| blob | 823d795b6de8af248e2792f2b3ebd1cd2ed61e7a |
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.
628 noerror comma-delim)
629 "Move to the start of the current statement or declaration, or to
630 the previous one if already at the beginning of one. Only
631 statements/declarations on the same level are considered, i.e. don't
632 move into or out of sexps (not even normal expression parentheses).
634 If point is already at the earliest statement within braces or parens,
635 this function doesn't move back into any whitespace preceding it; it
636 returns 'same in this case.
640 is within the continuation. If starting at such a token, move to the
641 corresponding statement start. If at the beginning of a statement,
642 move to the closest containing statement if there is any. This might
643 also stop at a continuation clause.
645 Labels are treated as part of the following statements if
646 IGNORE-LABELS is non-nil. (FIXME: Doesn't work if we stop at a known
647 statement start keyword.) Otherwise, each label is treated as a
648 separate statement.
650 Macros are ignored \(i.e. skipped over) unless point is within one, in
651 which case the content of the macro is treated as normal code. Aside
652 from any normal statement starts found in it, stop at the first token
655 macros before leaving them.
657 Return:
659 'same if stopped at the beginning of the current statement;
660 'up if stepped to a containing statement;
661 'previous if stepped to a preceding statement;
662 'beginning if stepped from a statement continuation clause to
663 its start clause; or
664 'macro if stepped to a macro start.
665 Note that 'same and not 'label is returned if stopped at the same
666 label without crossing the colon character.
668 LIM may be given to limit the search. If the search hits the limit,
669 point will be left at the closest following token, or at the start
670 position if that is less ('same is returned in this case).
672 NOERROR turns off error logging to `c-parsing-error'.
674 Normally only ';' and virtual semicolons are considered to delimit
675 statements, but if COMMA-DELIM is non-nil then ',' is treated
676 as a delimiter too.
678 Note that this function might do hidden buffer changes. See the
679 comment at the start of cc-engine.el for more info."
681 ;; The bulk of this function is a pushdown automaton that looks at statement
682 ;; boundaries and the tokens (such as "while") in c-opt-block-stmt-key. Its
683 ;; purpose is to keep track of nested statements, ensuring that such
684 ;; statements are skipped over in their entirety (somewhat akin to what C-M-p
685 ;; does with nested braces/brackets/parentheses).
686 ;;
687 ;; Note: The position of a boundary is the following token.
688 ;;
689 ;; Beginning with the current token (the one following point), move back one
690 ;; sexp at a time (where a sexp is, more or less, either a token or the
691 ;; entire contents of a brace/bracket/paren pair). Each time a statement
692 ;; boundary is crossed or a "while"-like token is found, update the state of
693 ;; the PDA. Stop at the beginning of a statement when the stack (holding
694 ;; nested statement info) is empty and the position has been moved.
695 ;;
696 ;; The following variables constitute the PDA:
697 ;;
698 ;; sym: This is either the "while"-like token (e.g. 'for) we've just
699 ;; scanned back over, 'boundary if we've just gone back over a
700 ;; statement boundary, or nil otherwise.
701 ;; state: takes one of the values (nil else else-boundary while
702 ;; while-boundary catch catch-boundary).
703 ;; nil means "no "while"-like token yet scanned".
704 ;; 'else, for example, means "just gone back over an else".
705 ;; 'else-boundary means "just gone back over a statement boundary
706 ;; immediately after having gone back over an else".
707 ;; saved-pos: A vector of either saved positions (tok ptok pptok, etc.) or
708 ;; of error reporting information.
709 ;; stack: The stack onto which the PDA pushes its state. Each entry
710 ;; consists of a saved value of state and saved-pos. An entry is
711 ;; pushed when we move back over a "continuation" token (e.g. else)
712 ;; and popped when we encounter the corresponding opening token
713 ;; (e.g. if).
714 ;;
715 ;;
716 ;; The following diagram briefly outlines the PDA.
717 ;;
718 ;; Common state:
719 ;; "else": Push state, goto state `else'.
720 ;; "while": Push state, goto state `while'.
721 ;; "catch" or "finally": Push state, goto state `catch'.
722 ;; boundary: Pop state.
723 ;; other: Do nothing special.
724 ;;
725 ;; State `else':
726 ;; boundary: Goto state `else-boundary'.
727 ;; other: Error, pop state, retry token.
728 ;;
729 ;; State `else-boundary':
730 ;; "if": Pop state.
731 ;; boundary: Error, pop state.
732 ;; other: See common state.
733 ;;
734 ;; State `while':
735 ;; boundary: Save position, goto state `while-boundary'.
736 ;; other: Pop state, retry token.
737 ;;
738 ;; State `while-boundary':
739 ;; "do": Pop state.
740 ;; boundary: Restore position if it's not at start, pop state. [*see below]
741 ;; other: See common state.
742 ;;
743 ;; State `catch':
744 ;; boundary: Goto state `catch-boundary'.
745 ;; other: Error, pop state, retry token.
746 ;;
747 ;; State `catch-boundary':
748 ;; "try": Pop state.
749 ;; "catch": Goto state `catch'.
750 ;; boundary: Error, pop state.
751 ;; other: See common state.
752 ;;
753 ;; [*] In the `while-boundary' state, we had pushed a 'while state, and were
754 ;; searching for a "do" which would have opened a do-while. If we didn't
755 ;; find it, we discard the analysis done since the "while", go back to this
756 ;; token in the buffer and restart the scanning there, this time WITHOUT
757 ;; pushing the 'while state onto the stack.
758 ;;
759 ;; In addition to the above there is some special handling of labels
760 ;; and macros.
764 macro-start
767 c-stmt-delim-chars-with-comma
768 c-stmt-delim-chars))
769 c-in-literal-cache c-maybe-labelp after-case:-pos saved
770 ;; Current position.
771 pos
772 ;; Position of last stmt boundary character (e.g. ;).
773 boundary-pos
774 ;; The position of the last sexp or bound that follows the
775 ;; first found colon, i.e. the start of the nonlabel part of
776 ;; the statement. It's `start' if a colon is found just after
777 ;; the start.
778 after-labels-pos
779 ;; Like `after-labels-pos', but the first such position inside
780 ;; a label, i.e. the start of the last label before the start
781 ;; of the nonlabel part of the statement.
782 last-label-pos
783 ;; The last position where a label is possible provided the
784 ;; statement started there. It's nil as long as no invalid
785 ;; label content has been found (according to
786 ;; `c-nonlabel-token-key'). It's `start' if no valid label
787 ;; content was found in the label. Note that we might still
788 ;; regard it a label if it starts with `c-label-kwds'.
789 label-good-pos
790 ;; Putative positions of the components of a bitfield declaration,
791 ;; e.g. "int foo : NUM_FOO_BITS ;"
792 bitfield-type-pos bitfield-id-pos bitfield-size-pos
793 ;; Symbol just scanned back over (e.g. 'while or 'boundary).
794 ;; See above.
795 sym
796 ;; Current state in the automaton. See above.
797 state
798 ;; Current saved positions. See above.
799 saved-pos
800 ;; Stack of conses (state . saved-pos).
801 stack
802 ;; Regexp which matches "for", "if", etc.
805 ;; Return value.
807 ;; Positions of the last three sexps or bounds we've stopped at.
808 tok ptok pptok)
818 ;; Try to skip back over unary operator characters, to register
819 ;; that we've moved.
823 ;; Protect post-++/-- operators just before a virtual semicolon.
827 ;; Skip back over any semicolon here. If it was a bare semicolon, we're
828 ;; done. Later on we ignore the boundaries for statements that don't
829 ;; contain any sexp. The only thing that is affected is that the error
830 ;; checking is a little less strict, and we really don't bother.
840 pos saved)
842 ;; Begin at start and not pos to detect macros if we stand
843 ;; directly after the #.
846 ;; Record this as the first token if not starting inside it.
849 ;; The following while loop goes back one sexp (balanced parens,
850 ;; etc. with contents, or symbol or suchlike) each iteration. This
851 ;; movement is accomplished with a call to c-backward-sexp approx 170
852 ;; lines below.
853 ;;
854 ;; The loop is exited only by throwing nil to the (catch 'loop ...):
855 ;; 1. On reaching the start of a macro;
856 ;; 2. On having passed a stmt boundary with the PDA stack empty;
857 ;; 3. On reaching the start of an Objective C method def;
858 ;; 4. From macro `c-bos-pop-state'; when the stack is empty;
859 ;; 5. From macro `c-bos-pop-state-and-retry' when the stack is empty.
863 ;; Are we in a macro, just after the opening #?
878 ;; Stop at the first token in the content of the macro.
882 ret 'macro
883 ignore-labels t))
886 ;; Do a round through the automaton if we've just passed a
887 ;; statement boundary or passed a "while"-like token.
895 ;; The PDA state handling.
896 ;;
897 ;; Refer to the description of the PDA in the opening
898 ;; comments. In the following OR form, the first leaf
899 ;; attempts to handles one of the specific actions detailed
900 ;; (e.g., finding token "if" whilst in state `else-boundary').
901 ;; We drop through to the second leaf (which handles common
902 ;; state) if no specific handler is found in the first cond.
903 ;; If a parsing error is detected (e.g. an "else" with no
904 ;; preceding "if"), we throw to the enclosing catch.
905 ;;
906 ;; Note that the (eq state 'else) means
907 ;; "we've just passed an else", NOT "we're looking for an
908 ;; else".
925 ;; Since this can cause backtracking we do a
926 ;; little more careful analysis to avoid it:
927 ;; If there's a label in front of the while
928 ;; it can't be part of a do-while.
957 ;; This is state common. We get here when the previous
958 ;; cond statement found no particular state handler.
960 ;; If we have a boundary at the start
961 ;; position we push a frame to go to the
962 ;; previous statement.
971 ;; Is this a real while, or a do-while?
972 ;; The next `when' triggers unless we are SURE that
973 ;; the `while' is not the tail end of a `do-while'.
976 ;; The following kludge is to prevent
977 ;; infinite recursion when called from
978 ;; c-awk-after-if-for-while-condition-p,
979 ;; or the like.
983 ;; Since this can cause backtracking we do a
984 ;; little more careful analysis to avoid it: If
985 ;; the while isn't followed by a (possibly
986 ;; virtual) semicolon it can't be a do-while.
995 ;; We're either past a statement boundary or at the
996 ;; start of a statement, so throw away any label data
997 ;; for the previous one.
999 last-label-pos nil
1000 c-maybe-labelp nil))))
1002 ;; Step to the previous sexp, but not if we crossed a
1003 ;; boundary, since that doesn't consume an sexp.
1007 ;; HERE IS THE SINGLE PLACE INSIDE THE PDA LOOP WHERE WE MOVE
1008 ;; BACKWARDS THROUGH THE SOURCE.
1012 ;; The end position of the area to search for statement
1013 ;; barriers in this round.
1016 ;; Go back over exactly one logical sexp, taking proper
1017 ;; account of macros and escaped EOLs.
1021 ;; Give up if we hit an unbalanced block. Since the
1022 ;; stack won't be empty the code below will report a
1023 ;; suitable error.
1026 ;; Have we moved into a macro?
1029 ;; Have we crossed a statement boundary? If not,
1030 ;; keep going back until we find one or a "real" sexp.
1037 ;; Have we just gone back over an escaped NL? This
1038 ;; doesn't count as a sexp.
1041 ;; Have we crossed a statement boundary?
1044 ;; Are we at a macro beginning?
1046 c-opt-cpp-prefix
1052 ;; Just gone back over a brace block?
1061 ;; Just gone back over some paren block?
1065 before-sws-pos)))
1068 ;; Just gone back over an ordinary symbol of some sort?
1074 ptok tok
1075 tok boundary-pos
1077 ;; Like a C "continue". Analyze the next sexp.
1080 ;; ObjC method def?
1087 ;; Might we have a bitfield declaration, "<type> <id> : <size>"?
1090 ;; The : <size> and <id> fields?
1100 ;; The <type> field?
1108 bitfield-id-pos nil)))))
1110 ;; Handle labels.
1113 ;; `c-crosses-statement-barrier-p' has found a colon, so we
1114 ;; might be in a label now. Have we got a real label
1115 ;; (including a case label) or something like C++'s "public:"?
1116 ;; A case label might use an expression rather than a token.
1119 ;; Catch C++'s inheritance construct "class foo : bar".
1133 ; been found.
1135 ;; We're in a potential label and it's the first
1136 ;; time we've found something that isn't allowed in
1137 ;; one.
1140 ;; We've moved back by a sexp, so update the token positions.
1142 pptok ptok
1143 ptok tok
1149 ;; If the stack isn't empty there might be errors to report.
1158 ;; Need to investigate closer whether we've crossed
1159 ;; between a substatement and its containing statement.
1163 pptok)
1165 ptok)
1173 after-labels-pos
1180 last-label-pos
1181 pos))
1183 ;; We're in a label. Maybe we should step to the statement
1184 ;; after it.
1189 ;; Might have jumped over several labels. Go to the last one.
1192 ;; Have we got "case <expression>:"?
1202 ;; Skip over the unary operators that can start the statement.
1205 ;; protect AWK post-inc/decrement operators, etc.
1210 ret)))
1213 "Return non-nil if there is a non-comment non-macro punctuation character
1214 between FROM and TO. FROM must not be in a string or comment. The returned
1215 value is the position of the first such character."
1226 "Return non-nil if buffer positions FROM to TO cross one or more
1227 statement or declaration boundaries. The returned value is actually
1228 the position of the earliest boundary char. FROM must not be within
1229 a string or comment.
1231 The variable `c-maybe-labelp' is set to the position of the first `:' that
1232 might start a label (i.e. not part of `::' and not preceded by `?'). If a
1233 single `?' is found, then `c-maybe-labelp' is cleared.
1235 For AWK, a statement which is terminated by an EOL (not a \; or a }) is
1237 the line. If this virtual semicolon is _at_ from, the function recognizes it.
1239 Note that this function might do hidden buffer changes. See the
1240 comment at the start of cc-engine.el for more info."
1242 ;; If the current language has CPP macros, insert # into skip-chars.
1245 c-opt-cpp-symbol ; usually "#"
1247 c-stmt-delim-chars))
1250 lit-range vsemi-pos)
1257 skip-chars
1261 ;; Virtual semicolon?
1271 ;; In a string/comment?
1278 ;; Ignore scope operators.
1282 ;; A question mark. Can't be a label, so stop
1283 ;; looking for more : and ?.
1286 ;; At a CPP construct or a "#" or "##" operator?
1297 ;; In trailing space after an as yet undetected virtual semicolon?
1300 ; unterminated string/regexp.
1305 nil))))))
1308 "Return non-nil if the point is at the first token in a statement
1309 or somewhere in the syntactic whitespace before it.
1312 Any kind of declaration-like construct that occur outside function
1315 Note that this function might do hidden buffer changes. See the
1316 comment at the start of cc-engine.el for more info."
1320 c-maybe-labelp)
1331 "Return non-nil if the point is at the first token in an expression or
1332 statement, or somewhere in the syntactic whitespace before it.
1335 grammar sense: It's any sequence of expression tokens except commas,
1336 unless they are enclosed inside parentheses of some kind. Also, an
1337 expression never continues past an enclosing parenthesis, but it might
1338 contain parenthesis pairs of any sort except braces.
1340 Since expressions never cross statement boundaries, this function also
1341 recognizes statement beginnings, just like `c-at-statement-start-p'.
1343 Note that this function might do hidden buffer changes. See the
1344 comment at the start of cc-engine.el for more info."
1349 c-maybe-labelp)
1357 \f
1358 ;; A set of functions that covers various idiosyncrasies in
1359 ;; implementations of `forward-comment'.
1361 ;; Note: Some emacsen considers incorrectly that any line comment
1362 ;; ending with a backslash continues to the next line. I can't think
1363 ;; of any way to work around that in a reliable way without changing
1364 ;; the buffer, though. Suggestions welcome. ;) (No, temporarily
1365 ;; changing the syntax for backslash doesn't work since we must treat
1366 ;; escapes in string literals correctly.)
1369 "Move forward past whitespace and the closest following comment, if any.
1370 Return t if a comment was found, nil otherwise. In either case, the
1371 point is moved past the following whitespace. Line continuations,
1372 i.e. a backslashes followed by line breaks, are treated as whitespace.
1373 The line breaks that end line comments are considered to be the
1374 comment enders, so the point will be put on the beginning of the next
1375 line if it moved past a line comment.
1377 This function does not do any hidden buffer changes."
1385 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1386 ;; forwards at eob.
1387 nil
1389 ;; Emacs includes the ending newline in a b-style (c++)
1390 ;; comment, but XEmacs doesn't. We depend on the Emacs
1391 ;; behavior (which also is symmetric).
1395 t))))
1398 "Move forward past all following whitespace and comments.
1399 Line continuations, i.e. a backslashes followed by line breaks, are
1400 treated as whitespace.
1402 Note that this function might do hidden buffer changes. See the
1403 comment at the start of cc-engine.el for more info."
1406 ;; If forward-comment in at least XEmacs 21 is given a large
1407 ;; positive value, it'll loop all the way through if it hits
1408 ;; eob.
1410 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1411 ;; forwards at eob.
1416 t))))
1419 "Move backward past whitespace and the closest preceding comment, if any.
1420 Return t if a comment was found, nil otherwise. In either case, the
1421 point is moved past the preceding whitespace. Line continuations,
1422 i.e. a backslashes followed by line breaks, are treated as whitespace.
1423 The line breaks that end line comments are considered to be the
1424 comment enders, so the point cannot be at the end of the same line to
1425 move over a line comment.
1427 This function does not do any hidden buffer changes."
1430 ;; When we got newline terminated comments, forward-comment in all
1431 ;; supported emacsen so far will stop at eol of each line not
1432 ;; ending with a comment when moving backwards. This corrects for
1433 ;; that, and at the same time handles line continuations.
1441 ;; Some emacsen (e.g. Emacs 19.34) return t when moving
1442 ;; backwards at bob.
1443 nil
1445 ;; Leave point after the closest following newline if we've
1446 ;; backed up over any above, since forward-comment won't move
1447 ;; backward over a line comment if point is at the end of the
1448 ;; same line.
1453 ;; If forward-comment above succeeded and we're at eol
1454 ;; then the newline we moved over above didn't end a
1455 ;; line comment, so we give it another go.
1458 t))
1460 ;; Emacs <= 20 and XEmacs move back over the closer of a
1461 ;; block comment that lacks an opener.
1464 t)))))
1467 "Move backward past all preceding whitespace and comments.
1468 Line continuations, i.e. a backslashes followed by line breaks, are
1469 treated as whitespace. The line breaks that end line comments are
1470 considered to be the comment enders, so the point cannot be at the end
1471 of the same line to move over a line comment. Unlike
1472 c-backward-syntactic-ws, this function doesn't move back over
1473 preprocessor directives.
1475 Note that this function might do hidden buffer changes. See the
1476 comment at the start of cc-engine.el for more info."
1480 ;; `forward-comment' in some emacsen (e.g. XEmacs 21.4)
1481 ;; return t when moving backwards at bob.
1487 ;; Cope specifically with ^M^J here -
1488 ;; forward-comment sometimes gets stuck after ^Ms,
1489 ;; sometimes after ^M^J.
1493 t)
1497 t))))
1498 moved-comment)
1500 ;; Emacs <= 20 and XEmacs move back over the
1501 ;; closer of a block comment that lacks an opener.
1503 t)
1505 ;; XEmacs treats line continuations as whitespace but
1506 ;; only in the backward direction, which seems a bit
1507 ;; odd. Anyway, this is necessary for Emacs.
1512 t))))))
1514 \f
1515 ;; Tools for skipping over syntactic whitespace.
1517 ;; The following functions use text properties to cache searches over
1518 ;; large regions of syntactic whitespace. It works as follows:
1519 ;;
1520 ;; o If a syntactic whitespace region contains anything but simple
1521 ;; whitespace (i.e. space, tab and line breaks), the text property
1522 ;; `c-in-sws' is put over it. At places where we have stopped
1523 ;; within that region there's also a `c-is-sws' text property.
1524 ;; That since there typically are nested whitespace inside that
1525 ;; must be handled separately, e.g. whitespace inside a comment or
1526 ;; cpp directive. Thus, from one point with `c-is-sws' it's safe
1527 ;; to jump to another point with that property within the same
1528 ;; `c-in-sws' region. It can be likened to a ladder where
1529 ;; `c-in-sws' marks the bars and `c-is-sws' the rungs.
1530 ;;
1531 ;; o The `c-is-sws' property is put on the simple whitespace chars at
1532 ;; a "rung position" and also maybe on the first following char.
1533 ;; As many characters as can be conveniently found in this range
1534 ;; are marked, but no assumption can be made that the whole range
1535 ;; is marked (it could be clobbered by later changes, for
1536 ;; instance).
1537 ;;
1538 ;; Note that some part of the beginning of a sequence of simple
1539 ;; whitespace might be part of the end of a preceding line comment
1540 ;; or cpp directive and must not be considered part of the "rung".
1541 ;; Such whitespace is some amount of horizontal whitespace followed
1542 ;; by a newline. In the case of cpp directives it could also be
1543 ;; two newlines with horizontal whitespace between them.
1544 ;;
1545 ;; The reason to include the first following char is to cope with
1546 ;; "rung positions" that doesn't have any ordinary whitespace. If
1547 ;; `c-is-sws' is put on a token character it does not have
1548 ;; `c-in-sws' set simultaneously. That's the only case when that
1549 ;; can occur, and the reason for not extending the `c-in-sws'
1550 ;; region to cover it is that the `c-in-sws' region could then be
1551 ;; accidentally merged with a following one if the token is only
1552 ;; one character long.
1553 ;;
1554 ;; o On buffer changes the `c-in-sws' and `c-is-sws' properties are
1555 ;; removed in the changed region. If the change was inside
1556 ;; syntactic whitespace that means that the "ladder" is broken, but
1557 ;; a later call to `c-forward-sws' or `c-backward-sws' will use the
1558 ;; parts on either side and use an ordinary search only to "repair"
1559 ;; the gap.
1560 ;;
1561 ;; Special care needs to be taken if a region is removed: If there
1562 ;; are `c-in-sws' on both sides of it which do not connect inside
1563 ;; the region then they can't be joined. If e.g. a marked macro is
1564 ;; broken, syntactic whitespace inside the new text might be
1565 ;; marked. If those marks would become connected with the old
1566 ;; `c-in-sws' range around the macro then we could get a ladder
1567 ;; with one end outside the macro and the other at some whitespace
1568 ;; within it.
1569 ;;
1570 ;; The main motivation for this system is to increase the speed in
1571 ;; skipping over the large whitespace regions that can occur at the
1572 ;; top level in e.g. header files that contain a lot of comments and
1573 ;; cpp directives. For small comments inside code it's probably
1574 ;; slower than using `forward-comment' straightforwardly, but speed is
1575 ;; not a significant factor there anyway.
1577 ; (defface c-debug-is-sws-face
1578 ; '((t (:background "GreenYellow")))
1579 ; "Debug face to mark the `c-is-sws' property.")
1580 ; (defface c-debug-in-sws-face
1581 ; '((t (:underline t)))
1582 ; "Debug face to mark the `c-in-sws' property.")
1584 ; (defun c-debug-put-sws-faces ()
1585 ; ;; Put the sws debug faces on all the `c-is-sws' and `c-in-sws'
1586 ; ;; properties in the buffer.
1587 ; (interactive)
1588 ; (save-excursion
1589 ; (c-save-buffer-state (in-face)
1590 ; (goto-char (point-min))
1591 ; (setq in-face (if (get-text-property (point) 'c-is-sws)
1592 ; (point)))
1593 ; (while (progn
1594 ; (goto-char (next-single-property-change
1595 ; (point) 'c-is-sws nil (point-max)))
1596 ; (if in-face
1597 ; (progn
1598 ; (c-debug-add-face in-face (point) 'c-debug-is-sws-face)
1599 ; (setq in-face nil))
1600 ; (setq in-face (point)))
1601 ; (not (eobp))))
1602 ; (goto-char (point-min))
1603 ; (setq in-face (if (get-text-property (point) 'c-in-sws)
1604 ; (point)))
1605 ; (while (progn
1606 ; (goto-char (next-single-property-change
1607 ; (point) 'c-in-sws nil (point-max)))
1608 ; (if in-face
1609 ; (progn
1610 ; (c-debug-add-face in-face (point) 'c-debug-in-sws-face)
1611 ; (setq in-face nil))
1612 ; (setq in-face (point)))
1613 ; (not (eobp)))))))
1616 ;;`(message ,@args)
1617 )
1620 ;; This macro does a hidden buffer change.
