 10. It's a continued statement or top level
9401 ;; construct of some kind.
9402 ;;
9403 ;; This function might do hidden buffer changes.
9410 ;; (CASE A removed.)
9411 ;; CASE B: open braces for class or brace-lists
9418 ;; CASE B.1: class-open
9425 ;; CASE B.2: brace-list-open
9430 indent-point t t t)))
9431 ;; The most semantically accurate symbol here is
9432 ;; brace-list-open, but we normally report it simply as a
9433 ;; statement-cont. The reason is that one normally adjusts
9434 ;; brace-list-open for brace lists as top-level constructs,
9435 ;; and brace lists inside statements is a completely different
9436 ;; context. C.f. case 5A.3.
9439 ;; Turn off the dwim above when we're
9440 ;; analyzing the nature of the brace
9441 ;; for the auto newline feature.
9442 'brace-list-open
9444 nil nil
9445 containing-sexp paren-state))
9447 ;; CASE B.3: The body of a function declared inside a normal
9448 ;; block. Can occur e.g. in Pike and when using gcc
9449 ;; extensions, but watch out for macros followed by blocks.
9450 ;; C.f. cases E, 16F and 17G.
9456 ;; Turn off recognition of constructs that lacks a
9457 ;; type in this case, since that's more likely to be
9458 ;; a macro followed by a block.
9461 containing-sexp paren-state))
9463 ;; CASE B.4: Continued statement with block open. The most
9464 ;; accurate analysis is perhaps `statement-cont' together with
9465 ;; `block-open' but we play DWIM and use `substatement-open'
9466 ;; instead. The rationale is that this typically is a macro
9467 ;; followed by a block which makes it very similar to a
9468 ;; statement with a substatement block.
9471 containing-sexp paren-state))
9472 ))
9474 ;; CASE C: iostream insertion or extraction operator
9478 ;; If there is no preceding streamop in the statement
9479 ;; then indent this line as a normal statement-cont.
9483 t))))
9485 ;; CASE E: In the "K&R region" of a function declared inside a
9486 ;; normal block. C.f. case B.3.
9488 ;; Check that the next token is a '{'. This works as
9489 ;; long as no language that allows nested function
9490 ;; definitions allows stuff like member init lists, K&R
9491 ;; declarations or throws clauses there.
9492 ;;
9493 ;; Note that we do a forward search for something ahead
9494 ;; of the indentation line here. That's not good since
9495 ;; the user might not have typed it yet. Unfortunately
9496 ;; it's exceedingly tricky to recognize a function
9497 ;; prototype in a code block without resorting to this.
9505 ;; Turn off recognition of constructs that lacks a
9506 ;; type in this case, since that's more likely to be
9507 ;; a macro followed by a block.
9510 containing-sexp paren-state))
9512 ;;CASE F: continued statement and the only preceding items are
9513 ;;annotations.
9521 t)
9528 ;; CASE G: a template list continuation?
9529 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9547 ;; In a nested template arglist.
9553 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9554 ;; template aware.
9557 ;; CASE D: continued statement.
9561 containing-sexp paren-state))
9562 )))
9564 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9565 ;; 2005/11/29).
9566 ;;;###autoload
9568 "Return the syntactic context of the current line."
9574 open-paren-in-column-0-is-defun-start
9575 ;; A whole ugly bunch of various temporary variables. Have
9576 ;; to declare them here since it's not possible to declare
9577 ;; a variable with only the scope of a cond test and the
9578 ;; following result clauses, and most of this function is a
9579 ;; single gigantic cond. :P
9580 literal char-before-ip before-ws-ip char-after-ip macro-start
9581 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9582 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9583 containing-<
9584 ;; The following record some positions for the containing
9585 ;; declaration block if we're directly within one:
9586 ;; `containing-decl-open' is the position of the open
9587 ;; brace. `containing-decl-start' is the start of the
9588 ;; declaration. `containing-decl-kwd' is the keyword
9589 ;; symbol of the keyword that tells what kind of block it
9590 ;; is.