1627 ;; This macro does a hidden buffer change.
1634 ;; This macro does a hidden buffer change.
1641 ;; This macro does a hidden buffer change.
1648 ;; This macro does a hidden buffer change.
1656 ;; Called from `after-change-functions'. Note that if
1657 ;; `c-forward-sws' or `c-backward-sws' are used outside
1658 ;; `c-save-buffer-state' or similar then this will remove the cache
1659 ;; properties right after they're added.
1660 ;;
1661 ;; This function does hidden buffer changes.
1664 ;; Adjust the end to remove the properties in any following simple
1665 ;; ws up to and including the next line break, if there is any
1666 ;; after the changed region. This is necessary e.g. when a rung
1667 ;; marked empty line is converted to a line comment by inserting
1668 ;; "//" before the line break. In that case the line break would
1669 ;; keep the rung mark which could make a later `c-backward-sws'
1670 ;; move into the line comment instead of over it.
1680 ;; Ensure that an `c-in-sws' range gets broken. Note that it isn't
1681 ;; safe to keep a range that was continuous before the change. E.g:
1682 ;;
1683 ;; #define foo
1684 ;; \
1685 ;; bar
1686 ;;
1687 ;; There can be a "ladder" between "#" and "b". Now, if the newline
1688 ;; after "foo" is removed then "bar" will become part of the cpp
1689 ;; directive instead of a syntactically relevant token. In that
1690 ;; case there's no longer syntactic ws from "#" to "b".
1697 ;; Used by `c-forward-syntactic-ws' to implement the unbounded search.
1698 ;;
1699 ;; This function might do hidden buffer changes.
1702 ;; unmarked part of the simple ws region.
1704 rung-is-marked next-rung-is-marked simple-ws-end
1705 ;; `safe-start' is set when it's safe to cache the start position.
1706 ;; It's not set if we've initially skipped over comments and line
1707 ;; continuations since we might have gone out through the end of a
1708 ;; macro then. This provision makes `c-forward-sws' not populate the
1709 ;; cache in the majority of cases, but otoh is `c-backward-sws' by far
1710 ;; more common.
1711 safe-start)
1713 ;; Skip simple ws and do a quick check on the following character to see
1714 ;; if it's anything that can't start syntactic ws, so we can bail out
1715 ;; early in the majority of cases when there just are a few ws chars.
1722 ;; Find the last rung position to avoid setting properties in all
1723 ;; the cases when the marked rung is complete.
1724 ;; (`next-single-property-change' is certain to move at least one
1725 ;; step forward.)
1728 ;; Got no marked rung here. Since the simple ws might have started
1729 ;; inside a line comment or cpp directive we must set `rung-pos' as
1730 ;; high as possible.
1740 ;; The following search is the main reason that `c-in-sws'
1741 ;; and `c-is-sws' aren't combined to one property.
1745 ;; If the `c-in-sws' region extended past the last
1746 ;; `c-is-sws' char we have to go back a bit.
1753 "c-forward-sws cached move %s -> %s (max %s)"
1760 ;; We'll loop here if there is simple ws after the last rung.
1761 ;; That means that there's been some change in it and it's
1762 ;; possible that we've stepped into another ladder, so extend
1763 ;; the previous one to join with it if there is one, and try to
1764 ;; use the cache again.
1766 "c-forward-sws extending rung with [%s..%s] (max %s)"
1769 ;; Remove any `c-in-sws' property from the last char of
1770 ;; the rung before we mark it with `c-is-sws', so that we
1771 ;; won't connect with the remains of a broken "ladder".
1777 last-put-in-sws-pos rung-pos)))
1784 ;; Skipped over comments. Don't cache at eob in case the buffer
1785 ;; is narrowed.
1796 ;; Skip a preprocessor directive.
1803 ;; Don't cache at eob in case the buffer is narrowed.
1806 ;; We've searched over a piece of non-white syntactic ws. See if this
1807 ;; can be cached.
1813 ;; Cache if we haven't skipped comments only, and if we started
1814 ;; either from a marked rung or from a completely uncached
1815 ;; position.
1820 ;; See if there's a marked rung in the encountered simple ws. If
1821 ;; so then we can cache, unless `safe-start' is nil. Even then
1822 ;; we need to do this to check if the cache can be used for the
1823 ;; next step.
1827 safe-start))
1831 "c-forward-sws caching [%s..%s] - [%s..%s] (max %s)"
1835 ;; Remove the properties for any nested ws that might be cached.
1836 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
1837 ;; anyway.
1845 last-put-in-sws-pos rung-pos))
1847 ;; Remove any `c-in-sws' property from the last char of
1848 ;; the rung before we mark it with `c-is-sws', so that we
1849 ;; won't connect with the remains of a broken "ladder".
1852 rung-end-pos))
1855 "c-forward-sws not caching [%s..%s] - [%s..%s] (max %s)"
1859 ;; Set `rung-pos' for the next rung. It's the same thing here as
1860 ;; initially, except that the rung position is set as early as
1861 ;; possible since we can't be in the ending ws of a line comment or
1862 ;; cpp directive now.
1869 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
1870 ;; another one after the point (which might occur when editing inside a
1871 ;; comment or macro).
1875 "c-forward-sws clearing at %s for cache separation"
1876 last-put-in-sws-pos)
1880 ;; If at eob we have to clear the last character before the end
1881 ;; instead since the buffer might be narrowed and there might
1882 ;; be a `c-in-sws' after (point-max). In this case it's
1883 ;; necessary to clear both properties.
1885 "c-forward-sws clearing thoroughly at %s for cache separation"
1888 last-put-in-sws-pos))))
1889 ))))
1892 ;; Used by `c-backward-syntactic-ws' to implement the unbounded search.
1893 ;;
1894 ;; This function might do hidden buffer changes.
1897 ;; part of the simple ws region.
1899 rung-is-marked simple-ws-beg cmt-skip-pos)
1901 ;; Skip simple horizontal ws and do a quick check on the preceding
1902 ;; character to see if it's anything that can't end syntactic ws, so we can
1903 ;; bail out early in the majority of cases when there just are a few ws
1904 ;; chars. Newlines are complicated in the backward direction, so we can't
1905 ;; skip over them.
1912 ;; Try to find a rung position in the simple ws preceding point, so that
1913 ;; we can get a cache hit even if the last bit of the simple ws has
1914 ;; changed recently.
1920 ;; `rung-pos' will be the earliest marked position, which means that
1921 ;; there might be later unmarked parts in the simple ws region.
1922 ;; It's not worth the effort to fix that; the last part of the
1923 ;; simple ws is also typically edited often, so it could be wasted.
1935 ;; The following search is the main reason that `c-in-sws'
1936 ;; and `c-is-sws' aren't combined to one property.
1940 ;; If the `c-in-sws' region extended past the first
1941 ;; `c-is-sws' char we have to go forward a bit.
1946 "c-backward-sws cached move %s <- %s (min %s)"
1958 t
1960 nil))
1962 ;; We'll loop here if there is simple ws before the first rung.
1963 ;; That means that there's been some change in it and it's
1964 ;; possible that we've stepped into another ladder, so extend
1965 ;; the previous one to join with it if there is one, and try to
1966 ;; use the cache again.
1968 "c-backward-sws extending rung with [%s..%s] (min %s)"
1971 ;; Remove any `c-in-sws' property from the last char of
1972 ;; the rung before we mark it with `c-is-sws', so that we
1973 ;; won't connect with the remains of a broken "ladder".
1976 rung-pos)
1980 last-put-in-sws-pos rung-pos))
1989 ;; Inside a cpp directive. See if it should be skipped over.
1992 ;; Move back over all line continuations in the region skipped
1993 ;; over by `c-backward-comments'. If we go past it then we
1994 ;; started inside the cpp directive.
2003 ;; Don't move past the cpp directive if we began inside
2004 ;; it. Note that the position at the end of the last line
2005 ;; of the macro is also considered to be within it.
2007 nil)
2009 ;; It's worthwhile to spend a little bit of effort on finding
2010 ;; the end of the macro, to get a good `simple-ws-beg'
2011 ;; position for the cache. Note that `c-backward-comments'
2012 ;; could have stepped over some comments before going into
2013 ;; the macro, and then `simple-ws-beg' must be kept on the
2014 ;; same side of those comments.
2021 ;; Might happen if comments after the macro were skipped
2022 ;; over.
2026 t)))
2031 simple-ws-beg)
2032 ;; Skipped over comments. Must put point at the end of
2033 ;; the simple ws at point since we might be after a line
2034 ;; comment or cpp directive that's been partially
2035 ;; narrowed out, and we can't risk marking the simple ws
2036 ;; at the end of it.
2038 t)))
2040 ;; We've searched over a piece of non-white syntactic ws. See if this
2041 ;; can be cached.
2046 ;; Cache if we started either from a marked rung or from a
2047 ;; completely uncached position.
2048 rung-is-marked
2051 ;; Cache if there's a marked rung in the encountered simple ws.
2059 "c-backward-sws caching [%s..%s] - [%s..%s] (min %s)"
2064 ;; Remove the properties for any nested ws that might be cached.
2065 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
2066 ;; anyway.
2071 ;; Remove any `c-in-sws' property from the last char of
2072 ;; the rung before we mark it with `c-is-sws', so that we
2073 ;; won't connect with the remains of a broken "ladder".
2076 rung-end-pos)
2079 last-put-in-sws-pos simple-ws-beg)
2080 rung-pos)
2085 "c-backward-sws not caching [%s..%s] - [%s..%s] (min %s)"
2091 ))
2093 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
2094 ;; another one before the point (which might occur when editing inside a
2095 ;; comment or macro).
2099 "c-backward-sws clearing at %s for cache separation"
2102 last-put-in-sws-pos))
2104 ;; If at bob and the buffer is narrowed, we have to clear the
2105 ;; character we're standing on instead since there might be a
2106 ;; `c-in-sws' before (point-min). In this case it's necessary
2107 ;; to clear both properties.
2109 "c-backward-sws clearing thoroughly at %s for cache separation"
2110 last-put-in-sws-pos)
2113 ))))
2115 \f
2116 ;; Other whitespace tools
2118 ;; Is the region (beg end) WS, and is there WS (or BOB/EOB) next to the
2119 ;; region? This is a "heuristic" function. .....
2120 ;;
2121 ;; The motivation for the second bit is to check whether removing this
2122 ;; region would coalesce two symbols.
2123 ;;
2124 ;; FIXME!!! This function doesn't check virtual semicolons in any way. Be
2125 ;; careful about using this function for, e.g. AWK. (2007/3/7)
2134 \f
2135 ;; A system for finding noteworthy parens before the point.
2138 ;; A maximum comfortable scanning distance, e.g. between
2139 ;; `c-state-cache-good-pos' and "HERE" (where we call c-parse-state). When
2140 ;; this distance is exceeded, we take "emergency measures", e.g. by clearing
2141 ;; the cache and starting again from point-min or a beginning of defun. This
2142 ;; value can be tuned for efficiency or set to a lower value for testing.
2146 ;; The state cache used by `c-parse-state' to cut down the amount of
2147 ;; searching. It's the result from some earlier `c-parse-state' call. See
2148 ;; `c-parse-state''s doc string for details of its structure.
2149 ;;
2150 ;; The use of the cached info is more effective if the next
2151 ;; `c-parse-state' call is on a line close by the one the cached state
2152 ;; was made at; the cache can actually slow down a little if the
2153 ;; cached state was made very far back in the buffer. The cache is
2154 ;; most effective if `c-parse-state' is used on each line while moving
2155 ;; forward.
2159 ;; This is a position where `c-state-cache' is known to be correct, or
2160 ;; nil (see below). It's a position inside one of the recorded unclosed
2161 ;; parens or the top level, but not further nested inside any literal or
2162 ;; subparen that is closed before the last recorded position.
2163 ;;
2164 ;; The exact position is chosen to try to be close to yet earlier than
2165 ;; the position where `c-state-cache' will be called next. Right now
2166 ;; the heuristic is to set it to the position after the last found
2167 ;; closing paren (of any type) before the line on which
2168 ;; `c-parse-state' was called. That is chosen primarily to work well
2169 ;; with refontification of the current line.
2170 ;;
2171 ;; 2009-07-28: When `c-state-point-min' and the last position where
2172 ;; `c-parse-state' or for which `c-invalidate-state-cache' was called, are
2173 ;; both in the same literal, there is no such "good position", and
2174 ;; c-state-cache-good-pos is then nil. This is the ONLY circumstance in which
2175 ;; it can be nil. In this case, `c-state-point-min-literal' will be non-nil.
2176 ;;
2177 ;; 2009-06-12: In a brace desert, c-state-cache-good-pos may also be in
2178 ;; the middle of the desert, as long as it is not within a brace pair
2179 ;; recorded in `c-state-cache' or a paren/bracket pair.
2181 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2182 ;; We maintain a simple cache of positions which aren't in a literal, so as to
2183 ;; speed up testing for non-literality.
2185 ;; The approximate interval between entries in `c-state-nonlit-pos-cache'.
2189 ;; A list of buffer positions which are known not to be in a literal or a cpp
2190 ;; construct. This is ordered with higher positions at the front of the list.
2191 ;; Only those which are less than `c-state-nonlit-pos-cache-limit' are valid.
2195 ;; An upper limit on valid entries in `c-state-nonlit-pos-cache'. This is
2196 ;; reduced by buffer changes, and increased by invocations of
2197 ;; `c-state-literal-at'.
2201 ;; A list of buffer positions which are known not to be in a literal. This is
2202 ;; ordered with higher positions at the front of the list. Only those which
2203 ;; are less than `c-state-semi-nonlit-pos-cache-limit' are valid.
2207 ;; An upper limit on valid entries in `c-state-semi-nonlit-pos-cache'. This is
2208 ;; reduced by buffer changes, and increased by invocations of
2209 ;; `c-state-literal-at'. FIXME!!!
2212 ;; Do a parse-partial-sexp from FROM to TO, returning either
2213 ;; (STATE TYPE (BEG . END)) if TO is in a literal; or
2214 ;; (STATE) otherwise,
2215 ;; where STATE is the parsing state at TO, TYPE is the type of the literal
2216 ;; (one of 'c, 'c++, 'string) and (BEG . END) is the boundaries of the literal.
2217 ;;
2218 ;; Unless NOT-IN-DELIMITER is non-nil, when TO is inside a two-character
2219 ;; comment opener, this is recognized as being in a comment literal.
2220 ;;
2221 ;; Only elements 3 (in a string), 4 (in a comment), 5 (following a quote),
2222 ;; 7 (comment type) and 8 (start of comment/string) (and possibly 9) of
2223 ;; STATE are valid.
2226 ty co-st)
2234 nil ; TARGETDEPTH
2235 nil ; STOPBEFORE
2236 s ; OLDSTATE
2253 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2254 ;; string, comment, or macro.
2255 ;;
2256 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2257 ;; MAY NOT contain any positions within macros, since macros are frequently
2258 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2259 ;; We cannot rely on this mechanism whilst determining a cache pos since
2260 ;; this function is also called from outwith `c-parse-state'.
2265 pos npos high-pos lit macro-beg macro-end)
2266 ;; Trim the cache to take account of buffer changes.
2278 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2284 ;; Test for being in a literal. If so, go to after it.
2291 ;; Test for being in a macro. If so, go to after it.
2306 ;; Add one extra element above HERE so as to to avoid the previous
2307 ;; expensive calculation when the next call is close to the current
2308 ;; one. This is especially useful when inside a large macro.
2315 pos))))
2318 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2319 ;; string or comment. It may be in a macro.
2324 pos npos high-pos lit macro-beg macro-end)
2325 ;; Trim the cache to take account of buffer changes.
2337 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2341 ;; Test for being in a literal. If so, go to after it.
2354 pos))))
2357 ;; If position HERE is inside a literal, return (START . END), the
2358 ;; boundaries of the literal (which may be outside the accessible bit of the
2359 ;; buffer). Otherwise, return nil.
2360 ;;
2361 ;; This function is almost the same as `c-literal-limits'. Previously, it
2362 ;; differed in that it was a lower level function, and that it rigorously
2363 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2364 ;; virtually identical to this function.
2372 ;; Return the start of the literal containing POS, or POS itself.
2374 pos))
2377 ;; Return a position outside of a string/comment/macro at or before POS.
2378 ;; STATE is the parse-partial-sexp state at POS.
2382 pos)))
2387 res))))
2389 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2390 ;; Stuff to do with point-min, and coping with any literal there.
2393 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2394 ;; narrowing is likely to affect the parens that are visible before the point.
2400 ;; These two variables define the literal, if any, containing point-min.
2401 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2402 ;; literal. If there's no literal there, they're both nil.
2406 ;; This is the earliest buffer-pos from which scanning can be done. It is
2407 ;; either the end of the literal containing point-min, or point-min itself.
2408 ;; It becomes nil if the buffer is changed earlier than this point.
2410 ;; Return the lowest valid scanning pos. This will be the end of the
2411 ;; literal enclosing point-min, or point-min itself.
2423 ;; Determine the properties of any literal containing POINT-MIN, setting the
2424 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2425 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2427 lit)
2442 c-state-point-min-lit-start nil
2443 c-state-min-scan-pos p-min)))))
2446 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2447 ;; A variable which signals a brace dessert - helpful for reducing the number
2448 ;; of fruitless backward scans.
2451 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2452 ;; that defun has searched backwards for a brace pair and not found one. Its
2453 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2454 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2455 ;; nil when at top level) and FROM is where the backward search started. It
2456 ;; is reset to nil in `c-invalidate-state-cache'.
2459 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2460 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2461 ;; list of like structure.
2463 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2464 ;; (default `c-state-cache') (or nil).
2471 ;; Return the address of the latest brace/bracket/paren (whether left or
2472 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2479 ;; Return the position just after the latest brace/bracket/paren (whether
2480 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2488 ;; From the state-cache, determine the buffer position from which we might
2489 ;; scan forward to HERE to update this cache. This position will be just
2490 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2491 ;; return the earliest position in the accessible region which isn't within
2492 ;; a literal. If the visible portion of the buffer is entirely within a
2493 ;; literal, return NIL.
2495 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2511 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2512 ;; Variables which keep track of preprocessor constructs.
2521 ;; These are the limits of the macro containing point at the previous call of
2522 ;; `c-parse-state', or nil.
2524 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2525 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2527 ;; Return the position of the opening paren/brace/bracket before HERE- which
2528 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2529 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2530 ;;
2531 ;; ............................................
2532 ;; | |
2533 ;; ( [ ( .........#macro.. ) ( ) ] )
2534 ;; ^ ^ ^ ^
2535 ;; | | | |
2536 ;; return HERE- HERE+ TOP
2537 ;;
2538 ;; If there aren't enough opening paren/brace/brackets, return the position
2539 ;; of the outermost one found, or HERE- if there are none. If there are no
2540 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2541 ;; must not be inside literals. Only the accessible portion of the buffer
2542 ;; will be scanned.
2544 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2545 ;; `here'. Go round the next loop each time we pass over such a ")". These
2546 ;; probably match "("s before `here-'.
2558 ;; PART 2: Scan back before `here-' searching for the "("s
2559 ;; matching/mismatching the ")"s found above. We only need to direct the
2560 ;; caller to scan when we've encountered unmatched right parens.
2569 pos))
2572 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2573 ;; to minimize the amount of scanning. HERE is the pertinent position in
2574 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2575 ;; its head trimmed) is known to be good, or nil if there is no such
2576 ;; position.
2577 ;;
2578 ;; The return value is a list, one of the following:
2579 ;;
2580 ;; o - ('forward START-POINT) - scan forward from START-POINT,
2581 ;; which is not less than the highest position in `c-state-cache' below HERE,
2582 ;; which is after GOOD-POS.
2583 ;; o - ('backward nil) - scan backwards (from HERE).
2584 ;; o - ('back-and-forward START-POINT) - like 'forward, but when HERE is earlier
2585 ;; than GOOD-POS.
2586 ;; o - ('IN-LIT nil) - point is inside the literal containing point-min.
2588 strategy ; 'forward, 'backward, or 'IN-LIT.
2589 start-point)
2601 start-point cache-pos)))
2605 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2606 ;; Routines which change `c-state-cache' and associated values.
2608 ;; The region (more precisely, point-min) has changed since we
2609 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2611 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2612 ;; It would be possible to do a better job here and recalculate the top
2613 ;; only.
2617 c-state-cache-good-pos c-state-min-scan-pos
2618 c-state-brace-pair-desert nil))
2620 ;; point-min has MOVED FORWARD.
2622 ;; Is the new point-min inside a (different) literal?
2627 ;; Cut off a bit of the tail from `c-state-cache'.
2629 pa)
2637 c-state-cache-good-pos c-state-min-scan-pos)
2640 )))
2645 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2646 ;; of nesting (not necessarily immediately preceding), push a cons onto
2647 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2648 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2649 ;; UPPER-LIM.
2650 ;;
2651 ;; Return non-nil when this has been done.
2652 ;;
2653 ;; The situation it copes with is this transformation:
2654 ;;
2655 ;; OLD: { (.) {...........}
2656 ;; ^ ^
2657 ;; FROM HERE
2658 ;;
2659 ;; NEW: { {....} (.) {.........
2660 ;; ^ ^ ^
2661 ;; LOWER BRACE PAIR HERE or HERE
2662 ;;
2663 ;; This routine should be fast. Since it can get called a LOT, we maintain
2664 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2665 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2675 ;; Don't start scanning in the middle of a CPP construct unless
2676 ;; it contains HERE - these constructs, in Emacs, are "commented
2677 ;; out" with category properties.
2680 macro-start-or-from
2681 from))
2685 ;; If we're essentially repeating a fruitless search, just give up.
2691 ;; DESERT-LIM. Avoid repeated searching through the cached desert.
2697 ;; CACHE-LIM. This limit will be necessary when an opening
2698 ;; paren at `cache-pos' has just had its matching close paren
2699 ;; inserted into the buffer. `cache-pos' continues to be a
2700 ;; search bound, even though the algorithm below would skip
2701 ;; over the new paren pair.
2710 ;; The top limit is EOB to ensure that `bra' is inside the
2711 ;; accessible part of the buffer at the next scan operation.
2714 ;; In the next pair of nested loops, the inner one moves back past a
2715 ;; pair of (mis-)matching parens or brackets; the outer one moves
2716 ;; back over a sequence of unmatched close brace/paren/bracket each
2717 ;; time round.
2733 ; brace, ignore it.
2736 ;; We've found the desired brace-pair.
2748 ;; We haven't found a brace pair. Record this in the cache.
2751 bra
2756 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2757 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2758 ;; "push" "a" brace pair onto `c-state-cache'.
2759 ;;
2760 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2761 ;; otherwise push it normally.
2762 ;;
2763 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2764 ;; latter is inside a macro, not being a macro containing
2765 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2766 ;; base pair. This latter case is assumed to be rare.
2767 ;;
2768 ;; Note: POINT is not preserved in this routine.
2778 c-state-cache)))
2779 ;; N.B. This defsubst codes one method for the simple, normal case,
2780 ;; and a more sophisticated, slower way for the general case. Don't
2781 ;; eliminate this defsubst - it's a speed optimization.
2785 ;; Scan the buffer from FROM to HERE, adding elements into `c-state-cache'
2786 ;; for braces etc. Return a candidate for `c-state-cache-good-pos'.
2787 ;;
2788 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2789 ;; any. Typically, it is immediately after it. It must not be inside a
2790 ;; literal.
2796 here)))
2799 ; Is actually the pos. to scan for a (/{/[ from,
2800 ; which sometimes is after a silly )/}/].
2803 ; good-pos.
2811 ;; Each time round the following loop, we enter a successively deeper
2812 ;; level of brace/paren nesting. (Except sometimes we "continue at
2813 ;; the existing level".) `pa+1' is a pos inside an opening
2814 ;; brace/paren/bracket, usually just after it.
2817 ;; Each time round the next loop moves forward over an opening then
2818 ;; a closing brace/bracket/paren. This loop is white hot, so it
2819 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2820 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2821 ;; call of `scan-lists' signals an error, which happens when there
2822 ;; are no more b/b/p's to scan.
2833 ;; We've just entered a deeper nesting level.
2835 ;; Insert the brace pair (if present) and the single open
2836 ;; paren/brace/bracket into `c-state-cache' It cannot be
2837 ;; inside a macro, except one around point, because of what
2838 ;; `c-neutralize-syntax-in-CPP' has done.
2840 ;; Insert the opening brace/bracket/paren position.
2842 ;; Clear admin stuff for the next more nested part of the scan.
2846 ;; All open p/b/b's at this nesting level, if any, have probably
2847 ;; been closed by matching/mismatching ones. We're probably
2848 ;; finished - we just need to check for having found an
2849 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2850 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2853 ;; Record the final, innermost, brace-pair if there is one.
2856 ;; Determine a good pos
2864 here-bol))
2873 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2874 ;; not be in it when it is amended for position HERE. This may involve
2875 ;; replacing a CONS element for a brace pair containing HERE with its car.
2876 ;; Additionally, the "outermost" open-brace entry before HERE will be
2877 ;; converted to a cons if the matching close-brace is below HERE.
2878 ;;
2879 ;; START-POINT is a "maximal" "safe position" - there must be no open
2880 ;; parens/braces/brackets between START-POINT and HERE.
2881 ;;
2882 ;; As a second thing, calculate the result of parse-partial-sexp at
2883 ;; PPS-POINT, w.r.t. START-POINT. The motivation here is that
2884 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2885 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2886 ;; needs to be FAST).
2887 ;;
2888 ;; Return a list (GOOD-POS SCAN-BACK-POS CONS-SEPARATED PPS-STATE), where
2889 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2890 ;; to be good (we aim for this to be as high as possible);
2891 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2892 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2893 ;; position to scan backwards from. It is the position of the "{" of the
2894 ;; last element to be removed from `c-state-cache', when that elt is a
2895 ;; cons, otherwise nil.
2896 ;; o - CONS-SEPARATED is t when a cons element in `c-state-cache' has been
2897 ;; replaced by its car because HERE lies inside the brace pair represented
2898 ;; by the cons.
2899 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2909 ; start-point or nil
2916 ; HERE), or nil.
2922 pps-state ; Will be 9 or 10 elements long.
2923 pos
2924 upper-lim ; ,beyond which `c-state-cache' entries are removed
2925 scan-back-pos
2926 cons-separated
2927 pair-beg pps-point-state target-depth)
2929 ;; Remove entries beyond HERE. Also remove any entries inside
2930 ;; a macro, unless HERE is in the same macro.
2935 here
2941 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2942 ;; RBrace and indicate we'll need to search backwards for a previous
2943 ;; brace pair.
2949 cons-separated t))
2951 ;; The next loop jumps forward out of a nested level of parens each
2952 ;; time round; the corresponding elements in `c-state-cache' are
2953 ;; removed. `pos' is just after the brace-pair or the open paren at
2954 ;; (car c-state-cache). There can be no open parens/braces/brackets
2955 ;; between `start-point'/`start-point-actual-macro-start' and HERE,
2956 ;; due to the interface spec to this function.
2959 in-macro-start)))
2961 ; marked by a `category' text property.
2962 start-point))
2973 ;; Do nothing when we've merely reached pps-point.
2974 )
2976 ;; Scan!
2980 target-depth
2981 nil pps-state))
2991 ;; We've just moved out of the paren pair containing the brace-pair
2992 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2993 ;; and is potentially where the open brace of a cons in
2994 ;; c-state-cache will be.
3000 ; containing Lparen
3004 ; together with enclosed brace pair.
3005 ;; (t nil) ; Ignore an unmated Rparen.
3006 )))
3011 nil nil ; TARGETDEPTH, STOPBEFORE
3012 pps-state)))
3014 ;; If the last paren pair we moved out of was actually a brace pair,
3015 ;; insert it into `c-state-cache'.
3020 c-state-cache)))
3025 ;; Strip stale elements of `c-state-cache' by moving backwards through the
3026 ;; buffer, and inform the caller of the scenario detected.
3027 ;;
3028 ;; HERE is the position we're setting `c-state-cache' for.
3029 ;; CACHE-POS (a locally bound variable) is just after the latest recorded
3030 ;; position in `c-state-cache' before HERE, or a position at or near
3031 ;; point-min which isn't in a literal.
3032 ;;
3033 ;; This function must only be called only when (> `c-state-cache-good-pos'
3034 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
3035 ;; optimized to eliminate (or minimize) scanning between these two
3036 ;; positions.
3037 ;;
3038 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
3039 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
3040 ;; could become so after missing elements are inserted into
3041 ;; `c-state-cache'. This is JUST AFTER an opening or closing
3042 ;; brace/paren/bracket which is already in `c-state-cache' or just before
3043 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
3044 ;; before `here''s line, or the start of the literal containing it.
3045 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
3046 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
3047 ;; to scan backwards from.
3048 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
3049 ;; POS and HERE which aren't recorded in `c-state-cache'.
3050 ;;
3051 ;; The comments in this defun use "paren" to mean parenthesis or square
3052 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
3053 ;;
3054 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
3055 ;; | | | | | |
3056 ;; CP E here D C good
3059 pa ren ; positions of "(" and ")"
3060 dropped-cons ; whether the last element dropped from `c-state-cache'
3061 ; was a cons (representing a brace-pair)
3062 good-pos ; see above.
3063 lit ; (START . END) of a literal containing some point.
3064 here-lit-start here-lit-end ; bounds of literal containing `here'
3065 ; or `here' itself.
3070 ;; Remove completely irrelevant entries from `c-state-cache'.
3076 ;; At this stage, (>= pos here);
3077 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3082 ;; CASE 1: The top of the cache is a brace pair which now encloses
3083 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3084 ;; knowledge of what's inside these braces, we have no alternative but
3085 ;; to direct the caller to scan the buffer from the opening brace.
3089 ; entry into a { entry, so the caller needs to
3090 ; search for a brace pair before the {.
3092 ;; `here' might be inside a literal. Check for this.
3097 ;; Has `here' just "newly entered" a macro?
3110 ;; `here' might be nested inside any depth of parens (or brackets but
3111 ;; not braces). Scan backwards to find the outermost such opening
3112 ;; paren, if there is one. This will be the scan position to return.
3119 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3120 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3121 ;; a brace pair preceding this, it will already be in `c-state-cache',
3122 ;; unless there was a brace pair after it, i.e. there'll only be one to
3123 ;; scan for if we've just deleted one.
3126 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3127 ;; Further forward scanning isn't needed, but we still need to find a
3128 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3141 ;; CASE 3: After a }/)/] before `here''s BOL.
3145 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3146 ;; literal containing it.
3151 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3152 ;; Externally visible routines.
3156 c-state-cache-good-pos 1
3157 c-state-nonlit-pos-cache nil
3158 c-state-nonlit-pos-cache-limit 1
3159 c-state-semi-nonlit-pos-cache nil
3160 c-state-semi-nonlit-pos-cache-limit 1
3161 c-state-brace-pair-desert nil
3162 c-state-point-min 1
3163 c-state-point-min-lit-type nil
3164 c-state-point-min-lit-start nil
3165 c-state-min-scan-pos 1
3166 c-state-old-cpp-beg nil
3167 c-state-old-cpp-end nil)
3170 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3171 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3172 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3173 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3174 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3175 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3176 ;; (defun c-state-dump ()
3177 ;; ;; For debugging.
3178 ;; ;(message
3179 ;; (concat
3180 ;; (c-sc-qde c-state-cache)
3181 ;; (c-sc-de c-state-cache-good-pos)
3182 ;; (c-sc-qde c-state-nonlit-pos-cache)
3183 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3184 ;; (c-sc-qde c-state-brace-pair-desert)
3185 ;; (c-sc-de c-state-point-min)
3186 ;; (c-sc-de c-state-point-min-lit-type)
3187 ;; (c-sc-de c-state-point-min-lit-start)
3188 ;; (c-sc-de c-state-min-scan-pos)
3189 ;; (c-sc-de c-state-old-cpp-beg)
3190 ;; (c-sc-de c-state-old-cpp-end)))
3191 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3194 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3195 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3196 ;; left in a consistent state.
3197 ;;
3198 ;; This is much like `c-whack-state-after', but it never changes a paren
3199 ;; pair element into an open paren element. Doing that would mean that the
3200 ;; new open paren wouldn't have the required preceding paren pair element.
3201 ;;
3202 ;; This function is called from c-after-change.
3204 ;; The caches of non-literals:
3205 ;; Note that we use "<=" for the possibility of the second char of a two-char
3206 ;; comment opener being typed; this would invalidate any cache position at
3207 ;; HERE.
3213 ;; `c-state-cache':
3214 ;; Case 1: if `here' is in a literal containing point-min, everything
3215 ;; becomes (or is already) nil.
3219 c-state-cache-good-pos nil
3220 c-state-min-scan-pos nil)
3222 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3223 ;; below `here'. To maintain its consistency, we may need to insert a new
3224 ;; brace pair.
3227 too-high-pa ; recorded {/(/[ next above here, or nil.
3228 dropped-cons ; was the last removed element a brace pair?
3229 pa)
3230 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3237 ;; Do we need to add in an earlier brace pair, having lopped one off?
3244 ;; The brace-pair desert marker:
3252 ;; Find and record all noteworthy parens between some good point earlier in
3253 ;; the file and point. That good point is at least the beginning of the
3254 ;; top-level construct we are in, or the beginning of the preceding
3255 ;; top-level construct if we aren't in one.
3256 ;;
3257 ;; The returned value is a list of the noteworthy parens with the last one
3258 ;; first. If an element in the list is an integer, it's the position of an
3259 ;; open paren (of any type) which has not been closed before the point. If
3260 ;; an element is a cons, it gives the position of a closed BRACE paren
3261 ;; pair[*]; the car is the start brace position and the cdr is the position
3262 ;; following the closing brace. Only the last closed brace paren pair
3263 ;; before each open paren and before the point is recorded, and thus the
3264 ;; state never contains two cons elements in succession. When a close brace
3265 ;; has no matching open brace (e.g., the matching brace is outside the
3266 ;; visible region), it is not represented in the returned value.
3267 ;;
3268 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3269 ;; This defun explicitly treats mismatching parens/braces/brackets as
3270 ;; matching. It is the open brace which makes it a "brace" pair.
3271 ;;
3272 ;; If POINT is within a macro, open parens and brace pairs within
3273 ;; THIS macro MIGHT be recorded. This depends on whether their
3274 ;; syntactic properties have been suppressed by
3275 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3276 ;;
3277 ;; Currently no characters which are given paren syntax with the
3278 ;; syntax-table property are recorded, i.e. angle bracket arglist
3279 ;; parens are never present here. Note that this might change.
3280 ;;
3281 ;; BUG: This function doesn't cope entirely well with unbalanced
3282 ;; parens in macros. (2008-12-11: this has probably been resolved
3283 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3284 ;; following case the brace before the macro isn't balanced with the
3285 ;; one after it:
3286 ;;
3287 ;; {
3288 ;; #define X {
3289 ;; }
3290 ;;
3291 ;; Note to maintainers: this function DOES get called with point
3292 ;; within comments and strings, so don't assume it doesn't!
3293 ;;
3294 ;; This function might do hidden buffer changes.
3297 open-paren-in-column-0-is-defun-start
3298 strategy ; 'forward, 'backward etc..
3299 ;; Candidate positions to start scanning from:
3300 cache-pos ; highest position below HERE already existing in
3301 ; cache (or 1).
3302 good-pos
3303 start-point ; (when scanning forward) a place below HERE where there
3304 ; are no open parens/braces between it and HERE.
3305 bopl-state
3306 res
3307 cons-separated
3309 ;; If POINT-MIN has changed, adjust the cache
3313 ;; Strategy?
3318 ;; SCAN!
3326 ; start-point)
3336 good-pos)))
3348 good-pos)))
3352 c-state-cache-good-pos nil))))
3354 c-state-cache)
3357 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3358 ;;
3359 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3360 ;; of all parens in preprocessor constructs, except for any such construct
3361 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3362 ;; worrying further about macros and template delimiters.
3364 ;; Emacs
3369 c-state-old-cpp-beg
3374 ;; XEmacs
3378 ;; If PLACE is non-nil, return a marker marking it, otherwise nil.
3379 ;; We (re)use MARKER.
3385 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3386 ;; description of the functionality and return value.
3387 ;;
3388 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3389 ;; of all parens in preprocessor constructs, except for any such construct
3390 ;; containing point. We can then call `c-parse-state-1' without worrying
3391 ;; further about macros and template delimiters.
3398 here-cpp-beg)
3400 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3401 ;; subsystem.
3404 ;; Emacs
3408 here-cpp-beg here-cpp-end
3412 ;; XEmacs
3416 c-state-old-cpp-end
3419 ;; Debug tool to catch cache inconsistencies. This is called from
3420 ;; 000tests.el.
3440 c-state-cache-good-pos
3441 c-state-nonlit-pos-cache
3442 c-state-nonlit-pos-cache-limit
3443 c-state-semi-nonlit-pos-cache
3444 c-state-semi-nonlit-pos-cache-limit
3445 c-state-brace-pair-desert
3446 c-state-point-min
3447 c-state-point-min-lit-type
3448 c-state-point-min-lit-start
3449 c-state-min-scan-pos
3450 c-state-old-cpp-beg
3451 c-state-old-cpp-end
3452 c-parse-state-point))))
3474 conses-not-ok))
3491 ;; The cache can actually go further back due to the ad-hoc way
3492 ;; the first paren is found, so try to whack off a bit of its
3493 ;; start before complaining.
3494 ;; (save-excursion
3495 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3496 ;; (c-beginning-of-defun-1)
3497 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3498 ;; (c-beginning-of-defun-1))
3499 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3500 ;; (message (concat "c-parse-state inconsistency at %s: "
3501 ;; "using cache: %s, from scratch: %s")
3502 ;; here res1 res2)))
3505 here res1 res2)
3511 here res1)
3516 res2 ; res1 correct a cascading series of errors ASAP
3517 ))
3523 'c-debug-parse-state
3531 \f
3533 ;; Whack off any state information from PAREN-STATE which lies
3534 ;; before BUFPOS. Not destructive on PAREN-STATE.
3537 car)
3548 ;; Whack off any state information from PAREN-STATE which lies at or
3549 ;; after BUFPOS. Not destructive on PAREN-STATE.
3554 ;; just check the car, because in a balanced brace
3555 ;; expression, it must be impossible for the corresponding
3556 ;; close brace to be before point, but the open brace to
3557 ;; be after.
3559 nil ; whack it off
3561 ;; its possible that the open brace is before
3562 ;; bufpos, but the close brace is after. In that
3563 ;; case, convert this to a non-cons element. The
3564 ;; rest of the state is before bufpos, so we're
3565 ;; done.
3567 ;; we know that both the open and close braces are
3568 ;; before bufpos, so we also know that everything else
3569 ;; on state is before bufpos.
3572 nil ; whack it off
3573 ;; it's before bufpos, so everything else should too.
3576 nil)))
3579 ;; Return the bufpos of the innermost enclosing open paren before
3580 ;; bufpos, or nil if none was found.
3590 enclosingp))
3593 ;; Return the bufpos of the outermost enclosing open paren, or nil
3594 ;; if none was found.
3601 pos))
3604 ;; Return the closest "safe" position recorded on PAREN-STATE that
3605 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3606 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3607 ;; find the closest limit before a given limit that might be nil.
3608 ;;
3609 ;; A "safe" position is a position at or after a recorded open
3610 ;; paren, or after a recorded close paren. The returned position is
3611 ;; thus either the first position after a close brace, or the first
3612 ;; position after an enclosing paren, or at the enclosing paren in
3613 ;; case BUFPOS is immediately after it.
3623 ;; See below.
3626 ;; elem is the position at and not after the opening paren, so
3627 ;; we can go forward one more step unless it's equal to
3628 ;; bufpos. This is useful in some cases avoid an extra paren
3629 ;; level between the safe position and bufpos.
3634 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3635 ;; goes to the closest previous point that is known to be outside
3636 ;; any string literal or comment. `c-state-cache' is used if it has
3637 ;; a position in the vicinity.
3639 elem
3642 ;; Note: Similar code in `c-safe-position'. The
3643 ;; difference is that we accept a safe position at
3644 ;; the point and don't bother to go forward past open
3645 ;; parens.
3660 ;; The position is far back. Try `c-beginning-of-defun-1'
3661 ;; (although we can't be entirely sure it will go to a position
3662 ;; outside a comment or string in current emacsen). FIXME:
3663 ;; Consult `syntax-ppss' here.
3668 \f
3669 ;; Tools for scanning identifiers and other tokens.
3672 "Return non-nil if the point is on or directly after an identifier.
3673 Keywords are recognized and not considered identifiers. If an
3674 identifier is detected, the returned value is its starting position.
3675 If an identifier ends at the point and another begins at it \(can only
3676 happen in Pike) then the point for the preceding one is returned.
3678 Note that this function might do hidden buffer changes. See the
3679 comment at the start of cc-engine.el for more info."
3681 ;; FIXME: Shouldn't this function handle "operator" in C++?
3688 ;; Check for a normal (non-keyword) identifier.
3694 ;; Handle the `<operator> syntax in Pike.
3698 t
3705 ;; Handle the "operator +" syntax in C++.
3722 )))
3725 ;; If the point is at the end of a symbol then skip backward to the
3726 ;; beginning of it. Don't move otherwise. Return non-nil if point
3727 ;; moved.
3728 ;;
3729 ;; This function might do hidden buffer changes.
3732 ;; Handle the `<operator> syntax in Pike.
3738 t
3740 nil)))))
3743 ;; Move to the beginning of the current token. Do not move if not
3744 ;; in the middle of one. BACK-LIMIT may be used to bound the
3745 ;; backward search; if given it's assumed to be at the boundary
3746 ;; between two tokens. Return non-nil if the point is moved, nil
3747 ;; otherwise.
3748 ;;
3749 ;; This function might do hidden buffer changes.
3756 ;; `c-nonsymbol-token-regexp' should always match
3757 ;; since we've skipped backward over punctuation
3758 ;; or paren syntax, but consume one char in case
3759 ;; it doesn't so that we don't leave point before
3760 ;; some earlier incorrect token.
3767 ;; Move to the end of the current token. Do not move if not in the
3768 ;; middle of one. BACK-LIMIT may be used to bound the backward
3769 ;; search; if given it's assumed to be at the boundary between two
3770 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3771 ;;
3772 ;; This function might do hidden buffer changes.
3780 ;; `c-nonsymbol-token-regexp' should always match since
3781 ;; we've skipped backward over punctuation or paren
3782 ;; syntax, but move forward in case it doesn't so that
3783 ;; we don't leave point earlier than we started with.
3799 "Move forward by tokens.
3800 A token is defined as all symbols and identifiers which aren't
3802 treated properly). Point is always either left at the beginning of a
3803 token or not moved at all. COUNT specifies the number of tokens to
3804 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3805 moves to the next token beginning only if not already at one. If
3806 BALANCED is true, move over balanced parens, otherwise move into them.
3807 Also, if BALANCED is true, never move out of an enclosing paren.
3809 LIMIT sets the limit for the movement and defaults to the point limit.
3810 The case when LIMIT is set in the middle of a token, comment or macro
3811 is handled correctly, i.e. the point won't be left there.
3813 Return the number of tokens left to move \(positive or negative). If
3814 BALANCED is true, a move over a balanced paren counts as one. Note
3815 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3816 be returned. Thus, a return value of 0 guarantees that point is at
3817 the requested position and a return value less \(without signs) than
3818 COUNT guarantees that point is at the beginning of some token.
3820 Note that this function might do hidden buffer changes. See the
3821 comment at the start of cc-engine.el for more info."
3828 c-jump-syntax-balanced
3829 c-jump-syntax-unbalanced))
3834 ;; If count is zero we should jump if in the middle of a token.
3841 ;; Skip whitespace. Count this as a move if we did in
3842 ;; fact move.
3846 ;; Moved out of bounds. Make sure the returned count isn't zero.
3851 ;; Use `condition-case' to avoid having the limit tests
3852 ;; inside the loop.
3860 t)
3863 t)
3864 ;; `c-nonsymbol-token-regexp' above should always
3865 ;; match if there are correct tokens. Try to
3866 ;; widen to see if the limit was set in the
3867 ;; middle of one, else fall back to treating
3868 ;; the offending thing as a one character token.
3873 nil)
3876 t))))
3886 count)))
3889 "Move backward by tokens.
3890 See `c-forward-token-2' for details."
3898 c-jump-syntax-balanced
3899 c-jump-syntax-unbalanced))
3903 ;; The count is zero so try to skip to the beginning of the
3904 ;; current token.
3908 ;; The limit is inside the same token, so return 1.
3911 ;; We're not in the middle of a token. If there's
3912 ;; whitespace after the point then we must move backward,
3913 ;; so set count to 1 in that case.
3915 ;; If we're looking at a '#' that might start a cpp
3916 ;; directive then we have to do a more elaborate check.
3929 ;; Use `condition-case' to avoid having to check for buffer
3930 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3939 ;; This can be very inefficient if there's a long
3940 ;; sequence of operator tokens without any separation.
3941 ;; That doesn't happen in practice, anyway.
3951 count)))
3954 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3956 characters are jumped over character by character. This function is
3957 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3962 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3964 characters are jumped over character by character. This function is
3965 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3969 \f
3970 ;; Tools for doing searches restricted to syntactically relevant text.
3973 paren-level not-inside-token
3974 lookbehind-submatch)
3975 "Like `re-search-forward', but only report matches that are found
3976 in syntactically significant text. I.e. matches in comments, macros
3977 or string literals are ignored. The start point is assumed to be
3978 outside any comment, macro or string literal, or else the content of
3979 that region is taken as syntactically significant text.
3981 If PAREN-LEVEL is non-nil, an additional restriction is added to
3982 ignore matches in nested paren sexps. The search will also not go
3983 outside the current list sexp, which has the effect that if the point
3984 should be moved to BOUND when no match is found \(i.e. NOERROR is
3985 neither nil nor t), then it will be at the closing paren if the end of
3986 the current list sexp is encountered first.
3988 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3989 ignored. Things like multicharacter operators and special symbols
3991 constants.
3993 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3994 subexpression in REGEXP. The end of that submatch is used as the
3995 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
3996 isn't used or if that subexpression didn't match then the start
3998 subexpression is never tested before the starting position, so it
4002 subexpression can match the end of nonwhite syntactic whitespace,
4003 i.e. the end of comments or cpp directives. This since the function
4004 skips over such things before resuming the search. It's on the other
4006 matches syntactic whitespace.
4008 Bug: Unbalanced parens inside cpp directives are currently not handled
4009 correctly \(i.e. they don't get ignored as they should) when
4010 PAREN-LEVEL is set.
4012 Note that this function might do hidden buffer changes. See the
4013 comment at the start of cc-engine.el for more info."
4018 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
4021 tmp
4022 ;; Start position for the last search.
4023 search-pos
4024 ;; The `parse-partial-sexp' state between the start position
4025 ;; and the point.
4026 state
4027 ;; The current position after the last state update. The next
4028 ;; `parse-partial-sexp' continues from here.
4030 ;; The position at which to check the state and the state
4031 ;; there. This is separate from `state-pos' since we might
4032 ;; need to back up before doing the next search round.
4033 check-pos check-state
4034 ;; Last position known to end a token.
4036 ;; Set when a valid match is found.
4037 found)
4055 state-pos check-pos paren-level nil state))
4057 check-state state))
4059 ;; NOTE: If we got a look behind subexpression and get
4060 ;; an insignificant match in something that isn't
4061 ;; syntactic whitespace (i.e. strings or in nested
4062 ;; parentheses), then we can never skip more than a
4063 ;; single character from the match start position
4064 ;; (i.e. `state-pos' here) before continuing the
4065 ;; search. That since the look behind subexpression
4066 ;; might match the end of the insignificant region in
4067 ;; the next search.
4071 ;; Match inside a line comment. Skip to eol. Use
4072 ;; `re-search-forward' instead of `skip-chars-forward' to get
4073 ;; the right bound behavior.
4077 ;; Match inside a block comment. Skip to the '*/'.
4084 ;; Match in the middle of the opener of a block or line
4085 ;; comment.
4090 ;; The last `parse-partial-sexp' above might have
4091 ;; stopped short of the real check position if the end
4092 ;; of the current sexp was encountered in paren-level
4093 ;; mode. The checks above are always false in that
4094 ;; case, and since they can do better skipping in
4095 ;; lookbehind-submatch mode, we do them before
4096 ;; checking the paren level.
4100 ;; Check the paren level first since we're short of the
4101 ;; syntactic checking position if the end of the
4102 ;; current sexp was encountered by `parse-partial-sexp'.
4105 ;; Inside a nested paren sexp.
4107 ;; See the NOTE above.
4109 ;; Skip out of the paren quickly.
4113 ;; Have exited the current paren sexp.
4116 ;; The last `parse-partial-sexp' call above
4117 ;; has left us just after the closing paren
4118 ;; in this case, so we can modify the bound
4119 ;; to leave the point at the right position
4120 ;; upon return.