9591 containing-decl-open
9592 containing-decl-start
9593 containing-decl-kwd
9594 ;; The open paren of the closest surrounding sexp or nil if
9595 ;; there is none.
9596 containing-sexp
9597 ;; The position after the closest preceding brace sexp
9598 ;; (nested sexps are ignored), or the position after
9599 ;; `containing-sexp' if there is none, or (point-min) if
9600 ;; `containing-sexp' is nil.
9601 lim
9602 ;; The paren state outside `containing-sexp', or at
9603 ;; `indent-point' if `containing-sexp' is nil.
9605 ;; There's always at most one syntactic element which got
9606 ;; an anchor pos. It's stored in syntactic-relpos.
9607 syntactic-relpos
9610 ;; Check if we're directly inside an enclosing declaration
9611 ;; level block.
9620 containing-sexp)
9621 t)))
9624 containing-sexp nil)
9629 ;; Init some position variables.
9642 ;; Ignore balanced paren. The next entry
9643 ;; can't be another one.
9646 ;; If there is no surrounding open paren then
9647 ;; put the last balanced pair back on paren-state.
9649 containing-sexp nil)))
9653 ;; If we're in a parenthesis list then ',' delimits the
9654 ;; "statements" rather than being an operator (with the
9655 ;; exception of the "for" clause). This difference is
9656 ;; typically only noticeable when statements are used in macro
9657 ;; arglists.
9661 ;; cache char before and after indent point, and move point to
9662 ;; the most likely position to perform the majority of tests
9671 ;; are we in a literal?
9674 ;; now figure out syntactic qualities of the current line
9677 ;; CASE 1: in a string.
9681 ;; CASE 2: in a C or C++ style comment.
9683 ;; This is a kludge for XEmacs where we use
9684 ;; `buffer-syntactic-context', which doesn't correctly
9685 ;; recognize "\*/" to end a block comment.
9686 ;; `parse-partial-sexp' which is used by
9687 ;; `c-literal-limits' will however do that in most
9688 ;; versions, which results in that we get nil from
9689 ;; `c-literal-limits' even when `c-in-literal' claims
9690 ;; we're inside a comment.
9694 ;; CASE 3: in a cpp preprocessor macro continuation.
9704 ;; If at the beginning of the body of a #define
9705 ;; directive then analyze as cpp-define-intro
9706 ;; only. Go on with the syntactic analysis
9707 ;; otherwise. in-macro-expr is set if we're in a
9708 ;; cpp expression, i.e. before the #define body
9709 ;; or anywhere in a non-#define directive.
9716 nil)))))
9720 ;; CASE 11: an else clause?
9724 containing-sexp paren-state))
9726 ;; CASE 12: while closure of a do/while construct?
9734 containing-sexp paren-state))
9736 ;; CASE 13: A catch or finally clause? This case is simpler
9737 ;; than if-else and do-while, because a block is required
9738 ;; after every try, catch and finally.
9746 t)
9750 t)
9753 t))
9758 containing-sexp paren-state))
9760 ;; CASE 18: A substatement we can recognize by keyword.
9770 ;; Ought to cache the result from the
9771 ;; c-beginning-of-statement-1 calls here.
9781 ;; Step up to the containing statement if we
9782 ;; stayed in the same one.
9790 ;; There was no containing statement after all.
9792 placeholder))
9794 ;; Require a parenthesis after these keywords.
9795 ;; Necessary to catch e.g. synchronized in Java,
9796 ;; which can be used both as statement and
9797 ;; modifier.
9803 ;; CASE 18A: Simple substatement.
9809 containing-sexp paren-state))
9815 containing-sexp paren-state))
9818 containing-sexp paren-state))))
9820 ;; CASE 18B: Some other substatement. This is shared
9821 ;; with case 10.