4122 nil)
4126 ;; Match inside a string.
4129 ;; See the NOTE above.
4131 ;; Skip to the end of the string before continuing.
4140 continue)))
4145 ;; Match inside a macro. Skip to the end of it.
4153 (< check-pos
4159 ;; Inside a token.
4161 ;; See the NOTE above.
4166 ;; A real match.
4168 nil)))
4170 ;; Should loop to search again, but take care to avoid
4171 ;; looping on the same spot.
4175 nil
4178 t))))
4184 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4191 ;; Search failed. Set point as appropriate.
4195 nil)))
4200 ;; Return the start of the literal point is in, or nil.
4201 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4202 ;; bound in the caller.
4204 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4205 ;; if it's outside comments and strings.
4208 ;; Pick a safe position as close to the point as possible.
4209 ;;
4210 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4211 ;; position.
4224 ;; Cache positions along the way to use if we have to back up more. We
4225 ;; cache every closing paren on the same level. If the paren cache is
4226 ;; relevant in this region then we're typically already on the same
4227 ;; level as the target position. Note that we might cache positions
4228 ;; after opening parens in case safe-pos is in a nested list. That's
4229 ;; both uncommon and harmless.
4237 ;; If the state contains the start of the containing sexp we cache that
4238 ;; position too, so that parse-partial-sexp in the next run has a bigger
4239 ;; chance of starting at the same level as the target position and thus
4240 ;; will get more good safe positions into the list.
4246 ;; Inside string or comment. Continue search at the
4247 ;; beginning of it.
4251 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4252 i.e. don't stop at positions inside syntactic whitespace or string
4253 literals. Preprocessor directives are also ignored, with the exception
4254 of the one that the point starts within, if any. If LIMIT is given,
4255 it's assumed to be at a syntactically relevant position.
4257 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4258 sexps, and the search will also not go outside the current paren sexp.
4259 However, if LIMIT or the buffer limit is reached inside a nested paren
4260 then the point will be left at the limit.
4262 Non-nil is returned if the point moved, nil otherwise.
4264 Note that this function might do hidden buffer changes. See the
4265 comment at the start of cc-engine.el for more info."
4268 state-2
4269 ;; A list of syntactically relevant positions in descending
4270 ;; order. It's used to avoid scanning repeatedly over
4271 ;; potentially large regions with `parse-partial-sexp' to verify
4272 ;; each position. Used in `c-ssb-lit-begin'
4273 safe-pos-list
4274 ;; The result from `c-beginning-of-macro' at the start position or the
4275 ;; start position itself if it isn't within a macro. Evaluated on
4276 ;; demand.
4277 start-macro-beg
4278 ;; The earliest position after the current one with the same paren
4279 ;; level. Used only when `paren-level' is set.
4280 lit-beg
4285 ;; The next loop "tries" to find the end point each time round,
4286 ;; loops when it hasn't succeeded.
4292 ;; Don't stop inside a literal.
4295 t)))
4307 ;; Not at the right level.
4310 ;; We stop above if we go out of a paren.
4311 ;; Now check whether it precedes or is
4312 ;; nested in the starting sexp.
4316 pps-end-pos paren-level-pos
4317 nil nil state-2))
4320 ;; We've stopped short of the starting position
4321 ;; so the hit was inside a nested list. Go up
4322 ;; until we are at the right level.
4331 nil)
4332 t))
4335 nil))
4337 ;; The hit was outside the list at the start
4338 ;; position. Go to the start of the list and exit.
4340 nil))
4343 ;; Inside a macro.
4351 t
4353 ;; It's inside the same macro we started in so it's
4354 ;; a relevant match.
4356 nil))))))
4360 ;; Skip syntactic ws afterwards so that we don't stop at the
4361 ;; end of a comment if `skip-chars' is something like "^/".
4365 ;; We might want to extend this with more useful return values in
4366 ;; the future.
4369 ;; The following is an alternative implementation of
4370 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4371 ;; track of the syntactic context. It turned out to be generally
4372 ;; slower than the one above which uses forward checks from earlier
4373 ;; safe positions.
4374 ;;
4375 ;;(defconst c-ssb-stop-re
4376 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4377 ;; ;; stop at to avoid going into comments and literals.
4378 ;; (concat
4379 ;; ;; Match comment end syntax and string literal syntax. Also match
4380 ;; ;; '/' for block comment endings (not covered by comment end
4381 ;; ;; syntax).
4382 ;; "\\s>\\|/\\|\\s\""
4383 ;; (if (memq 'gen-string-delim c-emacs-features)
4384 ;; "\\|\\s|"
4385 ;; "")
4386 ;; (if (memq 'gen-comment-delim c-emacs-features)
4387 ;; "\\|\\s!"
4388 ;; "")))
4389 ;;
4390 ;;(defconst c-ssb-stop-paren-re
4391 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4392 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4393 ;;
4394 ;;(defconst c-ssb-sexp-end-re
4395 ;; ;; Regexp matching the ending syntax of a complex sexp.
4396 ;; (concat c-string-limit-regexp "\\|\\s)"))
4397 ;;
4398 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4399 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4400 ;;i.e. don't stop at positions inside syntactic whitespace or string
4401 ;;literals. Preprocessor directives are also ignored. However, if the
4402 ;;point is within a comment, string literal or preprocessor directory to
4403 ;;begin with, its contents is treated as syntactically relevant chars.
4404 ;;If LIMIT is given, it limits the backward search and the point will be
4405 ;;left there if no earlier position is found.
4406 ;;
4407 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4408 ;;sexps, and the search will also not go outside the current paren sexp.
4409 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4410 ;;then the point will be left at the limit.
4411 ;;
4412 ;;Non-nil is returned if the point moved, nil otherwise.
4413 ;;
4414 ;;Note that this function might do hidden buffer changes. See the
4415 ;;comment at the start of cc-engine.el for more info."
4416 ;;
4417 ;; (save-restriction
4418 ;; (when limit
4419 ;; (narrow-to-region limit (point-max)))
4420 ;;
4421 ;; (let ((start (point)))
4422 ;; (catch 'done
4423 ;; (while (let ((last-pos (point))
4424 ;; (stop-pos (progn
4425 ;; (skip-chars-backward skip-chars)
4426 ;; (point))))
4427 ;;
4428 ;; ;; Skip back over the same region as
4429 ;; ;; `skip-chars-backward' above, but keep to
4430 ;; ;; syntactically relevant positions.
4431 ;; (goto-char last-pos)
4432 ;; (while (and
4433 ;; ;; `re-search-backward' with a single char regexp
4434 ;; ;; should be fast.
4435 ;; (re-search-backward
4436 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4437 ;; stop-pos 'move)
4438 ;;
4439 ;; (progn
4440 ;; (cond
4441 ;; ((looking-at "\\s(")
4442 ;; ;; `paren-level' is set and we've found the
4443 ;; ;; start of the containing paren.
4444 ;; (forward-char)
4445 ;; (throw 'done t))
4446 ;;
4447 ;; ((looking-at c-ssb-sexp-end-re)
4448 ;; ;; We're at the end of a string literal or paren
4449 ;; ;; sexp (if `paren-level' is set).
4450 ;; (forward-char)
4451 ;; (condition-case nil
4452 ;; (c-backward-sexp)
4453 ;; (error
4454 ;; (goto-char limit)
4455 ;; (throw 'done t))))
4456 ;;
4457 ;; (t
4458 ;; (forward-char)
4459 ;; ;; At the end of some syntactic ws or possibly
4460 ;; ;; after a plain '/' operator.
4461 ;; (let ((pos (point)))
4462 ;; (c-backward-syntactic-ws)
4463 ;; (if (= pos (point))
4464 ;; ;; Was a plain '/' operator. Go past it.
4465 ;; (backward-char)))))
4466 ;;
4467 ;; (> (point) stop-pos))))
4468 ;;
4469 ;; ;; Now the point is either at `stop-pos' or at some
4470 ;; ;; position further back if `stop-pos' was at a
4471 ;; ;; syntactically irrelevant place.
4472 ;;
4473 ;; ;; Skip additional syntactic ws so that we don't stop
4474 ;; ;; at the end of a comment if `skip-chars' is
4475 ;; ;; something like "^/".
4476 ;; (c-backward-syntactic-ws)
4477 ;;
4478 ;; (< (point) stop-pos))))
4479 ;;
4480 ;; ;; We might want to extend this with more useful return values
4481 ;; ;; in the future.
4482 ;; (/= (point) start))))
4484 \f
4485 ;; Tools for handling comments and string literals.
4488 "Return the type of literal point is in, if any.
4489 The return value is `c' if in a C-style comment, `c++' if in a C++
4490 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4491 is non-nil and in a preprocessor line, or nil if somewhere else.
4492 Optional LIM is used as the backward limit of the search. If omitted,
4493 or nil, `c-beginning-of-defun' is used.
4495 The last point calculated is cached if the cache is enabled, i.e. if
4496 `c-in-literal-cache' is bound to a two element vector.
4498 Note that this function might do hidden buffer changes. See the
4499 comment at the start of cc-engine.el for more info."
4510 "Return a cons of the beginning and end positions of the comment or
4511 string surrounding point (including both delimiters), or nil if point
4513 to start parsing from. If NEAR is non-nil, then the limits of any
4515 spaces and tabs between point and the literal. The search for such a
4516 literal is done first in forward direction. If NOT-IN-DELIMITER is
4517 non-nil, the case when point is inside a starting delimiter won't be
4518 recognized. This only has effect for comments which have starting
4519 delimiters with more than one character.
4521 Note that this function might do hidden buffer changes. See the
4522 comment at the start of cc-engine.el for more info."
4538 ;; Search forward for a literal.
4549 ;; Search backward.
4560 ;; Block comment. Due to the nature of line
4561 ;; comments, they will always be covered by the
4562 ;; normal case above.
4565 ;; If LIM is bogus, beg will be bogus.
4569 ))))
4571 ;; In case external callers use this; it did have a docstring.
4575 "If the argument is a cons of two buffer positions (such as returned by
4576 `c-literal-limits'), and that range contains a C++ style line comment,
4577 then an extended range is returned that contains all adjacent line
4578 comments (i.e. all comments that starts in the same column with no
4579 empty lines or non-whitespace characters between them). Otherwise the
4580 argument is returned.
4582 Note that this function might do hidden buffer changes. See the
4583 comment at the start of cc-engine.el for more info."
4594 ;; Got to take care in the backward direction to handle
4595 ;; comments which are preceded by code.
4609 range)
4613 "Convenience function that given the result of `c-literal-limits',
4614 returns nil or the type of literal that the range surrounds, one
4615 of the symbols 'c, 'c++ or 'string. It's much faster than using
4616 `c-in-literal' and is intended to be used when you need both the
4617 type of a literal and its limits.
4619 Note that this function might do hidden buffer changes. See the
4620 comment at the start of cc-engine.el for more info."
4631 range))
4634 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4635 ;; This doesn't preserve point.
4644 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4645 ;; (default point). This is done by going back further in the buffer then
4646 ;; searching forward for literals. The position found won't be in a
4647 ;; literal. We start searching for the sought position TRY-SIZE (default
4648 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4649 ;; :-)
4657 stack elt size
4660 ;; Move forward one literal each time round this loop.
4661 ;; Move forward to the start of a comment or string.
4663 pos
4664 start
4665 nil ; target-depth
4666 nil ; stop-before
4667 s ; state
4670 ;; Gather details of the non-literal-bit - starting pos and size.
4674 pos))
4678 ;; Move forward to the end of the comment/string.
4682 start
4683 nil ; target-depth
4684 nil ; stop-before
4685 s ; state
4689 ;; Now try and find enough non-literal characters recorded on the stack.
4690 ;; Go back one recorded literal each time round this loop.
4692 stack)
4697 ;; Have we found enough yet?
4707 ;; Return a buffer position about HOW-FAR non-literal characters forward
4708 ;; from START-POS (default point), which must not be inside a literal.
4715 ;; Scan over counted characters.
4717 pos
4719 nil ; target-depth
4720 nil ; stop-before
4721 s ; state
4725 ;; Scan over literal characters.
4728 pos
4730 nil ; target-depth
4731 nil ; stop-before
4732 s ; state
4737 \f
4738 ;; `c-find-decl-spots' and accompanying stuff.
4740 ;; Variables used in `c-find-decl-spots' to cache the search done for
4741 ;; the first declaration in the last call. When that function starts,
4742 ;; it needs to back up over syntactic whitespace to look at the last
4743 ;; token before the region being searched. That can sometimes cause
4744 ;; moves back and forth over a quite large region of comments and
4745 ;; macros, which would be repeated for each changed character when
4746 ;; we're called during fontification, since font-lock refontifies the
4747 ;; current line for each change. Thus it's worthwhile to cache the
4748 ;; first match.
4749 ;;
4750 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4751 ;; the syntactic whitespace less or equal to some start position.
4752 ;; There's no cached value if it's nil.
4753 ;;
4754 ;; `c-find-decl-match-pos' is the match position if
4755 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4756 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4767 ; (defface c-debug-decl-spot-face
4768 ; '((t (:background "Turquoise")))
4769 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4770 ; (defface c-debug-decl-sws-face
4771 ; '((t (:background "Khaki")))
4772 ; "Debug face to mark the syntactic whitespace between the declaration
4773 ; spots and the preceding token end.")
4789 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4790 ;; but it contains lots of free variables that refer to things
4791 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4792 ;; if there is a match, otherwise at `cfd-limit'.
4793 ;;
4794 ;; The macro moves point forward to the next putative start of a declaration
4795 ;; or cfd-limit. This decl start is the next token after a "declaration
4796 ;; prefix". The declaration prefix is the earlier of `cfd-prop-match' and
4797 ;; `cfd-re-match'. `cfd-match-pos' is set to the decl prefix.
4798 ;;
4799 ;; This macro might do hidden buffer changes.
4802 ;; Find the next property match position if we haven't got one already.
4813 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4814 ;; got one already.
4820 ;; Each time round, the next `while' moves forward over a pseudo match
4821 ;; of `c-decl-prefix-or-start-re' which is either inside a literal, or
4822 ;; is a ":" not preceded by "public", etc.. `cfd-re-match' and
4823 ;; `cfd-re-match-end' get set.
4830 ;; No match. Finish up and exit the loop.
4832 nil)
4835 ;; Matched the end of a token preceding a decl spot.
4839 ;; Matched a token that start a decl spot.
4842 c-literal-faces)
4843 ;; Pseudo match inside a comment or string literal. Skip out
4844 ;; of comments and string literals.
4856 ;; Found a ":" which isn't part of "public:", etc.
4857 t)
4860 ;; If our match was at the decl start, we have to back up over the
4861 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4862 ;; any decl spots in the syntactic ws.
4867 ;; Choose whichever match is closer to the start.
4870 cfd-re-match nil)
4872 cfd-prop-match nil))
4877 ;; Skip forward past comments only so we don't skip macros.
4879 ;; Set the position to continue at. We can avoid going over
4880 ;; the comments skipped above a second time, but it's possible
4881 ;; that the comment skipping has taken us past `cfd-prop-match'
4882 ;; since the property might be used inside comments.
4888 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4889 ;; label from the point to CFD-LIMIT.
4890 ;;
4891 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4892 ;; arguments: The first is the end position of the token preceding the spot,
4893 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4894 ;; the match is inside a macro. Point should be moved forward by at least
4895 ;; one token.
4896 ;;
4897 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4898 ;; it should return non-nil to ensure that the next search will find them.
4899 ;;
4900 ;; Such a spot is:
4901 ;; o The first token after bob.
4902 ;; o The first token after the end of submatch 1 in
4903 ;; `c-decl-prefix-or-start-re' when that submatch matches. This
4904 ;; submatch is typically a (L or R) brace or paren, a ;, or a ,.
4905 ;; o The start of each `c-decl-prefix-or-start-re' match when
4906 ;; submatch 1 doesn't match. This is, for example, the keyword
4907 ;; "class" in Pike.
4908 ;; o The start of a previously recognized declaration; "recognized"
4909 ;; means that the last char of the previous token has a `c-type'
4910 ;; text property with the value `c-decl-end'; this only holds
4911 ;; when `c-type-decl-end-used' is set.
4912 ;;
4913 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4914 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4915 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4916 ;;
4917 ;; If the match is inside a macro then the buffer is narrowed to the
4918 ;; end of it, so that CFD-FUN can investigate the following tokens
4919 ;; without matching something that begins inside a macro and ends
4920 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4921 ;; CFD-FACE-CHECKLIST checks exist.
4922 ;;
4923 ;; The spots are visited approximately in order from top to bottom.
4924 ;; It's however the positions where `c-decl-prefix-or-start-re'
4925 ;; matches and where `c-decl-end' properties are found that are in
4926 ;; order. Since the spots often are at the following token, they
4927 ;; might be visited out of order insofar as more spots are reported
4928 ;; later on within the syntactic whitespace between the match
4929 ;; positions and their spots.
4930 ;;
4931 ;; It's assumed that comments and strings are fontified in the
4932 ;; searched range.
4933 ;;
4934 ;; This is mainly used in fontification, and so has an elaborate
4935 ;; cache to handle repeated calls from the same start position; see
4936 ;; the variables above.
4937 ;;
4938 ;; All variables in this function begin with `cfd-' to avoid name
4939 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4940 ;;
4941 ;; This function might do hidden buffer changes.
4945 ;; The end of the token preceding the decl spot last found
4946 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4947 ;; no match.
4948 cfd-re-match
4949 ;; The end position of the last `c-decl-prefix-or-start-re'
4950 ;; match. If this is greater than `cfd-continue-pos', the
4951 ;; next regexp search is started here instead.
4953 ;; The end of the last `c-decl-end' found by
4954 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4955 ;; match. If searching for the property isn't needed then we
4956 ;; disable it by setting it to `cfd-limit' directly.
4958 ;; The end of the token preceding the decl spot last found by
4959 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4960 ;; bob. `cfd-limit' if there's no match. In other words,
4961 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4963 ;; The position to continue searching at.
4964 cfd-continue-pos
4965 ;; The position of the last "real" token we've stopped at.
4966 ;; This can be greater than `cfd-continue-pos' when we get
4967 ;; hits inside macros or at `c-decl-end' positions inside
4968 ;; comments.
4970 ;; The end position of the last entered macro.
4973 ;; Initialize by finding a syntactically relevant start position
4974 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4975 ;; search unless we're at bob.
4978 ;; Must back up a bit since we look for the end of the previous
4979 ;; statement or declaration, which is earlier than the first
4980 ;; returned match.
4982 ;; This `cond' moves back over any literals or macros. It has special
4983 ;; handling for when the region being searched is entirely within a
4984 ;; macro. It sets `cfd-continue-pos' (unless we've reached
4985 ;; `cfd-limit').
4987 ;; First we need to move to a syntactically relevant position.
4988 ;; Begin by backing out of comment or string literals.
4989 ;;
4990 ;; This arm of the cond actually triggers if we're in a literal,
4991 ;; and cfd-limit is at most at BONL.
4993 ;; This arm of the `and' moves backwards out of a literal when
4994 ;; the face at point is a literal face. In this case, its value
4995 ;; is always non-nil.
4997 ;; Try to use the faces to back up to the start of the
4998 ;; literal. FIXME: What if the point is on a declaration
4999 ;; inside a comment?
5005 ;; XEmacs doesn't fontify the quotes surrounding string
5006 ;; literals.
5015 ;; Don't trust the literal to contain only literal faces
5016 ;; (the font lock package might not have fontified the
5017 ;; start of it at all, for instance) so check that we have
5018 ;; arrived at something that looks like a start or else
5019 ;; resort to `c-literal-limits'.
5026 ;; The start is in a literal. If the limit is in the same
5027 ;; one we don't have to find a syntactic position etc. We
5028 ;; only check that if the limit is at or before bonl to save
5029 ;; time; it covers the by far most common case when font-lock
5030 ;; refontifies the current line only.
5041 ;; Completely inside a literal. Set up variables to trig the
5042 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
5043 ;; find a suitable start position.
5046 ;; Check if the region might be completely inside a macro, to
5047 ;; optimize that like the completely-inside-literal above.
5054 ;; (Maybe) completely inside a macro. Only need to trig the
5055 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
5056 ;; set things up.
5058 start-in-macro t))
5060 ;; The default arm of the `cond' moves back over any macro we're in
5061 ;; and over any syntactic WS. It sets `c-find-decl-syntactic-pos'.
5063 ;; Back out of any macro so we don't miss any declaration
5064 ;; that could follow after it.
5068 ;; Now we're at a proper syntactically relevant position so we
5069 ;; can use the cache. But first clear it if it applied
5070 ;; further down.
5075 ;; Don't have to do this if the cache is relevant here,
5076 ;; typically if the same line is refontified again. If
5077 ;; we're just some syntactic whitespace further down we can
5078 ;; still use the cache to limit the skipping.
5081 ;; If we hit `c-find-decl-syntactic-pos' and
5082 ;; `c-find-decl-match-pos' is set then we install the cached
5083 ;; values. If we hit `c-find-decl-syntactic-pos' and
5084 ;; `c-find-decl-match-pos' is nil then we know there's no decl
5085 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
5086 ;; and so we can continue the search from this point. If we
5087 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
5088 ;; the right spot to begin searching anyway.
5090 c-find-decl-match-pos)
5092 cfd-continue-pos syntactic-pos)
5097 ;; Always consider bob a match to get the first
5098 ;; declaration in the file. Do this separately instead of
5099 ;; letting `c-decl-prefix-or-start-re' match bob, so that
5100 ;; regexp always can consume at least one character to
5101 ;; ensure that we won't get stuck in an infinite loop.
5106 ;; Do an initial search now. In the bob case above it's
5107 ;; only done to search for a `c-decl-end' spot.
5113 ;; Advance `cfd-continue-pos' if it's before the start position.
5114 ;; The closest continue position that might have effect at or
5115 ;; after the start depends on what we started in. This also
5116 ;; finds a suitable start position in the special cases when the
5117 ;; region is completely within a literal or macro.
5122 ;; If we're in a macro then it's the closest preceding token
5123 ;; in the macro. Check this before `start-in-literal',
5124 ;; since if we're inside a literal in a macro, the preceding
5125 ;; token is earlier than any `c-decl-end' spot inside the
5126 ;; literal (comment).
5128 ;; The only syntactic ws in macros are comments.
5134 ;; If we're in a comment it can only be the closest
5135 ;; preceding `c-decl-end' position within that comment, if
5136 ;; any. Go back to the beginning of such a property so that
5137 ;; `c-find-decl-prefix-search' will find the end of it.
5138 ;; (Can't stop at the end and install it directly on
5139 ;; `cfd-prop-match' since that variable might be cleared
5140 ;; after `cfd-fun' below.)
5141 ;;
5142 ;; Note that if the literal is a string then the property
5143 ;; search will simply skip to the beginning of it right
5144 ;; away.
5156 ;; Didn't find any property inside the comment, so we can
5157 ;; skip it entirely. (This won't skip past a string, but
5158 ;; that'll be handled quickly by the next
5159 ;; `c-find-decl-prefix-search' anyway.)
5165 ;; If we started in normal code, the only match that might
5166 ;; apply before the start is what we already got in
5167 ;; `cfd-match-pos' so we can continue at the start position.
5168 ;; (Note that we don't get here if the first match is below
5169 ;; it.)
5172 ;; Delete found matches if they are before our new continue
5173 ;; position, so that `c-find-decl-prefix-search' won't back up
5174 ;; to them later on.
5182 ;; This is the normal case and we got a proper syntactic
5183 ;; position. If there's a match then it's always outside
5184 ;; macros and comments, so advance to the next token and set
5185 ;; `cfd-token-pos'. The loop below will later go back using
5186 ;; `cfd-continue-pos' to fix declarations inside the
5187 ;; syntactic ws.
5195 ;; Have one of the special cases when the region is completely
5196 ;; within a literal or macro. `cfd-continue-pos' is set to a
5197 ;; good start position for the search, so do it.
5200 ;; Now loop, one decl spot per iteration. We already have the first
5201 ;; match in `cfd-match-pos'.
5203 ;; Go forward over "false matches", one per iteration.
5208 ;; Kludge to filter out matches on the "<" that
5209 ;; aren't open parens, for the sake of languages
5210 ;; that got `c-recognize-<>-arglists' set.
5215 ;; If `cfd-continue-pos' is less or equal to
5216 ;; `cfd-token-pos', we've got a hit inside a macro
5217 ;; that's in the syntactic whitespace before the last
5218 ;; "real" declaration we've checked. If they're equal
5219 ;; we've arrived at the declaration a second time, so
5220 ;; there's nothing to do.
5224 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5225 ;; we're still searching for declarations embedded in
5226 ;; the syntactic whitespace. In that case we need
5227 ;; only to skip comments and not macros, since they
5228 ;; can't be nested, and that's already been done in
5229 ;; `c-find-decl-prefix-search'.
5234 ;; Continue if the following token fails the
5235 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5242 t)))
5253 ;; Narrow to the end of the macro if we got a hit inside
5254 ;; one, to avoid recognizing things that start inside the
5255 ;; macro and end outside it.
5257 ;; Not in the same macro as in the previous round.
5268 t
5271 t)
5272 ;; The matched token was the last thing in the macro,
5273 ;; so the whole match is bogus.
5282 ;; Restore limits if we did macro narrowing above.
5289 ; cfd-match-pos, etc.
5291 \f
5292 ;; A cache for found types.
5294 ;; Buffer local variable that contains an obarray with the types we've
5295 ;; found. If a declaration is recognized somewhere we record the
5296 ;; fully qualified identifier in it to recognize it as a type
5297 ;; elsewhere in the file too. This is not accurate since we do not
5298 ;; bother with the scoping rules of the languages, but in practice the
5299 ;; same name is seldom used as both a type and something else in a
5300 ;; file, and we only use this as a last resort in ambiguous cases (see
5301 ;; `c-forward-decl-or-cast-1').
5302 ;;
5303 ;; Not every type need be in this cache. However, things which have
5304 ;; ceased to be types must be removed from it.
5305 ;;
5306 ;; Template types in C++ are added here too but with the template
5307 ;; arglist replaced with "<>" in references or "<" for the one in the
5308 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5309 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5310 ;; template specs can be fairly sized programs in themselves) and
5311 ;; improves the hit ratio (it's a type regardless of the template
5312 ;; args; it's just not the same type, but we're only interested in
5313 ;; recognizing types, not telling distinct types apart). Note that
5314 ;; template types in references are added here too; from the example
5315 ;; above there will also be an entry "Foo<".
5320 ;; Clears `c-found-types'.
5324 ;; Add the given region as a type in `c-found-types'. If the region
5325 ;; doesn't match an existing type but there is a type which is equal
5326 ;; to the given one except that the last character is missing, then
5327 ;; the shorter type is removed. That's done to avoid adding all
5328 ;; prefixes of a type as it's being entered and font locked. This
5329 ;; doesn't cover cases like when characters are removed from a type
5330 ;; or added in the middle. We'd need the position of point when the
5331 ;; font locking is invoked to solve this well.
5332 ;;
5333 ;; This function might do hidden buffer changes.
5340 ;; Remove the "NAME" from c-found-types, if present.
5344 ;; Return non-nil if the given region contains a type in
5345 ;; `c-found-types'.
5346 ;;
5347 ;; This function might do hidden buffer changes.
5349 c-found-types))
5352 ;; Return all the types in `c-found-types' as a sorted list of
5353 ;; strings.
5357 type-list)))
5358 c-found-types)
5361 ;; Shut up the byte compiler.
5365 ;; An after change function which, in conjunction with the info in
5366 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5367 ;; from `c-found-types', should this type have become stale. For
5368 ;; example, this happens to "foo" when "foo \n bar();" becomes
5369 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5370 ;; the fontification.
5371 ;;
5372 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5373 ;; type?
5393 ;; Changing the amount of (already existing) whitespace - don't do anything.
5398 ;; The syntactic relationship which defined a "found type" has been
5399 ;; destroyed.
5402 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5403 )))
5405 \f
5406 ;; Setting and removing syntax properties on < and > in languages (C++
5407 ;; and Java) where they can be template/generic delimiters as well as
5408 ;; their normal meaning of "less/greater than".
5410 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5411 ;; be delimiters, they are marked as such with the category properties
5412 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5414 ;; STRATEGY:
5415 ;;
5416 ;; It is impossible to determine with certainty whether a <..> pair in
5417 ;; C++ is two comparison operators or is template delimiters, unless
5418 ;; one duplicates a lot of a C++ compiler. For example, the following
5419 ;; code fragment:
5420 ;;
5421 ;; foo (a < b, c > d) ;
5422 ;;
5423 ;; could be a function call with two integer parameters (each a
5424 ;; relational expression), or it could be a constructor for class foo
5425 ;; taking one parameter d of templated type "a < b, c >". They are
5426 ;; somewhat easier to distinguish in Java.
5427 ;;
5428 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5429 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5430 ;; individually when their context so indicated. This gave rise to
5431 ;; intractable problems when one of a matching pair was deleted, or
5432 ;; pulled into a literal.]
5433 ;;
5434 ;; At each buffer change, the syntax-table properties are removed in a
5435 ;; before-change function and reapplied, when needed, in an
5436 ;; after-change function. It is far more important that the
5437 ;; properties get removed when they they are spurious than that they
5438 ;; be present when wanted.
5439 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5441 ;; POS (default point) is at a < character. If it is marked with
5442 ;; open paren syntax-table text property, remove the property,
5443 ;; together with the close paren property on the matching > (if
5444 ;; any).
5459 ;; POS (default point) is at a > character. If it is marked with
5460 ;; close paren syntax-table property, remove the property, together
5461 ;; with the open paren property on the matching < (if any).
5476 ;; POS (default point) is at a < or > character. If it has an
5477 ;; open/close paren syntax-table property, remove this property both
5478 ;; from the current character and its partner (which will also be
5479 ;; thusly marked).
5489 ;; POS (default point) is at a < character. If it is both marked
5490 ;; with open/close paren syntax-table property, and has a matching >
5491 ;; (also marked) which is after LIM, remove the property both from
5492 ;; the current > and its partner. Return t when this happens, nil
5493 ;; when it doesn't.
5507 t)))
5510 ;; POS (default point) is at a > character. If it is both marked
5511 ;; with open/close paren syntax-table property, and has a matching <
5512 ;; (also marked) which is before LIM, remove the property both from
5513 ;; the current < and its partner. Return t when this happens, nil
5514 ;; when it doesn't.
5528 t)))
5530 ;; Set by c-common-init in cc-mode.el.
5535 ;; Unmark certain pairs of "< .... >" which are currently marked as
5536 ;; template/generic delimiters. (This marking is via syntax-table
5537 ;; text properties).
5538 ;;
5539 ;; These pairs are those which are in the current "statement" (i.e.,
5540 ;; the region between the {, }, or ; before BEG and the one after
5541 ;; END), and which enclose any part of the interval (BEG END).
5542 ;;
5543 ;; Note that in C++ (?and Java), template/generic parens cannot
5544 ;; enclose a brace or semicolon, so we use these as bounds on the
5545 ;; region we must work on.
5546 ;;
5547 ;; This function is called from before-change-functions (via
5548 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5549 ;; and point is undefined, both at entry and exit.
5550 ;;
5551 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5552 ;; 2010-01-29.
5556 new-beg new-end need-new-beg need-new-end)
5557 ;; Locate the barrier before the changed region
5562 ;; Remove the syntax-table/category properties from each pertinent <...>
5563 ;; pair. Firsly, the ones with the < before beg and > after beg.
5569 ;; Locate the barrier after END.
5574 ;; Remove syntax-table properties from the remaining pertinent <...>
5575 ;; pairs, those with a > after end and < before end.
5580 ;; Extend the fontification region, if needed.
5590 ;; This is called from `after-change-functions' when
5591 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5592 ;; chars with paren syntax become part of another operator like "<<"
5593 ;; or ">=".
5594 ;;
5595 ;; This function might do hidden buffer changes.
5628 \f
5629 ;; Handling of small scale constructs like types and names.
5631 ;; Dynamically bound variable that instructs `c-forward-type' to also
5632 ;; treat possible types (i.e. those that it normally returns 'maybe or
5633 ;; 'found for) as actual types (and always return 'found for them).
5634 ;; This means that it records them in `c-record-type-identifiers' if
5635 ;; that is set, and that it adds them to `c-found-types'.
5638 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5639 ;; mark up successfully parsed arglists with paren syntax properties on
5640 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5641 ;; `c-type' property of each argument separating comma.
5642 ;;
5643 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5644 ;; all arglists for side effects (i.e. recording types), otherwise it
5645 ;; exploits any existing paren syntax properties to quickly jump to the
5646 ;; end of already parsed arglists.
5647 ;;
5648 ;; Marking up the arglists is not the default since doing that correctly
5649 ;; depends on a proper value for `c-restricted-<>-arglists'.
5652 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5653 ;; not accept arglists that contain binary operators.
5654 ;;
5655 ;; This is primarily used to handle C++ template arglists. C++
5656 ;; disambiguates them by checking whether the preceding name is a
5657 ;; template or not. We can't do that, so we assume it is a template
5658 ;; if it can be parsed as one. That usually works well since
5659 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5660 ;; in almost all cases would be pointless.
5661 ;;
5662 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5663 ;; should let the comma separate the function arguments instead. And
5664 ;; in a context where the value of the expression is taken, e.g. in
5665 ;; "if (a < b || c > d)", it's probably not a template.
5668 ;; Dynamically bound variables that instructs
5669 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5670 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5671 ;; `c-forward-label' to record the ranges of all the type and
5672 ;; reference identifiers they encounter. They will build lists on
5673 ;; these variables where each element is a cons of the buffer
5674 ;; positions surrounding each identifier. This recording is only
5675 ;; activated when `c-record-type-identifiers' is non-nil.
5676 ;;
5677 ;; All known types that can't be identifiers are recorded, and also
5678 ;; other possible types if `c-promote-possible-types' is set.
5679 ;; Recording is however disabled inside angle bracket arglists that
5680 ;; are encountered inside names and other angle bracket arglists.
5681 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5682 ;; instead.
5683 ;;
5684 ;; Only the names in C++ template style references (e.g. "tmpl" in
5685 ;; "tmpl<a,b>::foo") are recorded as references, other references
5686 ;; aren't handled here.
5687 ;;
5688 ;; `c-forward-label' records the label identifier(s) on
5689 ;; `c-record-ref-identifiers'.
5693 ;; This variable will receive a cons cell of the range of the last
5694 ;; single identifier symbol stepped over by `c-forward-name' if it's
5695 ;; successful. This is the range that should be put on one of the
5696 ;; record lists above by the caller. It's assigned nil if there's no
5697 ;; such symbol in the name.
5702 ;; Always true.
5712 ;; Always true.
5720 ;; Dynamically bound variable that instructs `c-forward-type' to
5721 ;; record the ranges of types that only are found. Behaves otherwise
5722 ;; like `c-record-type-identifiers'.
5726 ;; Used internally in `c-forward-keyword-clause' to move forward
5727 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5728 ;; possibly is prefixed by keywords and their associated clauses.
5729 ;; Try with a type/name first to not trip up on those that begin
5730 ;; with a keyword. Return t if a known or found type is moved
5731 ;; over. The point is clobbered if nil is returned. If range
5732 ;; recording is enabled, the identifier is recorded on as a type
5733 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5734 ;;
5735 ;; This macro might do hidden buffer changes.
5740 nil
5747 t)))
5750 ;; Used internally in `c-forward-keyword-clause' to move forward
5751 ;; over a comma separated list of types or names using
5752 ;; `c-forward-keyword-prefixed-id'.
5753 ;;
5754 ;; This macro might do hidden buffer changes.
5765 ;; Submatch MATCH in the current match data is assumed to surround a
5766 ;; token. If it's a keyword, move over it and any immediately
5767 ;; following clauses associated with it, stopping at the start of
5768 ;; the next token. t is returned in that case, otherwise the point
5769 ;; stays and nil is returned. The kind of clauses that are
5770 ;; recognized are those specified by `c-type-list-kwds',
5771 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5772 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5773 ;; and `c-<>-arglist-kwds'.
5774 ;;
5775 ;; This function records identifier ranges on
5776 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5777 ;; `c-record-type-identifiers' is non-nil.
5778 ;;
5779 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5780 ;; apply directly after the keyword, the type list is moved over
5781 ;; only when there is no unaccounted token before it (i.e. a token
5782 ;; that isn't moved over due to some other keyword list). The
5783 ;; identifier ranges in the list are still recorded if that should
5784 ;; be done, though.
5785 ;;
5786 ;; This function might do hidden buffer changes.
5789 ;; The call to `c-forward-<>-arglist' below is made after
5790 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5791 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5792 ;; should therefore be nil.
5804 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5809 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5814 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5821 ;; Use `c-forward-type' on every identifier we can find
5822 ;; inside the paren, to record the types.
5847 ;; If we are at the colon already, we move over the type
5848 ;; list after it.
5854 ;; Not at the colon, so stop here. But the identifier
5855 ;; ranges in the type list later on should still be
5856 ;; recorded.
5859 ;; If a keyword matched both one of the types above and
5860 ;; this one, we match `c-colon-type-list-re' after the
5861 ;; clause matched above.
5868 ;; There's a type after the `c-colon-type-list-re' match
5869 ;; after a keyword in `c-colon-type-list-kwds'.
5873 t)))
5875 ;; cc-mode requires cc-fonts.
5879 ;; The point is assumed to be at a "<". Try to treat it as the open
5880 ;; paren of an angle bracket arglist and move forward to the
5881 ;; corresponding ">". If successful, the point is left after the
5882 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5883 ;; returned. If ALL-TYPES is t then all encountered arguments in
5884 ;; the arglist that might be types are treated as found types.
5885 ;;
5886 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5887 ;; function handles text properties on the angle brackets and argument
5888 ;; separating commas.
5889 ;;
5890 ;; `c-restricted-<>-arglists' controls how lenient the template
5891 ;; arglist recognition should be.
5892 ;;
5893 ;; This function records identifier ranges on
5894 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5895 ;; `c-record-type-identifiers' is non-nil.
5896 ;;
5897 ;; This function might do hidden buffer changes.
5900 ;; If `c-record-type-identifiers' is set then activate
5901 ;; recording of any found types that constitute an argument in
5902 ;; the arglist.
5910 ;; `nconc' doesn't mind that the tail of
5911 ;; `c-record-found-types' is t.
5914 t)
5917 nil)))
5920 ;; Recursive part of `c-forward-<>-arglist'.
5921 ;;
5922 ;; This function might do hidden buffer changes.
5924 ;; Cover this so that any recorded found type ranges are
5925 ;; automatically lost if it turns out to not be an angle
5926 ;; bracket arglist. It's propagated through the return value
5927 ;; on successful completion.
5929 ;; List that collects the positions after the argument
5930 ;; separating ',' in the arglist.
5931 arg-start-pos)
5932 ;; If the '<' has paren open syntax then we've marked it as an angle
5933 ;; bracket arglist before, so skip to the end.
5941 t
5942 ;; Got unmatched paren angle brackets. We don't clear the paren
5943 ;; syntax properties and retry, on the basis that it's very
5944 ;; unlikely that paren angle brackets become operators by code
5945 ;; manipulation. It's far more likely that it doesn't match due
5946 ;; to narrowing or some temporary change.
5948 nil))
5953 ;; go forward one non-alphanumeric character (group) per iteration of
5954 ;; this loop.
5960 ;; All encountered identifiers are types, so set the
5961 ;; promote flag and parse the type.
5984 ;; Note: These regexps exploit the match order in \| so
5985 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5987 ;; Stop on ',', '|', '&', '+' and '-' to catch
5988 ;; common binary operators that could be between
5989 ;; two comparison expressions "a<b" and "c>d".
5991 nil t t))
5995 ;; Either an operator starting with '>' or the end of
5996 ;; the angle bracket arglist.
6003 ;; The angle bracket arglist is finished.
6015 ;; Either an operator starting with '<' or a nested arglist.
6019 ;; The '<' begins a multi-char operator.
6023 ;; We're at a nested <.....>
6029 ;; There's always an identifier before an angle
6030 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
6031 ;; or `c-<>-arglist-kwds'.
6047 ;; It was an angle bracket arglist.
6050 ;; Record the identifier before the template as a type
6051 ;; or reference depending on whether the arglist is last
6052 ;; in a qualified identifier.
6062 ;; At a "less than" operator.
6065 )))
6072 ;; Just another argument. Record the position. The
6073 ;; type check stuff that made us stop at it is at
6074 ;; the top of the loop.
6078 ;; Got a character that can't be in an angle bracket
6079 ;; arglist argument. Abort using `throw', since
6080 ;; it's useless to try to find a surrounding arglist
6081 ;; if we're nested.
6087 ;; The point is assumed to be directly after a ">". Try to treat it
6088 ;; as the close paren of an angle bracket arglist and move back to
6089 ;; the corresponding "<". If successful, the point is left at
6090 ;; the "<" and t is returned, otherwise the point isn't moved and
6091 ;; nil is returned. ALL-TYPES is passed on to
6092 ;; `c-forward-<>-arglist'.
6093 ;;
6094 ;; If the optional LIMIT is given, it bounds the backward search.
6095 ;; It's then assumed to be at a syntactically relevant position.
6096 ;;
6097 ;; This is a wrapper around `c-forward-<>-arglist'. See that
6098 ;; function for more details.
6107 t
6108 ;; See corresponding note in `c-forward-<>-arglist'.
6110 nil)
6117 t
6118 ;; Stopped at bob or a char that isn't allowed in an
6119 ;; arglist, so we've failed.
6121 nil)
6126 ;; If we moved then the "<" was part of some
6127 ;; multicharacter token.
6128 t
6134 ;; Matched the arglist. Break the while.
6136 nil)
6138 ;; We started from a non-paren ">" inside an
6139 ;; arglist.
6141 nil)
6143 ;; Matched a shorter arglist. Can be a nested
6144 ;; one so continue looking.
6146 t))
6147 t))))))
6152 ;; Move forward over a complete name if at the beginning of one,
6153 ;; stopping at the next following token. A keyword, as such,
6154 ;; doesn't count as a name. If the point is not at something that
6155 ;; is recognized as a name then it stays put.
6156 ;;
6157 ;; A name could be something as simple as "foo" in C or something as
6158 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6159 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6160 ;; int>::*volatile const" in C++ (this function is actually little
6161 ;; more than a `looking-at' call in all modes except those that,
6162 ;; like C++, have `c-recognize-<>-arglists' set).
6163 ;;
6164 ;; Return
6165 ;; o - nil if no name is found;
6166 ;; o - 'template if it's an identifier ending with an angle bracket
6167 ;; arglist;
6168 ;; o - 'operator of it's an operator identifier;
6169 ;; o - t if it's some other kind of name.
6170 ;;
6171 ;; This function records identifier ranges on
6172 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6173 ;; `c-record-type-identifiers' is non-nil.
6174 ;;
6175 ;; This function might do hidden buffer changes.
6178 ;; Turn off `c-promote-possible-types' here since we might
6179 ;; call `c-forward-<>-arglist' and we don't want it to promote
6180 ;; every suspect thing in the arglist to a type. We're
6181 ;; typically called from `c-forward-type' in this case, and
6182 ;; the caller only wants the top level type that it finds to
6183 ;; be promoted.
6184 c-promote-possible-types)
6190 ;; Check for keyword. We go to the last symbol in
6191 ;; `c-identifier-key' first.
6204 ;; "template" is only valid inside an
6205 ;; identifier if preceded by "::".
6210 t))
6212 ;; Handle a C++ operator or template identifier.
6216 ;; Got "... ::template".
6220 res subres))))
6223 ;; Got a cast operator.
6227 ;; Now we should match a sequence of either
6228 ;; '*', '&' or a name followed by ":: *",
6229 ;; where each can be followed by a sequence
6230 ;; of `c-opt-type-modifier-key'.
6233 t)
6243 t))))
6251 ;; Got some other operator.
6259 nil)
6261 ;; `id-start' is equal to `id-end' if we've jumped over
6262 ;; an identifier that doesn't end with a symbol token.
6263 ;; That can occur e.g. for Java import directives on the
6264 ;; form "foo.bar.*".
6271 res t)))
6281 ;; Got a concatenated identifier. This handles the
6282 ;; cases with tricky syntactic whitespace that aren't
6283 ;; covered in `c-identifier-key'.
6286 t)
6290 ;; Maybe an angle bracket arglist.
6298 c-last-identifier-range nil)
6303 ;; Continue if there's an identifier concatenation
6304 ;; operator after the template argument.
6310 t)
6315 nil)))
6316 )))))
6319 res))
6322 ;; Move forward over a type spec if at the beginning of one,
6323 ;; stopping at the next following token. The keyword "typedef"
6324 ;; isn't part of a type spec here.
6325 ;;
6326 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6327 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6328 ;; The current (2009-03-10) intention is to convert all uses of
6329 ;; `c-forward-type' to call with this parameter set, then to
6330 ;; eliminate it.
6331 ;;
6332 ;; Return
6333 ;; o - t if it's a known type that can't be a name or other
6334 ;; expression;
6335 ;; o - 'known if it's an otherwise known type (according to
6336 ;; `*-font-lock-extra-types');
6337 ;; o - 'prefix if it's a known prefix of a type;
6338 ;; o - 'found if it's a type that matches one in `c-found-types';
6339 ;; o - 'maybe if it's an identifier that might be a type;
6340 ;; o - 'decltype if it's a decltype(variable) declaration; - or
6341 ;; o - nil if it can't be a type (the point isn't moved then).
6342 ;;
6343 ;; The point is assumed to be at the beginning of a token.
6344 ;;
6345 ;; Note that this function doesn't skip past the brace definition
6346 ;; that might be considered part of the type, e.g.
6347 ;; "enum {a, b, c} foo".
6348 ;;
6349 ;; This function records identifier ranges on
6350 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6351 ;; `c-record-type-identifiers' is non-nil.
6352 ;;
6353 ;; This function might do hidden buffer changes.
6361 ;; Skip leading type modifiers. If any are found we know it's a
6362 ;; prefix of a type.
6381 ; "typedef".
6389 ;; In many languages the name can be used without the
6390 ;; prefix, so we add it to `c-found-types'.
6393 c-last-identifier-range)
6415 id-range c-last-identifier-range))))
6427 id-end)
6429 ;; Looking at a primitive or known type identifier. We've
6430 ;; checked for a name first so that we don't go here if the
6431 ;; known type match only is a prefix of another name.
6442 ;; There might be more keywords for the type.
6470 ;; A normal identifier.
6480 ;; It's an identifier that has been used as
6481 ;; a type somewhere else.
6482 'found
6483 ;; It's an identifier that might be a type.
6486 ;; A template is a type.
6490 ;; Otherwise it's an operator identifier, which is not a type.
6495 ;; Skip trailing type modifiers. If any are found we know it's
6496 ;; a type.
6503 ;; Step over any type suffix operator. Do not let the existence
6504 ;; of these alter the classification of the found type, since
6505 ;; these operators typically are allowed in normal expressions
6506 ;; too.
6513 ;; Look for a trailing operator that concatenates the type
6514 ;; with a following one, and if so step past that one through
6515 ;; a recursive call. Note that we don't record concatenated
6516 ;; types in `c-found-types' - it's the component types that
6517 ;; are recorded when appropriate.
6520 c-promote-possible-types))
6521 ;; If we can't promote then set `c-record-found-types' so that
6522 ;; we can merge in the types from the second part afterwards if
6523 ;; it turns out to be a known type there.
6526 subres)
6535 ;; If either operand certainly is a type then both are, but we
6536 ;; don't let the existence of the operator itself promote two
6537 ;; uncertain types to a certain one.
6556 ;; Merge in the ranges of any types found by the second
6557 ;; `c-forward-type'.
6559 ;; `nconc' doesn't mind that the tail of
6560 ;; `c-record-found-types' is t.
6562 c-record-type-identifiers))))
6570 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6572 res))
6575 ;; Used for Java code only at the moment. Assumes point is on the
6576 ;; @, moves forward an annotation. returns nil if there is no
6577 ;; annotation at point.
6584 t)))
6587 ;; Pull the next open brace from PS (which has the form of paren-state),
6588 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
6596 ;; Point is just after a closing brace/parenthesis. Try to parse this as a
6597 ;; C++ member initializer list, going back to just after the introducing ":"
6598 ;; and returning t. Otherwise return nil, leaving point unchanged.
6623 res))
6626 ;; Test whether we are in a C++ member initializer list, and if so, go back
6627 ;; to the introducing ":", returning the position of the opening paren of
6628 ;; the function's arglist. Otherwise return nil, leaving point unchanged.
6631 res)
6669 res))
6671 \f
6672 ;; Handling of large scale constructs like statements and declarations.
6674 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6675 ;; defsubst or perhaps even a defun, but it contains lots of free
6676 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6678 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6679 ;; of types when parsing a declaration, which means that it
6680 ;; sometimes consumes the identifier in the declaration as a type.