9823 char-after-ip
9824 placeholder
9825 lim
9826 paren-state)))
9828 ;; CASE 14: A case or default label
9834 containing-sexp))
9837 ;; Got a bogus label at the top level. In lack of better
9838 ;; alternatives, anchor it on (point-min).
9841 ;; CASE 15: any other label
9848 containing-decl-open
9849 containing-decl-start
9850 containing-decl-kwd
9851 paren-state))
9852 ;; Append access-label with the same anchor point as
9853 ;; inclass gets.
9859 containing-sexp))
9864 ;; If the surrounding statement is a switch then
9865 ;; let's analyze all labels as switch labels, so
9866 ;; that they get lined up consistently.
9867 'case-label
9873 ;; A label on the top level. Treat it as a class
9874 ;; context. (point-min) is the closest we get to the
9875 ;; class open brace.
9878 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9879 ;; 17E.
9882 containing-sexp
9883 ;; Have to turn on the heuristics after
9884 ;; the point even though it doesn't work
9885 ;; very well. C.f. test case class-16.pike.
9886 t))
9891 ;; It's a statement block or an anonymous class.
9893 ;; It's a Pike lambda. Check whether we are between the
9894 ;; lambda keyword and the argument list or at the defun
9895 ;; opener.
9897 'inline-open
9903 paren-state)
9907 ;; CASE 5: Line is inside a declaration level block or at top level.
9911 ;; CASE 5A: we are looking at a defun, brace list, class,
9912 ;; or inline-inclass method opening brace
9919 ;; CASE 5A.1: Non-class declaration block open.
9934 ;; Special case for extern-lang-open.
9935 'extern-lang-open
9937 nil t containing-sexp paren-state))
9939 ;; CASE 5A.2: we are looking at a class opening brace
9948 ;; CASE 5A.3: brace list open
9973 ))))
9977 ;; We're in Java and have found that the open brace
9978 ;; belongs to a "new Foo[]" initialization list,
9979 ;; which means the brace list is part of an
9980 ;; expression and not a top level definition. We
9981 ;; therefore treat it as any topmost continuation
9982 ;; even though the semantically correct symbol still
9983 ;; is brace-list-open, on the same grounds as in
9984 ;; case B.2.
9990 ;; CASE 5A.4: inline defun open
9996 containing-decl-open
9997 containing-decl-start
9998 containing-decl-kwd
9999 paren-state))
10001 ;; CASE 5A.5: ordinary defun open
10009 ;; Bogus to use bol here, but it's the legacy. (Resolved,
10010 ;; 2007-11-09)
10011 ))))
10013 ;; CASE 5R: Member init list. (Used to be part of CASE 5B.1)
10014 ;; Note there is no limit on the backward search here, since member
10015 ;; init lists can, in practice, be very large.
10021 ;; There is no preceding member init clause.
10022 ;; Indent relative to the beginning of indentation
10023 ;; for the topmost-intro line that contains the
10024 ;; prototype's open paren.
10027 ;; Indent relative to the first member init clause.
10032 ;; CASE 5B: After a function header but before the body (or
10033 ;; the ending semicolon if there's no body).
10040 ;; CASE 5B.1: Member init list.
10042 ;; There is no preceding member init clause.
10043 ;; Indent relative to the beginning of indentation
10044 ;; for the topmost-intro line that contains the
10045 ;; prototype's open paren.
10049 ;; CASE 5B.2: K&R arg decl intro
10056 containing-decl-open
10057 containing-decl-start
10058 containing-decl-kwd
10059 paren-state)))
10061 ;; CASE 5B.4: Nether region after a C++ or Java func
10062 ;; decl, which could include a `throws' declaration.