6681 ;; This is used to "backtrack" and make the last type be treated as
6682 ;; an identifier instead.
6685 ;; These identifiers are bound only in the inner let.
6687 identifier-start type-start
6688 got-parens nil
6689 got-identifier t
6690 got-suffix t
6691 got-suffix-after-parens id-start
6695 t
6696 backup-at-type))
6698 id-start backup-id-start)
6700 id-start start-pos))
6702 ;; When these flags already are set we've found specifiers that
6703 ;; unconditionally signal these attributes - backtracking doesn't
6704 ;; change that. So keep them set in that case.
6711 ;; This identifier is bound only in the inner let.
6715 ;; Move forward over a declaration or a cast if at the start of one.
6716 ;; The point is assumed to be at the start of some token. Nil is
6717 ;; returned if no declaration or cast is recognized, and the point
6718 ;; is clobbered in that case.
6719 ;;
6720 ;; If a declaration is parsed:
6721 ;;
6722 ;; The point is left at the first token after the first complete
6723 ;; declarator, if there is one. The return value is a cons where
6724 ;; the car is the position of the first token in the declarator. (See
6725 ;; below for the cdr.)
6726 ;; Some examples:
6727 ;;
6728 ;; void foo (int a, char *b) stuff ...
6729 ;; car ^ ^ point
6730 ;; float (*a)[], b;
6731 ;; car ^ ^ point
6732 ;; unsigned int a = c_style_initializer, b;
6733 ;; car ^ ^ point
6734 ;; unsigned int a (cplusplus_style_initializer), b;
6735 ;; car ^ ^ point (might change)
6736 ;; class Foo : public Bar {}
6737 ;; car ^ ^ point
6738 ;; class PikeClass (int a, string b) stuff ...
6739 ;; car ^ ^ point
6740 ;; enum bool;
6741 ;; car ^ ^ point
6742 ;; enum bool flag;
6743 ;; car ^ ^ point
6744 ;; void cplusplus_function (int x) throw (Bad);
6745 ;; car ^ ^ point
6746 ;; Foo::Foo (int b) : Base (b) {}
6747 ;; car ^ ^ point
6748 ;;
6749 ;; auto foo = 5;
6750 ;; car ^ ^ point
6751 ;; auto cplusplus_11 (int a, char *b) -> decltype (bar):
6752 ;; car ^ ^ point
6753 ;;
6754 ;;
6755 ;;
6756 ;; The cdr of the return value is non-nil when a
6757 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6758 ;; Specifically it is a dotted pair (A . B) where B is t when a
6759 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6760 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6761 ;; specifier is present. I.e., (some of) the declared
6762 ;; identifier(s) are types.
6763 ;;
6764 ;; If a cast is parsed:
6765 ;;
6766 ;; The point is left at the first token after the closing paren of
6767 ;; the cast. The return value is `cast'. Note that the start
6768 ;; position must be at the first token inside the cast parenthesis
6769 ;; to recognize it.
6770 ;;
6771 ;; PRECEDING-TOKEN-END is the first position after the preceding
6772 ;; token, i.e. on the other side of the syntactic ws from the point.
6773 ;; Use a value less than or equal to (point-min) if the point is at
6774 ;; the first token in (the visible part of) the buffer.
6775 ;;
6776 ;; CONTEXT is a symbol that describes the context at the point:
6777 ;; 'decl In a comma-separated declaration context (typically
6778 ;; inside a function declaration arglist).
6779 ;; '<> In an angle bracket arglist.
6780 ;; 'arglist Some other type of arglist.
6781 ;; nil Some other context or unknown context. Includes
6782 ;; within the parens of an if, for, ... construct.
6783 ;;
6784 ;; LAST-CAST-END is the first token after the closing paren of a
6785 ;; preceding cast, or nil if none is known. If
6786 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6787 ;; the position after the closest preceding call where a cast was
6788 ;; matched. In that case it's used to discover chains of casts like
6789 ;; "(a) (b) c".
6790 ;;
6791 ;; This function records identifier ranges on
6792 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6793 ;; `c-record-type-identifiers' is non-nil.
6794 ;;
6795 ;; This function might do hidden buffer changes.
6798 ;; first type, i.e. after any leading specifiers. It might
6799 ;; also point at the beginning of the preceding syntactic
6800 ;; whitespace.
6802 ;; Set to the result of `c-forward-type'.
6803 at-type
6804 ;; The position of the first token in what we currently
6805 ;; believe is the type in the declaration or cast, after any
6806 ;; specifiers and their associated clauses.
6807 type-start
6808 ;; The position of the first token in what we currently
6809 ;; believe is the declarator for the first identifier. Set
6810 ;; when the type is found, and moved forward over any
6811 ;; `c-decl-hangon-kwds' and their associated clauses that
6812 ;; occurs after the type.
6813 id-start
6814 ;; These store `at-type', `type-start' and `id-start' of the
6815 ;; identifier before the one in those variables. The previous
6816 ;; identifier might turn out to be the real type in a
6817 ;; declaration if the last one has to be the declarator in it.
6818 ;; If `backup-at-type' is nil then the other variables have
6819 ;; undefined values.
6820 backup-at-type backup-type-start backup-id-start
6821 ;; This stores `kwd-sym' of the symbol before the current one.
6822 ;; This is needed to distinguish the C++11 version of "auto" from
6823 ;; the pre C++11 meaning.
6824 backup-kwd-sym
6825 ;; Set if we've found a specifier (apart from "typedef") that makes
6826 ;; the defined identifier(s) types.
6827 at-type-decl
6828 ;; Set if we've a "typedef" keyword.
6829 at-typedef
6830 ;; Set if we've found a specifier that can start a declaration
6831 ;; where there's no type.
6832 maybe-typeless
6833 ;; Save the value of kwd-sym between loops of the "Check for a
6834 ;; type" loop. Needed to distinguish a C++11 "auto" from a pre
6835 ;; C++11 one.
6836 prev-kwd-sym
6837 ;; If a specifier is found that also can be a type prefix,
6838 ;; these flags are set instead of those above. If we need to
6839 ;; back up an identifier, they are copied to the real flag
6840 ;; variables. Thus they only take effect if we fail to
6841 ;; interpret it as a type.
6842 backup-at-type-decl backup-maybe-typeless
6843 ;; Whether we've found a declaration or a cast. We might know
6844 ;; this before we've found the type in it. It's 'ids if we've
6845 ;; found two consecutive identifiers (usually a sure sign, but
6846 ;; we should allow that in labels too), and t if we've found a
6847 ;; specifier keyword (a 100% sure sign).
6848 at-decl-or-cast
6849 ;; Set when we need to back up to parse this as a declaration
6850 ;; but not as a cast.
6851 backup-if-not-cast
6852 ;; For casts, the return position.
6853 cast-end
6854 ;; Have we got a new-style C++11 "auto"?
6855 new-style-auto
6856 ;; Save `c-record-type-identifiers' and
6857 ;; `c-record-ref-identifiers' since ranges are recorded
6858 ;; speculatively and should be thrown away if it turns out
6859 ;; that it isn't a declaration or cast.
6866 ;; Check for a type. Unknown symbols are treated as possible
6867 ;; types, but they could also be specifiers disguised through
6868 ;; macros like __INLINE__, so we recognize both types and known
6869 ;; specifiers after them too.
6873 ;; Look for a specifier keyword clause.
6886 ;; Found a known or possible type or a prefix of a known type.
6895 ;; Got two identifiers with nothing but whitespace
6896 ;; between them. That can only happen in declarations.
6900 ;; If the previous identifier is a found type we
6901 ;; record it as a real one; it might be some sort of
6902 ;; alias for a prefix like "unsigned".
6909 backup-type-start type-start
6910 backup-id-start id-start
6911 backup-kwd-sym kwd-sym
6912 at-type found-type
6913 type-start start
6915 ;; The previous ambiguous specifier/type turned out
6916 ;; to be a type since we've parsed another one after
6917 ;; it, so clear these backup flags.
6918 backup-at-type-decl nil
6919 backup-maybe-typeless nil))
6923 ;; Handle known specifier keywords and
6924 ;; `c-decl-hangon-kwds' which can occur after known
6925 ;; types.
6928 ;; It's a hang-on keyword that can occur anywhere.
6932 ;; Move the identifier start position if
6933 ;; we've passed a type.
6935 ;; Otherwise treat this as a specifier and
6936 ;; move the fallback position.
6940 ;; It's an ordinary specifier so we know that
6941 ;; anything before this can't be the type.
6943 start-pos kwd-clause-end)
6946 ;; It's ambiguous whether this keyword is a
6947 ;; specifier or a type prefix, so set the backup
6948 ;; flags. (It's assumed that `c-forward-type'
6949 ;; moved further than `c-forward-keyword-clause'.)
6957 ;; This test only happens after we've scanned a type.
6958 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6963 ;; Haven't matched a type so it's an unambiguous
6964 ;; specifier keyword and we know we're in a
6965 ;; declaration.
6971 ;; If the type isn't known we continue so that we'll jump
6972 ;; over all specifiers and type identifiers. The reason
6973 ;; to do this for a known type prefix is to make things
6974 ;; like "unsigned INT16" work.
6979 ;; If a known type was found, we still need to skip over any
6980 ;; hangon keyword clauses after it. Otherwise it has already
6981 ;; been done in the loop above.
6987 ;; A prefix type is itself a primitive type when it's not
6988 ;; followed by another type.
6992 ;; Got no type but set things up to continue anyway to handle
6993 ;; the various cases when a declaration doesn't start with a
6994 ;; type.
6999 ;; If it's C++ then check if the last "type" ends on the form
7000 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
7001 ;; (con|de)structor.
7012 ;; Cheating in the handling of syntactic ws below.
7019 name))
7020 ;; It is a (con|de)structor name. In that case the
7021 ;; declaration is typeless so zap out any preceding
7022 ;; identifier(s) that we might have taken as types.
7025 backup-at-type nil
7026 id-start type-start))))))
7028 ;; Check for and step over a type decl expression after the thing
7029 ;; that is or might be a type. This can't be skipped since we
7030 ;; need the correct end position of the declarator for
7031 ;; `max-type-decl-end-*'.
7033 ;; True if there's a non-open-paren match of
7034 ;; `c-type-decl-prefix-key'.
7035 got-prefix
7036 ;; True if the declarator is surrounded by a parenthesis pair.
7037 got-parens
7038 ;; True if there is an identifier in the declarator.
7039 got-identifier
7040 ;; True if there's a non-close-paren match of
7041 ;; `c-type-decl-suffix-key'.
7042 got-suffix
7043 ;; True if there's a prefix match outside the outermost
7044 ;; paren pair that surrounds the declarator.
7045 got-prefix-before-parens
7046 ;; True if there's a suffix match outside the outermost
7047 ;; paren pair that surrounds the declarator. The value is
7048 ;; the position of the first suffix match.
7049 got-suffix-after-parens
7050 ;; True if we've parsed the type decl to a token that is
7051 ;; known to end declarations in this context.
7052 at-decl-end
7053 ;; The earlier values of `at-type' and `type-start' if we've
7054 ;; shifted the type backwards.
7055 identifier-type identifier-start
7056 ;; If `c-parse-and-markup-<>-arglists' is set we need to
7057 ;; turn it off during the name skipping below to avoid
7058 ;; getting `c-type' properties that might be bogus. That
7059 ;; can happen since we don't know if
7060 ;; `c-restricted-<>-arglists' will be correct inside the
7061 ;; arglist paren that gets entered.
7062 c-parse-and-markup-<>-arglists
7063 ;; Start of the identifier for which `got-identifier' was set.
7064 name-start)
7068 ;; Skip over type decl prefix operators. (Note similar code in
7069 ;; `c-font-lock-declarators'.)
7079 ;; If the third submatch matches in C++ then
7080 ;; we're looking at an identifier that's a
7081 ;; prefix only if it specifies a member pointer.
7086 ;; We only check for a trailing "::" and
7087 ;; let the "*" that should follow be
7088 ;; matched in the next round.
7090 ;; It turned out to be the real identifier,
7091 ;; so stop.
7092 nil))
7093 t))
7107 ;; Skip over an identifier.
7114 ;; Skip over type decl suffix operators.
7121 t)
7125 t)
7131 ;; No suffix matched. We might have matched the
7132 ;; identifier as a type and the open paren of a
7133 ;; function arglist as a type decl prefix. In that
7134 ;; case we should "backtrack": Reinterpret the last
7135 ;; type as the identifier, move out of the arglist and
7136 ;; continue searching for suffix operators.
7137 ;;
7138 ;; Do this even if there's no preceding type, to cope
7139 ;; with old style function declarations in K&R C,
7140 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
7141 ;; style declarations. That isn't applicable in an
7142 ;; arglist context, though.
7147 maybe-typeless
7148 backup-maybe-typeless
7155 t))
7164 at-type))
7165 ;; Have found no identifier but `c-typeless-decl-kwds' has
7166 ;; matched so we know we're inside a declaration. The
7167 ;; preceding type must be the identifier instead.
7170 ;; Prepare the "-> type;" for fontification later on.
7180 at-decl-or-cast
7183 ;; CASE 1
7185 ;; Encountered something inside parens that isn't matched by
7186 ;; the `c-type-decl-*' regexps, so it's not a type decl
7187 ;; expression. Try to skip out to the same paren depth to
7188 ;; not confuse the cast check below.
7190 ;; If we've found a specifier keyword then it's a
7191 ;; declaration regardless.
7199 ;; Now we've collected info about various characteristics of
7200 ;; the construct we're looking at. Below follows a decision
7201 ;; tree based on that. It's ordered to check more certain
7202 ;; signs before less certain ones.
7207 ;; CASE 2
7210 ;; Got another identifier directly after the type, so it's a
7211 ;; declaration.
7216 ;; (not got-suffix-after-parens)
7218 maybe-typeless
7219 backup-maybe-typeless
7224 ;; Got a declaration of the form "foo bar (gnu);" or "bar
7225 ;; (gnu);" where we've recognized "bar" as the type and "gnu"
7226 ;; as the declarator. In this case it's however more likely
7227 ;; that "bar" is the declarator and "gnu" a function argument
7228 ;; or initializer (if `c-recognize-paren-inits' is set),
7229 ;; since the parens around "gnu" would be superfluous if it's
7230 ;; a declarator. Shift the type one step backward.
7233 ;; Found no identifier.
7238 ;; CASE 3
7240 ;; Got a plain list of identifiers. If a colon follows it's
7241 ;; a valid label, or maybe a bitfield. Otherwise the last
7242 ;; one probably is the declared identifier and we should
7243 ;; back up to the previous type, providing it isn't a cast.
7247 ;; If we've found a specifier keyword then it's a
7248 ;; declaration regardless.
7259 ;; CASE 4
7263 ;; Got a plain list of identifiers followed by some suffix.
7264 ;; If this isn't a cast then the last identifier probably is
7265 ;; the declared one and we should back up to the previous
7266 ;; type.
7270 ;; CASE 5
7272 ;; If the type is known we know that there can't be any
7273 ;; identifier somewhere else, and it's only in declarations in
7274 ;; e.g. function prototypes and in casts that the identifier may
7275 ;; be left out.
7279 ;; Only got a single identifier (parsed as a type so far).
7280 ;; CASE 6
7282 ;; Check that the identifier isn't at the start of an
7283 ;; expression.
7284 at-decl-end
7287 ;; Inside an arglist that contains declarations. If K&R
7288 ;; style declarations and parenthesis style initializers
7289 ;; aren't allowed then the single identifier must be a
7290 ;; type, else we require that it's known or found
7291 ;; (primitive types are handled above).
7296 ;; Inside a template arglist. Accept known and found
7297 ;; types; other identifiers could just as well be
7298 ;; constants in C++.
7301 ;; CASE 7
7302 ;; Can't be a valid declaration or cast, but if we've found a
7303 ;; specifier it can't be anything else either, so treat it as
7304 ;; an invalid/unfinished declaration or cast.
7312 maybe-typeless
7313 backup-maybe-typeless
7317 c-after-suffixed-type-maybe-decl-key))))))
7318 ;; Got an empty paren pair and a preceding type that probably
7319 ;; really is the identifier. Shift the type backwards to make
7320 ;; the last one the identifier. This is analogous to the
7321 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7322 ;; above.
7323 ;;
7324 ;; Exception: In addition to the conditions in that
7325 ;; "backtracking" code, do not shift backward if we're not
7326 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7327 ;; Since there's no preceding type, the shift would mean that
7328 ;; the declaration is typeless. But if the regexp doesn't match
7329 ;; then we will simply fall through in the tests below and not
7330 ;; recognize it at all, so it's better to try it as an abstract
7331 ;; declarator instead.
7334 ;; Still no identifier.
7335 ;; CASE 8
7337 ;; Require `got-prefix' together with either `got-parens' or
7338 ;; `got-suffix' to recognize it as an abstract declarator:
7339 ;; `got-parens' only is probably an empty function call.
7340 ;; `got-suffix' only can build an ordinary expression together
7341 ;; with the preceding identifier which we've taken as a type.
7342 ;; We could actually accept on `got-prefix' only, but that can
7343 ;; easily occur temporarily while writing an expression so we
7344 ;; avoid that case anyway. We could do a better job if we knew
7345 ;; the point when the fontification was invoked.
7348 ;; CASE 9
7352 got-suffix-after-parens
7354 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7355 ;; normal function call after all (or perhaps a C++ style object
7356 ;; instantiation expression).
7359 ;; CASE 10
7361 ;; By now we've located the type in the declaration that we know
7362 ;; we're in.
7365 ;; CASE 11
7373 ;; Shift the type backward in the case that there's a
7374 ;; single identifier inside parens. That can only
7375 ;; occur in K&R style function declarations so it's
7376 ;; more likely that it really is a function call.
7377 ;; Therefore we only do this after
7378 ;; `c-after-suffixed-type-decl-key' has matched.
7380 got-suffix-after-parens))
7381 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7384 ;; CASE 12
7387 ;; It's a declaration if a known type precedes it and it can't be a
7388 ;; function call.
7391 ;; If we get here we can't tell if this is a type decl or a normal
7392 ;; expression by looking at it alone. (That's under the assumption
7393 ;; that normal expressions always can look like type decl expressions,
7394 ;; which isn't really true but the cases where it doesn't hold are so
7395 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7396 ;; the effort to look for them.)
7398 ;;; 2008-04-16: commented out the next form, to allow the function to recognize
7399 ;;; "foo (int bar)" in CC (an implicit type (in class foo) without a semicolon)
7400 ;;; as a(n almost complete) declaration, enabling it to be fontified.
7401 ;; CASE 13
7402 ;; (unless (or at-decl-end (looking-at "=[^=]"))
7403 ;; If this is a declaration it should end here or its initializer(*)
7404 ;; should start here, so check for allowed separation tokens. Note
7405 ;; that this rule doesn't work e.g. with a K&R arglist after a
7406 ;; function header.
7407 ;;
7408 ;; *) Don't check for C++ style initializers using parens
7409 ;; since those already have been matched as suffixes.
7410 ;;
7411 ;; If `at-decl-or-cast' is then we've found some other sign that
7412 ;; it's a declaration or cast, so then it's probably an
7413 ;; invalid/unfinished one.
7414 ;; (throw 'at-decl-or-cast at-decl-or-cast))
7416 ;; Below are tests that only should be applied when we're certain to
7417 ;; not have parsed halfway through an expression.
7419 ;; CASE 14
7421 ;; The expression starts with a known type so treat it as a
7422 ;; declaration.
7425 ;; CASE 15
7427 ;; In C++ we check if the identifier is a known type, since
7428 ;; (con|de)structors use the class name as identifier.
7429 ;; We've always shifted over the identifier as a type and
7430 ;; then backed up again in this case.
7431 identifier-type
7434 ;; `at-type' probably won't be 'found for
7435 ;; destructors since the "~" is then part of the
7436 ;; type name being checked against the list of
7437 ;; known types, so do a check without that
7438 ;; operator.
7443 c-identifier-syntax-table
7447 ;; We have already parsed the type earlier,
7448 ;; so it'd be possible to cache the end
7449 ;; position instead of redoing it here, but
7450 ;; then we'd need to keep track of another
7451 ;; position everywhere.
7459 ;; CASE 16
7461 at-type
7465 ;; Got something like "foo * bar;". Since we're not inside an
7466 ;; arglist it would be a meaningless expression because the
7467 ;; result isn't used. We therefore choose to recognize it as
7468 ;; a declaration. Do not allow a suffix since it could then
7469 ;; be a function call.
7472 ;; CASE 17
7477 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7478 ;; be an odd expression or it could be a declaration. Treat
7479 ;; it as a declaration if "a" has been used as a type
7480 ;; somewhere else (if it's a known type we won't get here).
7483 ;; CASE 18
7489 ;; Got a type followed by an abstract declarator. If `got-prefix'
7490 ;; is set it's something like "a *" without anything after it. If
7491 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7492 ;; or similar, which we accept only if the context rules out
7493 ;; expressions.
7496 ;; If we had a complete symbol table here (which rules out
7497 ;; `c-found-types') we should return t due to the disambiguation rule
7498 ;; (in at least C++) that anything that can be parsed as a declaration
7499 ;; is a declaration. Now we're being more defensive and prefer to
7500 ;; highlight things like "foo (bar);" as a declaration only if we're
7501 ;; inside an arglist that contains declarations.
7502 ;; CASE 19
7505 ;; The point is now after the type decl expression.
7508 ;; Check for a cast.
7511 c-cast-parens
7513 ;; Should be the first type/identifier in a cast paren.
7517 ;; The closing paren should follow.
7522 ;; There should be a primary expression after it.
7531 ;; Check if the expression begins with a prefix keyword.
7534 ;; Expression begins with an ambiguous operator. Treat
7535 ;; it as a cast if it's a type decl or if we've
7536 ;; recognized the type somewhere else.
7539 ;; Unless it's a keyword, it's the beginning of a primary
7540 ;; expression.
7542 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7543 ;; that it matched a whole one so that we don't e.g. confuse
7544 ;; the operator '-' with '->'. It's ok if it matches further,
7545 ;; though, since it e.g. can match the float '.5' while the
7546 ;; operator regexp only matches '.'.
7550 ;; There should either be a cast before it or something that isn't an
7551 ;; identifier or close paren.
7559 ;; It's a symbol. Accept it only if it's one of the
7560 ;; keywords that can precede an expression (without
7561 ;; surrounding parens).
7564 ;; Check that it isn't a close paren (block close is ok,
7565 ;; though).
7567 ;; Check that it isn't a nonsymbol identifier.
7570 ;; Handle the cast.
7580 ;; We're at a declaration. Highlight the type and the following
7581 ;; declarators.
7587 ;; Make sure to propagate the `c-decl-arg-start' property to
7588 ;; the next argument if it's set in this one, to cope with
7589 ;; interactive refontification.
7592 ;; Record the type's coordinates in `c-record-type-identifiers' for
7593 ;; later fontification.
7595 ;; There seems no reason to exclude a token from
7596 ;; fontification just because it's "a known type that can't
7597 ;; be a name or other expression". 2013-09-18.
7598 )
7609 ;; False alarm. Restore the recorded ranges.
7611 c-record-ref-identifiers save-rec-ref-ids)
7612 nil))))
7615 ;; Assuming that point is at the beginning of a token, check if it starts a
7616 ;; label and if so move over it and return non-nil (t in default situations,
7617 ;; specific symbols (see below) for interesting situations), otherwise don't
7618 ;; move and return nil. "Label" here means "most things with a colon".
7619 ;;
7620 ;; More precisely, a "label" is regarded as one of:
7621 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7622 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7623 ;; bare "case", should the colon be missing. We return t;
7624 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7625 ;; return t;
7626 ;; (iv) One of QT's "extended" C++ variants of
7627 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7628 ;; Returns the symbol `qt-2kwds-colon'.
7629 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7630 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7631 ;; colon). Currently (2006-03), this applies only to Objective C's
7632 ;; keywords "@private", "@protected", and "@public". Returns t.
7633 ;;
7634 ;; One of the things which will NOT be recognized as a label is a bit-field
7635 ;; element of a struct, something like "int foo:5".
7636 ;;
7637 ;; The end of the label is taken to be just after the colon, or the end of
7638 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7639 ;; after the end on return. The terminating char gets marked with
7640 ;; `c-decl-end' to improve recognition of the following declaration or
7641 ;; statement.
7642 ;;
7643 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7644 ;; label, if any, has already been marked up like that.
7645 ;;
7646 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7647 ;; after the preceding token, i.e. on the other side of the
7648 ;; syntactic ws from the point. Use a value less than or equal to
7649 ;; (point-min) if the point is at the first token in (the visible
7650 ;; part of) the buffer.
7651 ;;
7652 ;; The optional LIMIT limits the forward scan for the colon.
7653 ;;
7654 ;; This function records the ranges of the label symbols on
7655 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7656 ;; non-nil.
7657 ;;
7658 ;; This function might do hidden buffer changes.
7661 label-end
7662 qt-symbol-idx
7663 macro-start ; if we're in one.
7664 label-type
7665 kwd)
7667 ;; "case" or "default" (Doesn't apply to AWK).
7670 ;; Record only the keyword itself for fontification, since in
7671 ;; case labels the following is a constant expression and not
7672 ;; a label.
7676 ;; Find the label end.
7680 ;; Stop on chars that aren't allowed in expressions,
7681 ;; and on operator chars that would be meaningless
7682 ;; there. FIXME: This doesn't cope with ?: operators.
7690 t)
7692 ;; It's an unfinished label. We consider the keyword enough
7693 ;; to recognize it as a label, so that it gets fontified.
7694 ;; Leave the point at the end of it, but don't put any
7695 ;; `c-decl-end' marker.
7697 t))))
7699 ;; @private, @protected, @public, in Objective C, or similar.
7702 ;; For a `c-opt-extra-label-key' match, we record the whole
7703 ;; thing for fontification. That's to get the leading '@' in
7704 ;; Objective-C protection labels fontified.
7711 ;; All other cases of labels.
7714 ;; A colon label must have something before the colon.
7717 ;; Check that we're not after a token that can't precede a label.
7719 ;; Trivially succeeds when there's no preceding token.
7720 ;; Succeeds when we're at a virtual semicolon.
7729 ;; Check if we're after a label, if we're after a closing
7730 ;; paren that belong to statement, and with
7731 ;; `c-label-prefix-re'. It's done in different order
7732 ;; depending on `assume-markup' since the checks have
7733 ;; different expensiveness.
7760 ;; Might be after another label, so check it recursively.
7764 ;; Essentially the same as the
7765 ;; `c-syntactic-re-search-forward' regexp below.
7771 ;; Note: the following should work instead of the
7772 ;; narrow-to-region above. Investigate why not,
7773 ;; sometime. ACM, 2006-03-31.
7774 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7775 ;; macro-start t)
7777 ;; If the caller turned on recording for us,
7778 ;; it shouldn't apply when we check the
7779 ;; preceding label.
7780 c-record-type-identifiers)
7781 ;; A label can't start at a cpp directive. Check for
7782 ;; this, since c-forward-syntactic-ws would foul up on it.
7787 ;; Point is still at the beginning of the possible label construct.
7788 ;;
7789 ;; Check that the next nonsymbol token is ":", or that we're in one
7790 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7791 ;; arguments. FIXME: Should build this regexp from the language
7792 ;; constants.
7794 ;; public: protected: private:
7804 ;; QT double keyword like "protected slots:" or goto target.
7823 'qt-1kwd-colon
7835 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7845 t)
7848 ;; Record the identifiers in the label for fontification, unless
7849 ;; it begins with `c-label-kwds' in which case the following
7850 ;; identifiers are part of a (constant) expression that
7851 ;; shouldn't be fontified.
7863 ;; Not a label.
7865 label-type))
7868 ;; Assuming the point is at the beginning of a token, try to move
7869 ;; forward to the end of the Objective-C directive that starts
7870 ;; there. Return t if a directive was fully recognized, otherwise
7871 ;; the point is moved as far as one could be successfully parsed and
7872 ;; nil is returned.
7873 ;;
7874 ;; This function records identifier ranges on
7875 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7876 ;; `c-record-type-identifiers' is non-nil.
7877 ;;
7878 ;; This function might do hidden buffer changes.
7881 start-char
7883 lim
7884 ;; Turn off recognition of angle bracket arglists while parsing
7885 ;; types here since the protocol reference list might then be
7886 ;; considered part of the preceding name or superclass-name.
7897 t)
7906 ;; Handle the name of the class itself.
7908 ;; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7909 ;; at EOB.
7916 ;; Look for ": superclass-name" or "( category-name )".
7927 ;; Look for a protocol reference list.
7933 t))))
7939 t)
7942 nil)))
7945 ;; Go to the first non-whitespace after the colon that starts a
7946 ;; multiple inheritance introduction. Optional LIM is the farthest
7947 ;; back we should search.
7948 ;;
7949 ;; This function might do hidden buffer changes.
7957 ;; Return nil if we aren't in a method definition, otherwise the
7958 ;; position of the initial [+-].
7959 ;;
7960 ;; This function might do hidden buffer changes.
7966 ))
7968 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7970 ;; Return non-nil if point is within a gcc \"asm\" block.
7971 ;;
7972 ;; This should be called with point inside an argument list.
7973 ;;
7974 ;; Only one level of enclosing parentheses is considered, so for
7975 ;; instance `nil' is returned when in a function call within an asm
7976 ;; operand.
7977 ;;
7978 ;; This function might do hidden buffer changes.
7990 a function.
7992 More precisely, being at the top-level means that point is either
7993 outside any enclosing block (such as a function definition), or
7994 directly inside a class, namespace or other block that contains
7995 another declaration level.
7997 If point is not at the top-level (e.g. it is inside a method
7998 definition), then nil is returned. Otherwise, if point is at a
7999 top-level not enclosed within a class definition, t is returned.
8000 Otherwise, a 2-vector is returned where the zeroth element is the
8001 buffer position of the start of the class declaration, and the first
8002 element is the buffer position of the enclosing class's opening
8003 brace.
8005 Note that this function might do hidden buffer changes. See the
8006 comment at the start of cc-engine.el for more info."
8012 ;; Return non-nil if the point is in the region after the argument
8013 ;; list of a function and its opening brace (or semicolon in case it
8014 ;; got no body). If there are K&R style argument declarations in
8015 ;; that region, the point has to be inside the first one for this
8016 ;; function to recognize it.
8017 ;;
8018 ;; If successful, the point is moved to the first token after the
8019 ;; function header (see `c-forward-decl-or-cast-1' for details) and
8020 ;; the position of the opening paren of the function arglist is
8021 ;; returned.
8022 ;;
8023 ;; The point is clobbered if not successful.
8024 ;;
8025 ;; LIM is used as bound for backward buffer searches.
8026 ;;
8027 ;; This function might do hidden buffer changes.
8040 ;; There should not be a '=' or ',' between beg and the
8041 ;; start of the declaration since that means we were in the
8042 ;; "expression part" of the declaration.
8047 ;; Check that there's an arglist paren in the
8048 ;; declaration.
8051 ;; The declarator is a paren expression, so skip past it
8052 ;; so that we don't get stuck on that instead of the
8053 ;; function arglist.
8057 ;; Don't trip up on "operator ()".
8064 ;; Return the position of the first argument declaration if point is
8065 ;; inside a K&R style argument declaration list, nil otherwise.
8066 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
8067 ;; position that bounds the backward search for the argument list.
8068 ;;
8069 ;; Point must be within a possible K&R region, e.g. just before a top-level
8070 ;; "{". It must be outside of parens and brackets. The test can return
8071 ;; false positives otherwise.
8072 ;;
8073 ;; This function might do hidden buffer changes.
8077 ;; If we're in a macro, our search range is restricted to it. Narrow to
8078 ;; the searchable range.
8082 before-lparen after-rparen
8084 ; we give up
8087 ;; Search backwards for the defun's argument list. We give up if we
8088 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
8089 ;; a knr region) or BOB.
8090 ;;
8091 ;; The criterion for a paren structure being the arg list is:
8092 ;; o - there is non-WS stuff after it but before any "{"; AND
8093 ;; o - the token after it isn't a ";" AND
8094 ;; o - it is preceded by either an identifier (the function name) or
8095 ;; a macro expansion like "DEFUN (...)"; AND
8096 ;; o - its content is a non-empty comma-separated list of identifiers
8097 ;; (an empty arg list won't have a knr region).
8098 ;;
8099 ;; The following snippet illustrates these rules:
8100 ;; int foo (bar, baz, yuk)
8101 ;; int bar [] ;
8102 ;; int (*baz) (my_type) ;
8103 ;; int (*) (void) (*yuk) (void) ;
8104 ;; {
8115 ; parens/brackets.
8119 ;; We're inside a paren. Could it be our argument list....?
8125 ;; FIXME!!! What about macros between the parens? 2007/01/20
8128 ;; It can't be the arg list if next token is ; or {
8133 ;; Is the thing preceding the list an identifier (the
8134 ;; function name), or a macro expansion?
8144 ;; Have we got a non-empty list of comma-separated
8145 ;; identifiers?
8159 ;; ...Yes. We've identified the function's argument list.
8165 ;; ...No. The current parens aren't the function's arg list.
8172 ;; skip forward over conditional at point, including any predicate
8173 ;; statements in parentheses. No error checking is performed.
8174 ;;
8175 ;; This function might do hidden buffer changes.
8177 ;; else if()
8182 ;; do, else, try, finally
8187 ;; for, if, while, switch, catch, synchronized, foreach
8191 ;; If looking at the token after a conditional then return the
8192 ;; position of its start, otherwise return nil.
8193 ;;
8194 ;; This function might do hidden buffer changes.
8205 ;; If the point is after an operator identifier that isn't handled
8206 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
8207 ;; position of the start of that identifier is returned. nil is
8208 ;; returned otherwise. The point may be anywhere in the syntactic
8209 ;; whitespace after the last token of the operator identifier.
8210 ;;
8211 ;; This function might do hidden buffer changes.
8223 ;; Assuming point is at a brace that opens a statement block of some
8224 ;; kind, move to the proper anchor point for that block. It might
8225 ;; need to be adjusted further by c-add-stmt-syntax, but the
8226 ;; position at return is suitable as start position for that
8227 ;; function.
8228 ;;
8229 ;; This function might do hidden buffer changes.
8236 ;; Assuming point is at a brace that opens the block of a top level
8237 ;; declaration of some kind, move to the proper anchor point for
8238 ;; that block.
8239 ;;
8240 ;; This function might do hidden buffer changes.
8245 ;; Search forward for the end of the "header" of the current
8246 ;; declaration. That's the position where the definition body
8247 ;; starts, or the first variable initializer, or the ending
8248 ;; semicolon. I.e. search forward for the closest following
8249 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
8250 ;; _after_ the first found token, or at point-max if none is found.
8251 ;;
8252 ;; This function might do hidden buffer changes.
8257 ;; In C++ we need to take special care to handle operator
8258 ;; tokens and those pesky template brackets.
8263 ;; Handle operator identifiers, i.e. ignore any
8264 ;; operator token preceded by "operator".
8271 t
8273 nil)))))
8282 ;; Go to the beginning of the current declaration, or the beginning
8283 ;; of the previous one if already at the start of it. Point won't
8284 ;; be moved out of any surrounding paren. Return a cons cell of the
8285 ;; form (MOVE . KNR-POS). MOVE is like the return value from
8286 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
8287 ;; style argument declarations (and they are to be recognized) then
8288 ;; KNR-POS is set to the start of the first such argument
8289 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
8290 ;; position that bounds the backward search.
8291 ;;
8292 ;; NB: Cases where the declaration continues after the block, as in
8293 ;; "struct foo { ... } bar;", are currently recognized as two
8294 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
8295 ;;
8296 ;; This function might do hidden buffer changes.
8302 ;; `c-beginning-of-statement-1' stops at a block start, but we
8303 ;; want to continue if the block doesn't begin a top level
8304 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8305 ;; or an open paren.
8307 ;; Go back one "statement" each time round the loop until we're just
8308 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8309 ;; an ObjC method. This will move over a multiple declaration whose
8310 ;; components are comma separated.
8312 ;; Must check with c-opt-method-key in ObjC mode.
8322 ;; Check that we don't move from the first thing in a
8323 ;; macro to its header.
8329 move tentative-move))
8334 ;; Handle K&R argdecls. Back up after the "statement" jumped
8335 ;; over by `c-beginning-of-statement-1', unless it was the
8336 ;; function body, in which case we're sitting on the opening
8337 ;; brace now. Then test if we're in a K&R argdecl region and
8338 ;; that we started at the other side of the first argdecl in
8339 ;; it.
8349 'previous
8351 knr-argdecl-start))
8354 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8355 ;; statement, so the result will be 'previous if we've moved over any.
8356 ;; So change our result back to 'same if necessary.
8357 ;;
8358 ;; If they were brace list initializers we might not have moved over a
8359 ;; declaration boundary though, so change it to 'same if we've moved
8360 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8361 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8362 ;; potentially can search over a large amount of text.). Take special
8363 ;; pains not to get mislead by C++'s "operator=", and the like.
8366 c++-template-syntax-table
8372 ; no more, or get to one which isn't an "operator ="
8375 c-overloadable-operators-regexp
8376 c-opt-op-identifier-prefix
8391 ;; Assuming point is at the start of a declaration (as detected by
8392 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8393 ;; `c-beginning-of-decl-1', this function handles the case when a
8394 ;; block is followed by identifiers in e.g. struct declarations in C
8395 ;; or C++. If a proper end was found then t is returned, otherwise
8396 ;; point is moved as far as possible within the current sexp and nil
8397 ;; is returned. This function doesn't handle macros; use
8398 ;; `c-end-of-macro' instead in those cases.
8399 ;;
8400 ;; This function might do hidden buffer changes.
8403 c++-template-syntax-table
8411 ;; Stopped at the ';' in a K&R argdecl section which is
8412 ;; detected using the same criteria as in
8413 ;; `c-beginning-of-decl-1'. Move to the following block
8414 ;; start.
8418 ;; Encountered a block in the declaration. Jump over it.
8429 ;; Check for `c-opt-block-decls-with-vars-key'
8430 ;; before the first paren.
8433 c-opt-block-decls-with-vars-key
8435 lim t t t)
8438 ;; Check that the first following paren is
8439 ;; the block.
8441 lim t t t)
8443 ;; The declaration doesn't have any of the
8444 ;; `c-opt-block-decls-with-vars' keywords in the
8445 ;; beginning, so it ends here at the end of the block.
8453 nil)))
8456 ;; Assuming the point is at an open brace, check if it starts a
8457 ;; block that contains another declaration level, i.e. that isn't a
8458 ;; statement block or a brace list, and if so return non-nil.
8459 ;;
8460 ;; If the check is successful, the return value is the start of the
8461 ;; keyword that tells what kind of construct it is, i.e. typically
8462 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8463 ;; the point will be at the start of the construct, before any
8464 ;; leading specifiers, otherwise it's at the returned position.
8465 ;;
8466 ;; The point is clobbered if the check is unsuccessful.
8467 ;;
8468 ;; CONTAINING-SEXP is the position of the open of the surrounding
8469 ;; paren, or nil if none.
8470 ;;
8471 ;; The optional LIMIT limits the backward search for the start of
8472 ;; the construct. It's assumed to be at a syntactically relevant
8473 ;; position.
8474 ;;
8475 ;; If any template arglists are found in the searched region before
8476 ;; the open brace, they get marked with paren syntax.
8477 ;;
8478 ;; This function might do hidden buffer changes.
8485 ;; Could be at the end of a template arglist.
8493 ;; Note: Can't get bogus hits inside template arglists below since they
8494 ;; have gotten paren syntax above.
8496 ;; If `goto-start' is set we begin by searching for the
8497 ;; first possible position of a leading specifier list.
8498 ;; The `c-decl-block-key' search continues from there since
8499 ;; we know it can't match earlier.
8502 open-brace t t)
8504 t)
8505 t)
8511 ;; Exclude cases where we matched what would ordinarily
8512 ;; be a block declaration keyword, except where it's not
8513 ;; legal because it's part of a "compound keyword" like
8514 ;; "enum class". Of course, if c-after-brace-list-key
8515 ;; is nil, we can skip the test.
8525 ;; Found a keyword that can't be a type?
8528 ;; Can be a type too, in which case it's the return type of a
8529 ;; function (under the assumption that no declaration level
8530 ;; block construct starts with a type).
8533 ;; Jumped over a type, but it could be a declaration keyword
8534 ;; followed by the declared identifier that we've jumped over
8535 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8536 ;; then we should be at the declarator now, so check for a
8537 ;; valid declarator start.
8538 ;;
8539 ;; Note: This doesn't cope with the case when a declared
8540 ;; identifier is followed by e.g. '(' in a language where '('
8541 ;; also might be part of a declarator expression. Currently
8542 ;; there's no such language.
8546 ;; In Pike a list of modifiers may be followed by a brace
8547 ;; to make them apply to many identifiers. Note that the
8548 ;; match data will be empty on return in this case.
8554 ;; Use this variant to avoid yet another special regexp.
8558 t)))
8560 ;; Got a match.
8563 ;; Back up over any preceding specifiers and their clauses
8564 ;; by going forward from `first-specifier-pos', which is the
8565 ;; earliest possible position where the specifier list can
8566 ;; start.
8572 ;; Accept any plain symbol token on the ground that
8573 ;; it's a specifier masked through a macro (just
8574 ;; like `c-forward-decl-or-cast-1' skip forward over
8575 ;; such tokens).
8576 ;;
8577 ;; Could be more restrictive wrt invalid keywords,
8578 ;; but that'd only occur in invalid code so there's
8579 ;; no use spending effort on it.
8585 ;; Can't parse a declaration preamble and is still
8586 ;; before `kwd-start'. That means `first-specifier-pos'
8587 ;; was in some earlier construct. Search again.
8591 ;; Got no preamble before the block declaration keyword.
8597 kwd-start)))
8600 ;; Check if the closest containing paren sexp is a declaration
8601 ;; block, returning a 2 element vector in that case. Aref 0
8602 ;; contains the bufpos at boi of the class key line, and aref 1
8603 ;; contains the bufpos of the open brace. This function is an
8604 ;; obsolete wrapper for `c-looking-at-decl-block'.
8605 ;;
8606 ;; This function might do hidden buffer changes.
8614 nil))
8619 ;; Return the buffer position of the most enclosing decl-block brace (in the
8620 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8621 ;; none was found.
8630 open-brace))
8633 ;; Return the position of the L-brace if point is inside a brace list
8634 ;; initialization of an array, etc. This is an approximate function,
8635 ;; designed for speed over accuracy. It will not find every bracelist, but
8636 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8637 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8638 ;; is everywhere else.
8647 b-pos)))
8650 ;; We're at the token after what might be a type prefixed with a colon. Try
8651 ;; moving backward over this type and the colon. On success, return t and
8652 ;; leave point before colon; on failure, leave point unchanged. Will clobber
8653 ;; match data.
8670 nil)
8672 t)
8674 t)
8678 t)))
8681 ;; We're at a "{". Move back to the enum-like keyword that starts this
8682 ;; declaration and return t, otherwise don't move and return nil.
8684 up-sexp-pos before-identifier)
8700 t)
8706 t)
8712 ;; return the buffer position of the beginning of the brace list
8713 ;; statement if we're inside a brace list, otherwise return nil.
8714 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8715 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8716 ;; braces
8717 ;;
8718 ;; N.B.: This algorithm can potentially get confused by cpp macros
8719 ;; placed in inconvenient locations. It's a trade-off we make for
8720 ;; speed.
8721 ;;
8722 ;; This function might do hidden buffer changes.
8724 ;; This will pick up brace list declarations.
8728 ;; this will pick up array/aggregate init lists, even if they are nested.
8731 ;; Pike can have class definitions anywhere, so we must
8732 ;; check for the class key here.
8734 c-decl-block-key))
8735 bufpos braceassignp lim next-containing macro-start)
8737 containing-sexp)
8748 ;; We're in an in-expression block of some kind. Do not
8749 ;; check nesting. We deliberately set the limit to the
8750 ;; containing sexp, so that c-looking-at-inexpr-block
8751 ;; doesn't check for an identifier before it.
8753 ;; see if the open brace is preceded by = or [...] in
8754 ;; this statement, but watch out for operator=
8757 ;; Checks to do only on the first sexp before the brace.
8760 ;; In Java, an initialization brace list may follow
8761 ;; directly after "new Foo[]", so check for a "new"
8762 ;; earlier.
8768 ;; Carry on looking if this is an
8769 ;; identifier (may contain "." in Java)
8770 ;; or another "[]" sexp.
8773 ;; Checks to do on all sexps before the brace, up to the
8774 ;; beginning of the statement.
8782 ;; We've seen a =, but must check earlier tokens so
8783 ;; that it isn't something that should be ignored.
8789 ;; Check for operator =
8792 nil)
8793 ;; Check for `<opchar>= in Pike.
8796 ;; Special case for Pikes
8797 ;; `[]=, since '[' is not in
8798 ;; the punctuation class.
8801 nil)
8803 ;; make sure we're not in a C++ template
8804 ;; argument assignment
8816 ))))
8817 nil)
8824 ;; We've hit the beginning of the aggregate list.
8829 ;; Brace lists can't contain a semicolon, so we're done.
8834 ;; We've a macro whose expansion starts with the '{'.
8835 ;; Heuristically, if we have a ';' in it we've not got a
8836 ;; brace list, otherwise we have.
8843 containing-sexp nil)
8846 ;; Go up one level
8848 lim nil
8849 next-containing nil)))))
8851 bufpos))
8852 ))
8855 ;; If we're looking at the start of a pike-style list, i.e., `({Â })',
8856 ;; `([Â ])', `(<Â >)', etc., a cons of a cons of its starting and ending
8857 ;; positions and its entry in c-special-brace-lists is returned, nil
8858 ;; otherwise. The ending position is nil if the list is still open.
8859 ;; LIM is the limit for forward search. The point may either be at
8860 ;; the `(' or at the following paren character. Tries to check the
8861 ;; matching closer, but assumes it's correct if no balanced paren is
8862 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8863 ;; a special brace list).
8864 ;;
8865 ;; This function might do hidden buffer changes.
8870 inner-beg end type)
8885 nil)))))
8895 ;; Check balancing of the inner paren
8896 ;; below.
8899 t)
8900 ;; If the inner char isn't a paren then
8901 ;; we can't check balancing, so just
8902 ;; check the char before the outer
8903 ;; closing paren.
8918 ;; Return non-nil if between two statements or declarations, assuming
8919 ;; point is not inside a literal or comment.
8920 ;;
8921 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8922 ;; are recommended instead.
8923 ;;
8924 ;; This function might do hidden buffer changes.
8929 ;; Return non-nil if we're looking at the beginning of a block
8930 ;; inside an expression. The value returned is actually a cons of
8931 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8932 ;; position of the beginning of the construct.
8933 ;;
8934 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8935 ;; position of the closest containing list. If it's nil, the
8936 ;; containing paren isn't used to decide whether we're inside an
8937 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8938 ;; needs to be farther back.
8939 ;;
8940 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8941 ;; brace block might be done. It should only be used when the
8942 ;; construct can be assumed to be complete, i.e. when the original
8943 ;; starting position was further down than that.
8944 ;;
8945 ;; This function might do hidden buffer changes.
8950 ;; Look at the character after point only as a last resort
8951 ;; when we can't disambiguate.
8971 ;; If the class definition is at the start of
8972 ;; a statement, we don't consider it an
8973 ;; in-expression class.
8981 ;; Also, in Pike we treat it as an
8982 ;; in-expression class if it's used in an
8983 ;; object clone expression.
8999 nil)
9007 ;; Accept several square bracket sexps for
9008 ;; Java array initializations.
9016 block-follows
9017 containing-sexp
9026 nil
9029 res))))
9032 ;; Returns non-nil if we're looking at the end of an in-expression
9033 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
9034 ;; PAREN-STATE is the paren state relevant at the current position.
9035 ;;
9036 ;; This function might do hidden buffer changes.
9038 ;; We currently only recognize a block.
9041 containing-sexp)
9050 paren-state)
9051 containing-sexp)))))
9054 ;; Is there a "virtual semicolon" at POS or point?
9055 ;; (See cc-defs.el for full details of "virtual semicolons".)
9056 ;;
9057 ;; This is true when point is at the last non syntactic WS position on the
9058 ;; line, there is a macro call last on the line, and this particular macro's
9059 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
9060 ;; semicolon.
9068 c-macro-with-semi-re
9071 ;; Check we've got nothing after this except comments and empty lines
9072 ;; joined by escaped EOLs.
9085 nil)
9093 ;; Check for one of the listed macros being before point.
9097 t))
9105 \f
9106 ;; `c-guess-basic-syntax' and the functions that precedes it below
9107 ;; implements the main decision tree for determining the syntactic
9108 ;; analysis of the current line of code.
9110 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
9111 ;; auto newline analysis.