10066 )))
10068 ;; CASE 5C: inheritance line. could be first inheritance
10069 ;; line, or continuation of a multiple inheritance
10077 ;; watch out for scope operator
10082 )))
10091 ;; for Java
10096 cont done)
10101 done t))
10105 )
10107 injava-inher)
10110 ))
10113 ;; CASE 5C.1: non-hanging colon on an inher intro
10117 ;; don't add inclass symbol since relative point already
10118 ;; contains any class offset
10119 )
10121 ;; CASE 5C.2: hanging colon on an inher intro
10127 containing-decl-open
10128 containing-decl-start
10129 containing-decl-kwd
10130 paren-state)))
10132 ;; CASE 5C.3: in a Java implements/extends
10146 )))
10148 ;; CASE 5C.4: a continued inheritance line
10152 ;; don't add inclass symbol since relative point already
10153 ;; contains any class offset
10154 )))
10156 ;; CASE 5P: AWK pattern or function or continuation
10157 ;; thereof.
10163 'topmost-intro-cont
10165 nil nil
10166 containing-sexp paren-state))
10168 ;; CASE 5D: this could be a top-level initialization, a
10169 ;; member init list continuation, or a template argument
10170 ;; list continuation.
10172 ;; Note: We use the fact that lim is always after any
10173 ;; preceding brace sexp.
10183 t))
10188 t)
10194 nil)
10195 t))
10197 ;; NB: No c-after-special-operator-id stuff in this
10198 ;; clause - we assume only C++ needs it.
10203 ;; CASE 5D.3: perhaps a template list continuation?
10218 ;; In a nested template arglist.
10224 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
10225 ;; template aware.
10228 ;; CASE 5D.4: perhaps a multiple inheritance line?
10240 t)
10245 ;; CASE 5D.5: Continuation of the "expression part" of a
10246 ;; top level construct. Or, perhaps, an unrecognized construct.
10258 ;; A preceding comma at the top level means that a
10259 ;; new variable declaration starts here. Use
10260 ;; topmost-intro-cont for it, for consistency with
10261 ;; the first variable declaration. C.f. case 5N.
10264 nil nil containing-sexp paren-state))
10265 ))
10267 ;; CASE 5F: Close of a non-class declaration level block.
10271 ;; This is inconsistent: Should use `containing-decl-open'
10272 ;; here if it's at boi, like in case 5J.
10276 ;; Special case for compatibility with the
10277 ;; extern-lang syntactic symbols.
10278 'extern-lang-close
10281 nil t
10283 paren-state))
10285 ;; CASE 5G: we are looking at the brace which closes the
10286 ;; enclosing nested class decl
10291 containing-decl-open
10292 containing-decl-start
10293 containing-decl-kwd
10294 paren-state))
10296 ;; CASE 5H: we could be looking at subsequent knr-argdecls
10303 ;; Do an extra check to avoid tripping up on
10304 ;; statements that occur in invalid contexts
10305 ;; (e.g. in macro bodies where we don't really
10306 ;; know the context of what we're looking at).
10313 ;; CASE 5I: ObjC method definition.
10318 ;; Handle the case when it's the first (non-comment)
10319 ;; thing in the buffer. Can't look for a 'same return
10320 ;; value from cbos1 since ObjC directives currently
10321 ;; aren't recognized fully, so that we get 'same
10322 ;; instead of 'previous if it moved over a preceding
10323 ;; directive.
10327 ;; CASE 5N: At a variable declaration that follows a class
10328 ;; definition or some other block declaration that doesn't
10329 ;; end at the closing '}'. C.f. case 5D.5.
10338 ;; Speed up the backward search a bit.
10343 ;; The '}' is unbalanced.
10344 nil
10349 containing-sexp paren-state))
10351 ;; NOTE: The point is at the end of the previous token here.
10353 ;; CASE 5J: we are at the topmost level, make
10354 ;; sure we skip back past any access specifiers
10356 ;; A macro continuation line is never at top level.
10378 nil)))))
10379 ;; For historic reasons we anchor at bol of the last
10380 ;; line of the previous declaration. That's clearly
10381 ;; highly bogus and useless, and it makes our lives hard
10382 ;; to remain compatible. :P
10393 ;; Special case for compatibility with the
10394 ;; extern-lang syntactic symbols.
10395 'inextern-lang
10398 nil t
10400 paren-state))
10402 containing-decl-open
10403 containing-decl-start
10404 containing-decl-kwd
10405 paren-state)))
10407 macro-start
10412 ;; CASE 5K: we are at an ObjC method definition
10413 ;; continuation line.
10422 ;; CASE 5L: we are at the first argument of a template
10423 ;; arglist that begins on the previous line.
10431 ;; CASE 5Q: we are at a statement within a macro.
10436 ;;CASE 5N: We are at a topmost continuation line and the only
10437 ;;preceding items are annotations.
10444 t)
10450 ;; CASE 5M: we are at a topmost continuation line
10461 ))
10463 ;; (CASE 6 has been removed.)
10465 ;; CASE 7: line is an expression, not a statement. Most
10466 ;; likely we are either in a function prototype or a function
10467 ;; call argument list
10475 ;; CASE 7A: we are looking at the arglist closing paren.
10476 ;; C.f. case 7F.
10488 paren-state))
10490 ;; CASE 7B: Looking at the opening brace of an
10491 ;; in-expression block or brace list. C.f. cases 4, 16A
10492 ;; and 17E.
10496 paren-state))
10500 placeholder
10503 paren-state)
10504 containing-sexp)))
10505 ;; placeholder is nil if it's a block directly in
10506 ;; a function arglist. That makes us skip out of
10507 ;; this case.
10508 )))
10513 paren-state)
10517 ;; CASE 7C: we are looking at the first argument in an empty
10518 ;; argument list. Use arglist-close if we're actually
10519 ;; looking at a close paren or bracket.
10531 paren-state))
10533 ;; CASE 7D: we are inside a conditional test clause. treat
10534 ;; these things as statements
10544 ))
10546 ;; CASE 7E: maybe a continued ObjC method call. This is the
10547 ;; case when we are inside a [] bracketed exp, and what
10548 ;; precede the opening bracket is not an identifier.
10556 )))
10558 ;; CASE 7F: we are looking at an arglist continuation line,
10559 ;; but the preceding argument is on the same line as the
10560 ;; opening paren. This case includes multi-line
10561 ;; mathematical paren groupings, but we could be on a
10562 ;; for-list continuation line. C.f. case 7A.
10579 paren-state))
10581 ;; CASE 7G: we are looking at just a normal arglist
10582 ;; continuation line
10585 ))
10587 ;; CASE 8: func-local multi-inheritance line
10597 ;; CASE 8A: non-hanging colon on an inher intro
10602 ;; CASE 8B: hanging colon on an inher intro
10606 ;; CASE 8C: a continued inheritance line
10610 )))
10612 ;; CASE 9: we are inside a brace-list
10622 ;; CASE 9A: In the middle of a special brace list opener.
10642 ;; CASE 9B: brace-list-close brace
10644 ;; Check special brace list closer.
10651 ;; We were between the special close char and the `)'.
10654 ;; We were before the special close char.
10658 ;; Normal brace list check.
10669 ;; Prepare for the rest of the cases below by going to the
10670 ;; token following the opening brace
10683 ;; CASE 9C: we're looking at the first line in a brace-list
10694 ;; CASE 9D: this is just a later brace-list-entry or
10695 ;; brace-entry-open
10704 ))
10705 ))))
10707 ;; CASE 10: A continued statement or top level construct.
10717 ;; This is shared with case 18.
10719 char-after-ip
10720 placeholder
10721 containing-sexp
10722 paren-state))
10724 ;; CASE 16: block close brace, possibly closing the defun or
10725 ;; the class
10727 ;; From here on we have the next containing sexp in lim.
10732 ;; CASE 16E: Closing a statement block? This catches
10733 ;; cases where it's preceded by a statement keyword,
10734 ;; which works even when used in an "invalid" context,
10735 ;; e.g. a macro argument.
10740 ;; CASE 16A: closing a lambda defun or an in-expression
10741 ;; block? C.f. cases 4, 7B and 17E.
10744 nil))
10746 'inline-close
10756 paren-state)
10760 ;; CASE 16B: does this close an inline or a function in
10761 ;; a non-class declaration level block?
10768 nil))
10778 ;; CASE 16F: Can be a defun-close of a function declared
10779 ;; in a statement block, e.g. in Pike or when using gcc
10780 ;; extensions, but watch out for macros followed by
10781 ;; blocks. Let it through to be handled below.
10782 ;; C.f. cases B.3 and 17G.
10788 ;; Turn off recognition of constructs that
10789 ;; lacks a type in this case, since that's more
10790 ;; likely to be a macro followed by a block.
10797 ;; CASE 16C: If there is an enclosing brace then this is
10798 ;; a block close since defun closes inside declaration
10799 ;; level blocks have been handled above.
10801 ;; If the block is preceded by a case/switch label on
10802 ;; the same line, we anchor at the first preceding label
10803 ;; at boi. The default handling in c-add-stmt-syntax
10804 ;; really fixes it better, but we do like this to keep
10805 ;; the indentation compatible with version 5.28 and
10806 ;; earlier. C.f. case 17H.
10813 ;; c-backward-to-block-anchor not necessary here; those
10814 ;; situations are handled in case 16E above.
10817 ;; CASE 16D: Only top level defun close left.
10823 paren-state))
10824 ))
10826 ;; CASE 19: line is an expression, not a statement, and is directly
10827 ;; contained by a template delimiter. Most likely, we are in a
10828 ;; template arglist within a statement. This case is based on CASE
10829 ;; 7. At some point in the future, we may wish to create more
10830 ;; syntactic symbols such as `template-intro',
10831 ;; `template-cont-nonempty', etc., and distinguish between them as we
10832 ;; do for `arglist-intro' etc. (2009-12-07).
10838 ; '<') before indent-point.
10846 paren-state))
10848 ;; CASE 17: Statement or defun catchall.
10851 ;; Back up statements until we find one that starts at boi.
10854 containing-sexp)))
10858 nil)
10863 ;; CASE 17B: continued statement
10867 containing-sexp paren-state))
10869 ;; CASE 17A: After a case/default label?
10877 'statement-case-open
10879 nil t containing-sexp paren-state))
10881 ;; CASE 17D: any old statement
10888 containing-sexp paren-state)
10892 ;; CASE 17I: Inside a substatement block.
10894 ;; The following tests are all based on containing-sexp.
10896 ;; From here on we have the next containing sexp in lim.
10901 lim paren-state)
10905 ;; CASE 17E: first statement in an in-expression block.
10906 ;; C.f. cases 4, 7B and 16A.
10909 nil))
10911 'defun-block-intro
10920 paren-state)
10926 ;; CASE 17F: first statement in an inline, or first
10927 ;; statement in a top-level defun. we can tell this is it
10928 ;; if there are no enclosing braces that haven't been
10929 ;; narrowed out by a class (i.e. don't use bod here).
10932 paren-state)))
10938 nil))))
10943 ;; CASE 17G: First statement in a function declared inside
10944 ;; a normal block. This can occur in Pike and with
10945 ;; e.g. the gcc extensions, but watch out for macros
10946 ;; followed by blocks. C.f. cases B.3 and 16F.
10952 ;; Turn off recognition of constructs that lacks
10953 ;; a type in this case, since that's more likely
10954 ;; to be a macro followed by a block.
10960 lim paren-state))
10962 ;; CASE 17H: First statement in a block.
10964 ;; If the block is preceded by a case/switch label on the
10965 ;; same line, we anchor at the first preceding label at
10966 ;; boi. The default handling in c-add-stmt-syntax is
10967 ;; really fixes it better, but we do like this to keep the
10968 ;; indentation compatible with version 5.28 and earlier.
10969 ;; C.f. case 16C.
10976 ;; c-backward-to-block-anchor not necessary here; those
10977 ;; situations are handled in case 17I above.
10979 lim paren-state))
10982 ))
10983 )
10985 ;; now we need to look at any modifiers
10989 ;; are we looking at a comment only line?
10994 ;; we might want to give additional offset to friends (in C++).
10999 ;; Set syntactic-relpos.
11005 nil)
11006 t))
11009 ;; Start of or a continuation of a preprocessor directive?
11024 ;; If inside a cpp expression, i.e. anywhere in a
11025 ;; cpp directive except a #define body, we only let
11026 ;; through the syntactic analysis that is internal
11027 ;; in the expression. That means the arglist
11028 ;; elements, if they are anchored inside the cpp
11029 ;; expression.
11038 ;; Inside a #define body and the syntactic analysis is
11039 ;; anchored on the start of the #define. In this case
11040 ;; we add cpp-define-intro to get the extra
11041 ;; indentation of the #define body.
11044 ;; return the syntax
11045 c-syntactic-context)))
11047 \f
11048 ;; Indentation calculation.
11051 ;; offset can be a number, a function, a variable, a list, or one of
11052 ;; the symbols + or -
11053 ;;
11054 ;; This function might do hidden buffer changes.
11073 langelem symbol))
11079 offset symbol)
11080 nil)
11094 Error evaluating offset %S for %s: \
11095 Cannot combine absolute offset %S with relative %S in `%s' method"
11101 Error evaluating offset %S for %s: \
11102 Cannot combine relative offset %S with absolute %S in `%s' method"
11107 res))
11125 Error evaluating offset %S for %s: \
11126 Cannot combine absolute offsets %S and %S in `add' method"
11130 res))
11139 res))))
11146 nil))))
11150 res
11152 offset symbol res)
11153 nil)))
11156 ;; Get offset from LANGELEM which is a list beginning with the
11157 ;; syntactic symbol and followed by any analysis data it provides.
11158 ;; That data may be zero or more elements, but if at least one is
11159 ;; given then the first is the anchor position (or nil). The symbol
11160 ;; is matched against `c-offsets-alist' and the offset calculated
11161 ;; from that is returned.
11162 ;;
11163 ;; This function might do hidden buffer changes.
11173 offset
11177 ))
11180 ;; This is a compatibility wrapper for `c-calc-offset' in case
11181 ;; someone is calling it directly. It takes an old style syntactic
11182 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
11183 ;; new list form.
11184 ;;
11185 ;; This function might do hidden buffer changes.
11192 ;; Calculate the syntactic indentation from a syntactic description
11193 ;; as returned by `c-guess-syntax'.
11194 ;;
11195 ;; Note that topmost-intro always has an anchor position at bol, for
11196 ;; historical reasons. It's often used together with other symbols
11197 ;; that has more sane positions. Since we always use the first
11198 ;; found anchor position, we rely on that these other symbols always
11199 ;; precede topmost-intro in the LANGELEMS list.
11200 ;;
11201 ;; This function might do hidden buffer changes.
11209 ;; Got an absolute column that overrides any indentation
11210 ;; we've collected so far, but not the relative
11211 ;; indentation we might get for the nested structures
11212 ;; further down the langelems list.
11216 ;; Got a relative change of the current calculated
11217 ;; indentation.
11220 ;; Use the anchor position from the first syntactic
11221 ;; element with one.
11231 indent)))
11233 \f
11236 ;; Local Variables:
11237 ;; indent-tabs-mode: t
11238 ;; tab-width: 8
11239 ;; End:
11240 ;;; cc-engine.el ends here