9115 ;; BRACEPOS is the position of a brace in a construct like "namespace
9116 ;; Bar {". Return the anchor point in this construct; this is the
9117 ;; earliest symbol on the brace's line which isn't earlier than
9118 ;; "namespace".
9119 ;;
9120 ;; Currently (2007-08-17), "like namespace" means "matches
9121 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
9122 ;; or anything like that.
9132 ;; A simple function to prepend a new syntax element to
9133 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
9134 ;; should always be dynamically bound but since we read it first
9135 ;; we'll fail properly anyway if this function is misused.
9137 c-syntactic-context)))
9140 ;; Like `c-add-syntax' but appends to the end of the syntax list.
9141 ;; (Normally not necessary.)
9146 syntax-extra-args
9147 stop-at-boi-only
9148 containing-sexp
9149 paren-state)
9150 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
9151 ;; needed with further syntax elements of the types `substatement',
9152 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
9153 ;; `defun-block-intro'.
9154 ;;
9155 ;; Do the generic processing to anchor the given syntax symbol on
9156 ;; the preceding statement: Skip over any labels and containing
9157 ;; statements on the same line, and then search backward until we
9158 ;; find a statement or block start that begins at boi without a
9159 ;; label or comment.
9160 ;;
9161 ;; Point is assumed to be at the prospective anchor point for the
9162 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
9163 ;; skip past open parens and containing statements. Most of the added
9164 ;; syntax elements will get the same anchor point - the exception is
9165 ;; for an anchor in a construct like "namespace"[*] - this is as early
9166 ;; as possible in the construct but on the same line as the {.
9167 ;;
9168 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
9169 ;;
9170 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
9171 ;; syntax symbol. They are appended after the anchor point.
9172 ;;
9173 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
9174 ;; if the current statement starts there.
9175 ;;
9176 ;; Note: It's not a problem if PAREN-STATE "overshoots"
9177 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
9178 ;;
9179 ;; This function might do hidden buffer changes.
9182 ;; This is by far the most common case, so let's give it special
9183 ;; treatment.
9188 ;; Set when we're on a label, so that we don't stop there.
9189 ;; FIXME: To be complete we should check if we're on a label
9190 ;; now at the start.
9191 on-label)
9193 ;; Use point as the anchor point for "namespace", "extern", etc.
9197 syntax-extra-args)
9199 ;; Loop while we have to back out of containing blocks.
9204 ;; Loop while we have to back up statements.
9206 on-label
9209 ;; Skip past any comments that stands between the
9210 ;; statement start and boi.
9218 ;; Skip to the beginning of this statement or backward
9219 ;; another one.
9227 ;; If we didn't move we're at the start of a block and
9228 ;; have to continue outside it.
9237 ;; Special case to avoid deeper and deeper indentation
9238 ;; of "else if" clauses.
9239 )
9244 ;; If stop-at-boi-only is nil, we shouldn't back up
9245 ;; over previous or containing statements to try to
9246 ;; reach boi, so go back to the last position and
9247 ;; exit.
9254 ;; If we've moved into another statement then we
9255 ;; should no longer try to stop in the middle of a
9256 ;; line.
9259 ;; Record this as a substatement if we skipped up one
9260 ;; level.
9263 )))
9265 containing-sexp)
9267 ;; Now we have to go out of this block.
9270 ;; Don't stop in the middle of a special brace list opener
9271 ;; like "({".
9283 ;; Analyze the construct in front of the block we've stepped out
9284 ;; from and add the right syntactic element for it.
9287 step-type)
9290 ;; Stepped out of a parenthesis block, so we're in an
9291 ;; expression now.
9303 ;; Stepped out of an in-expression statement. This
9304 ;; syntactic element won't get an anchor pos.
9307 ;; A parenthesis normally belongs to an arglist.
9315 on-label nil))
9317 ;; Stepped out of a brace block.
9329 containing-sexp)))
9331 'defun-block-intro
9333 nil))
9337 c-other-decl-block-key-in-symbols-alist))
9344 ;; Always done if the open brace was at boi. The
9345 ;; c-beginning-of-statement-1 call above is necessary
9346 ;; anyway, to decide the type of block-intro to add.
9349 ))
9351 ;; Fill in the current point as the anchor for all the symbols
9352 ;; added above.
9361 )))
9364 containing-decl-open
9365 containing-decl-start
9366 containing-decl-kwd
9367 paren-state)
9368 ;; The inclass and class-close syntactic symbols are added in
9369 ;; several places and some work is needed to fix everything.
9370 ;; Therefore it's collected here.
9371 ;;
9372 ;; This function might do hidden buffer changes.
9377 containing-decl-open)
9379 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9380 ;; here, but we have to do like this for compatibility.
9390 char-after-ip
9391 beg-of-same-or-containing-stmt
9392 containing-sexp
9393 paren-state)
9394 ;; This function contains the decision tree reached through both
9395 ;; cases 18 and 10. It's a continued statement or top level
9396 ;; construct of some kind.
9397 ;;
9398 ;; This function might do hidden buffer changes.
9405 ;; (CASE A removed.)
9406 ;; CASE B: open braces for class or brace-lists
9413 ;; CASE B.1: class-open
9420 ;; CASE B.2: brace-list-open
9425 indent-point t t t)))
9426 ;; The most semantically accurate symbol here is
9427 ;; brace-list-open, but we normally report it simply as a
9428 ;; statement-cont. The reason is that one normally adjusts
9429 ;; brace-list-open for brace lists as top-level constructs,
9430 ;; and brace lists inside statements is a completely different
9431 ;; context. C.f. case 5A.3.
9434 ;; Turn off the dwim above when we're
9435 ;; analyzing the nature of the brace
9436 ;; for the auto newline feature.
9437 'brace-list-open
9439 nil nil
9440 containing-sexp paren-state))
9442 ;; CASE B.3: The body of a function declared inside a normal
9443 ;; block. Can occur e.g. in Pike and when using gcc
9444 ;; extensions, but watch out for macros followed by blocks.
9445 ;; C.f. cases E, 16F and 17G.
9451 ;; Turn off recognition of constructs that lacks a
9452 ;; type in this case, since that's more likely to be
9453 ;; a macro followed by a block.
9456 containing-sexp paren-state))
9458 ;; CASE B.4: Continued statement with block open. The most
9459 ;; accurate analysis is perhaps `statement-cont' together with
9460 ;; `block-open' but we play DWIM and use `substatement-open'
9461 ;; instead. The rationale is that this typically is a macro
9462 ;; followed by a block which makes it very similar to a
9463 ;; statement with a substatement block.
9466 containing-sexp paren-state))
9467 ))
9469 ;; CASE C: iostream insertion or extraction operator
9473 ;; If there is no preceding streamop in the statement
9474 ;; then indent this line as a normal statement-cont.
9478 t))))
9480 ;; CASE E: In the "K&R region" of a function declared inside a
9481 ;; normal block. C.f. case B.3.
9483 ;; Check that the next token is a '{'. This works as
9484 ;; long as no language that allows nested function
9485 ;; definitions allows stuff like member init lists, K&R
9486 ;; declarations or throws clauses there.
9487 ;;
9488 ;; Note that we do a forward search for something ahead
9489 ;; of the indentation line here. That's not good since
9490 ;; the user might not have typed it yet. Unfortunately
9491 ;; it's exceedingly tricky to recognize a function
9492 ;; prototype in a code block without resorting to this.
9500 ;; Turn off recognition of constructs that lacks a
9501 ;; type in this case, since that's more likely to be
9502 ;; a macro followed by a block.
9505 containing-sexp paren-state))
9507 ;;CASE F: continued statement and the only preceding items are
9508 ;;annotations.
9516 t)
9523 ;; CASE G: a template list continuation?
9524 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9542 ;; In a nested template arglist.
9548 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9549 ;; template aware.
9552 ;; CASE D: continued statement.
9556 containing-sexp paren-state))
9557 )))
9559 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9560 ;; 2005/11/29).
9561 ;;;###autoload
9563 "Return the syntactic context of the current line."
9569 open-paren-in-column-0-is-defun-start
9570 ;; A whole ugly bunch of various temporary variables. Have
9571 ;; to declare them here since it's not possible to declare
9572 ;; a variable with only the scope of a cond test and the
9573 ;; following result clauses, and most of this function is a
9574 ;; single gigantic cond. :P
9575 literal char-before-ip before-ws-ip char-after-ip macro-start
9576 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9577 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9578 containing-<
9579 ;; The following record some positions for the containing
9580 ;; declaration block if we're directly within one:
9581 ;; `containing-decl-open' is the position of the open
9582 ;; brace. `containing-decl-start' is the start of the
9583 ;; declaration. `containing-decl-kwd' is the keyword
9584 ;; symbol of the keyword that tells what kind of block it
9585 ;; is.
9586 containing-decl-open
9587 containing-decl-start
9588 containing-decl-kwd
9589 ;; The open paren of the closest surrounding sexp or nil if
9590 ;; there is none.
9591 containing-sexp
9592 ;; The position after the closest preceding brace sexp
9593 ;; (nested sexps are ignored), or the position after
9594 ;; `containing-sexp' if there is none, or (point-min) if
9595 ;; `containing-sexp' is nil.
9596 lim
9597 ;; The paren state outside `containing-sexp', or at
9598 ;; `indent-point' if `containing-sexp' is nil.
9600 ;; There's always at most one syntactic element which got
9601 ;; an anchor pos. It's stored in syntactic-relpos.
9602 syntactic-relpos
9605 ;; Check if we're directly inside an enclosing declaration
9606 ;; level block.
9615 containing-sexp)
9616 t)))
9619 containing-sexp nil)
9624 ;; Init some position variables.
9637 ;; Ignore balanced paren. The next entry
9638 ;; can't be another one.
9641 ;; If there is no surrounding open paren then
9642 ;; put the last balanced pair back on paren-state.
9644 containing-sexp nil)))
9648 ;; If we're in a parenthesis list then ',' delimits the
9649 ;; "statements" rather than being an operator (with the
9650 ;; exception of the "for" clause). This difference is
9651 ;; typically only noticeable when statements are used in macro
9652 ;; arglists.
9656 ;; cache char before and after indent point, and move point to
9657 ;; the most likely position to perform the majority of tests
9666 ;; are we in a literal?
9669 ;; now figure out syntactic qualities of the current line
9672 ;; CASE 1: in a string.
9676 ;; CASE 2: in a C or C++ style comment.
9678 ;; This is a kludge for XEmacs where we use
9679 ;; `buffer-syntactic-context', which doesn't correctly
9680 ;; recognize "\*/" to end a block comment.
9681 ;; `parse-partial-sexp' which is used by
9682 ;; `c-literal-limits' will however do that in most
9683 ;; versions, which results in that we get nil from
9684 ;; `c-literal-limits' even when `c-in-literal' claims
9685 ;; we're inside a comment.
9689 ;; CASE 3: in a cpp preprocessor macro continuation.
9699 ;; If at the beginning of the body of a #define
9700 ;; directive then analyze as cpp-define-intro
9701 ;; only. Go on with the syntactic analysis
9702 ;; otherwise. in-macro-expr is set if we're in a
9703 ;; cpp expression, i.e. before the #define body
9704 ;; or anywhere in a non-#define directive.
9711 nil)))))
9715 ;; CASE 11: an else clause?
9719 containing-sexp paren-state))
9721 ;; CASE 12: while closure of a do/while construct?
9729 containing-sexp paren-state))
9731 ;; CASE 13: A catch or finally clause? This case is simpler
9732 ;; than if-else and do-while, because a block is required
9733 ;; after every try, catch and finally.
9741 t)
9745 t)
9748 t))
9753 containing-sexp paren-state))
9755 ;; CASE 18: A substatement we can recognize by keyword.
9765 ;; Ought to cache the result from the
9766 ;; c-beginning-of-statement-1 calls here.
9776 ;; Step up to the containing statement if we
9777 ;; stayed in the same one.
9785 ;; There was no containing statement after all.
9787 placeholder))
9789 ;; Require a parenthesis after these keywords.
9790 ;; Necessary to catch e.g. synchronized in Java,
9791 ;; which can be used both as statement and
9792 ;; modifier.
9798 ;; CASE 18A: Simple substatement.
9804 containing-sexp paren-state))
9810 containing-sexp paren-state))
9813 containing-sexp paren-state))))
9815 ;; CASE 18B: Some other substatement. This is shared
9816 ;; with case 10.
9818 char-after-ip
9819 placeholder
9820 lim
9821 paren-state)))
9823 ;; CASE 14: A case or default label
9829 containing-sexp))
9832 ;; Got a bogus label at the top level. In lack of better
9833 ;; alternatives, anchor it on (point-min).
9836 ;; CASE 15: any other label
9843 containing-decl-open
9844 containing-decl-start
9845 containing-decl-kwd
9846 paren-state))
9847 ;; Append access-label with the same anchor point as
9848 ;; inclass gets.
9854 containing-sexp))
9859 ;; If the surrounding statement is a switch then
9860 ;; let's analyze all labels as switch labels, so
9861 ;; that they get lined up consistently.
9862 'case-label
9868 ;; A label on the top level. Treat it as a class
9869 ;; context. (point-min) is the closest we get to the
9870 ;; class open brace.
9873 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9874 ;; 17E.
9877 containing-sexp
9878 ;; Have to turn on the heuristics after
9879 ;; the point even though it doesn't work
9880 ;; very well. C.f. test case class-16.pike.
9881 t))
9886 ;; It's a statement block or an anonymous class.
9888 ;; It's a Pike lambda. Check whether we are between the
9889 ;; lambda keyword and the argument list or at the defun
9890 ;; opener.
9892 'inline-open
9898 paren-state)
9902 ;; CASE 5: Line is inside a declaration level block or at top level.
9906 ;; CASE 5A: we are looking at a defun, brace list, class,
9907 ;; or inline-inclass method opening brace
9914 ;; CASE 5A.1: Non-class declaration block open.
9929 ;; Special case for extern-lang-open.
9930 'extern-lang-open
9932 nil t containing-sexp paren-state))
9934 ;; CASE 5A.2: we are looking at a class opening brace
9943 ;; CASE 5A.3: brace list open
9968 ))))
9972 ;; We're in Java and have found that the open brace
9973 ;; belongs to a "new Foo[]" initialization list,
9974 ;; which means the brace list is part of an
9975 ;; expression and not a top level definition. We
9976 ;; therefore treat it as any topmost continuation
9977 ;; even though the semantically correct symbol still
9978 ;; is brace-list-open, on the same grounds as in
9979 ;; case B.2.
9985 ;; CASE 5A.4: inline defun open
9991 containing-decl-open
9992 containing-decl-start
9993 containing-decl-kwd
9994 paren-state))
9996 ;; CASE 5A.5: ordinary defun open
10004 ;; Bogus to use bol here, but it's the legacy. (Resolved,
10005 ;; 2007-11-09)
10006 ))))
10008 ;; CASE 5R: Member init list. (Used to be part of CASE 5B.1)
10009 ;; Note there is no limit on the backward search here, since member
10010 ;; init lists can, in practice, be very large.
10016 ;; There is no preceding member init clause.
10017 ;; Indent relative to the beginning of indentation
10018 ;; for the topmost-intro line that contains the
10019 ;; prototype's open paren.
10022 ;; Indent relative to the first member init clause.
10027 ;; CASE 5B: After a function header but before the body (or
10028 ;; the ending semicolon if there's no body).
10035 ;; CASE 5B.1: Member init list.
10037 ;; There is no preceding member init clause.
10038 ;; Indent relative to the beginning of indentation
10039 ;; for the topmost-intro line that contains the
10040 ;; prototype's open paren.
10044 ;; CASE 5B.2: K&R arg decl intro
10051 containing-decl-open
10052 containing-decl-start
10053 containing-decl-kwd
10054 paren-state)))
10056 ;; CASE 5B.4: Nether region after a C++ or Java func
10057 ;; decl, which could include a `throws' declaration.
10061 )))
10063 ;; CASE 5C: inheritance line. could be first inheritance
10064 ;; line, or continuation of a multiple inheritance
10072 ;; watch out for scope operator
10077 )))
10086 ;; for Java
10091 cont done)
10096 done t))
10100 )
10102 injava-inher)
10105 ))
10108 ;; CASE 5C.1: non-hanging colon on an inher intro
10112 ;; don't add inclass symbol since relative point already
10113 ;; contains any class offset
10114 )
10116 ;; CASE 5C.2: hanging colon on an inher intro
10122 containing-decl-open
10123 containing-decl-start
10124 containing-decl-kwd
10125 paren-state)))
10127 ;; CASE 5C.3: in a Java implements/extends
10141 )))
10143 ;; CASE 5C.4: a continued inheritance line
10147 ;; don't add inclass symbol since relative point already
10148 ;; contains any class offset
10149 )))
10151 ;; CASE 5P: AWK pattern or function or continuation
10152 ;; thereof.
10158 'topmost-intro-cont
10160 nil nil
10161 containing-sexp paren-state))
10163 ;; CASE 5D: this could be a top-level initialization, a
10164 ;; member init list continuation, or a template argument
10165 ;; list continuation.
10167 ;; Note: We use the fact that lim is always after any
10168 ;; preceding brace sexp.
10178 t))
10183 t)
10189 nil)
10190 t))
10192 ;; NB: No c-after-special-operator-id stuff in this
10193 ;; clause - we assume only C++ needs it.
10198 ;; CASE 5D.3: perhaps a template list continuation?
10213 ;; In a nested template arglist.
10219 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
10220 ;; template aware.
10223 ;; CASE 5D.4: perhaps a multiple inheritance line?
10235 t)
10240 ;; CASE 5D.5: Continuation of the "expression part" of a
10241 ;; top level construct. Or, perhaps, an unrecognized construct.
10253 ;; A preceding comma at the top level means that a
10254 ;; new variable declaration starts here. Use
10255 ;; topmost-intro-cont for it, for consistency with
10256 ;; the first variable declaration. C.f. case 5N.
10259 nil nil containing-sexp paren-state))
10260 ))
10262 ;; CASE 5F: Close of a non-class declaration level block.
10266 ;; This is inconsistent: Should use `containing-decl-open'
10267 ;; here if it's at boi, like in case 5J.
10271 ;; Special case for compatibility with the
10272 ;; extern-lang syntactic symbols.
10273 'extern-lang-close
10276 nil t
10278 paren-state))
10280 ;; CASE 5G: we are looking at the brace which closes the
10281 ;; enclosing nested class decl
10286 containing-decl-open
10287 containing-decl-start
10288 containing-decl-kwd
10289 paren-state))
10291 ;; CASE 5H: we could be looking at subsequent knr-argdecls
10298 ;; Do an extra check to avoid tripping up on
10299 ;; statements that occur in invalid contexts
10300 ;; (e.g. in macro bodies where we don't really
10301 ;; know the context of what we're looking at).
10308 ;; CASE 5I: ObjC method definition.
10313 ;; Handle the case when it's the first (non-comment)
10314 ;; thing in the buffer. Can't look for a 'same return
10315 ;; value from cbos1 since ObjC directives currently
10316 ;; aren't recognized fully, so that we get 'same
10317 ;; instead of 'previous if it moved over a preceding
10318 ;; directive.
10322 ;; CASE 5N: At a variable declaration that follows a class
10323 ;; definition or some other block declaration that doesn't
10324 ;; end at the closing '}'. C.f. case 5D.5.
10333 ;; Speed up the backward search a bit.
10338 ;; The '}' is unbalanced.
10339 nil
10344 containing-sexp paren-state))
10346 ;; NOTE: The point is at the end of the previous token here.
10348 ;; CASE 5J: we are at the topmost level, make
10349 ;; sure we skip back past any access specifiers
10351 ;; A macro continuation line is never at top level.
10373 nil)))))
10374 ;; For historic reasons we anchor at bol of the last
10375 ;; line of the previous declaration. That's clearly
10376 ;; highly bogus and useless, and it makes our lives hard
10377 ;; to remain compatible. :P
10388 ;; Special case for compatibility with the
10389 ;; extern-lang syntactic symbols.
10390 'inextern-lang
10393 nil t
10395 paren-state))
10397 containing-decl-open
10398 containing-decl-start
10399 containing-decl-kwd
10400 paren-state)))
10402 macro-start
10407 ;; CASE 5K: we are at an ObjC method definition
10408 ;; continuation line.
10417 ;; CASE 5L: we are at the first argument of a template
10418 ;; arglist that begins on the previous line.
10426 ;; CASE 5Q: we are at a statement within a macro.
10431 ;;CASE 5N: We are at a topmost continuation line and the only
10432 ;;preceding items are annotations.
10439 t)
10445 ;; CASE 5M: we are at a topmost continuation line
10456 ))
10458 ;; (CASE 6 has been removed.)
10460 ;; CASE 7: line is an expression, not a statement. Most
10461 ;; likely we are either in a function prototype or a function
10462 ;; call argument list
10470 ;; CASE 7A: we are looking at the arglist closing paren.
10471 ;; C.f. case 7F.
10483 paren-state))
10485 ;; CASE 7B: Looking at the opening brace of an
10486 ;; in-expression block or brace list. C.f. cases 4, 16A
10487 ;; and 17E.
10491 paren-state))
10495 placeholder
10498 paren-state)
10499 containing-sexp)))
10500 ;; placeholder is nil if it's a block directly in
10501 ;; a function arglist. That makes us skip out of
10502 ;; this case.
10503 )))
10508 paren-state)
10512 ;; CASE 7C: we are looking at the first argument in an empty
10513 ;; argument list. Use arglist-close if we're actually
10514 ;; looking at a close paren or bracket.
10526 paren-state))
10528 ;; CASE 7D: we are inside a conditional test clause. treat
10529 ;; these things as statements
10539 ))
10541 ;; CASE 7E: maybe a continued ObjC method call. This is the
10542 ;; case when we are inside a [] bracketed exp, and what
10543 ;; precede the opening bracket is not an identifier.
10551 )))
10553 ;; CASE 7F: we are looking at an arglist continuation line,
10554 ;; but the preceding argument is on the same line as the
10555 ;; opening paren. This case includes multi-line
10556 ;; mathematical paren groupings, but we could be on a
10557 ;; for-list continuation line. C.f. case 7A.
10574 paren-state))
10576 ;; CASE 7G: we are looking at just a normal arglist
10577 ;; continuation line
10580 ))
10582 ;; CASE 8: func-local multi-inheritance line
10592 ;; CASE 8A: non-hanging colon on an inher intro
10597 ;; CASE 8B: hanging colon on an inher intro
10601 ;; CASE 8C: a continued inheritance line
10605 )))
10607 ;; CASE 9: we are inside a brace-list
10617 ;; CASE 9A: In the middle of a special brace list opener.
10637 ;; CASE 9B: brace-list-close brace
10639 ;; Check special brace list closer.
10646 ;; We were between the special close char and the `)'.
10649 ;; We were before the special close char.
10653 ;; Normal brace list check.
10664 ;; Prepare for the rest of the cases below by going to the
10665 ;; token following the opening brace
10678 ;; CASE 9C: we're looking at the first line in a brace-list
10689 ;; CASE 9D: this is just a later brace-list-entry or
10690 ;; brace-entry-open
10699 ))
10700 ))))
10702 ;; CASE 10: A continued statement or top level construct.
10712 ;; This is shared with case 18.
10714 char-after-ip
10715 placeholder
10716 containing-sexp
10717 paren-state))
10719 ;; CASE 16: block close brace, possibly closing the defun or
10720 ;; the class
10722 ;; From here on we have the next containing sexp in lim.
10727 ;; CASE 16E: Closing a statement block? This catches
10728 ;; cases where it's preceded by a statement keyword,
10729 ;; which works even when used in an "invalid" context,
10730 ;; e.g. a macro argument.
10735 ;; CASE 16A: closing a lambda defun or an in-expression
10736 ;; block? C.f. cases 4, 7B and 17E.
10739 nil))
10741 'inline-close
10751 paren-state)
10755 ;; CASE 16B: does this close an inline or a function in
10756 ;; a non-class declaration level block?
10763 nil))
10773 ;; CASE 16F: Can be a defun-close of a function declared
10774 ;; in a statement block, e.g. in Pike or when using gcc
10775 ;; extensions, but watch out for macros followed by
10776 ;; blocks. Let it through to be handled below.
10777 ;; C.f. cases B.3 and 17G.
10783 ;; Turn off recognition of constructs that
10784 ;; lacks a type in this case, since that's more
10785 ;; likely to be a macro followed by a block.
10792 ;; CASE 16C: If there is an enclosing brace then this is
10793 ;; a block close since defun closes inside declaration
10794 ;; level blocks have been handled above.
10796 ;; If the block is preceded by a case/switch label on
10797 ;; the same line, we anchor at the first preceding label
10798 ;; at boi. The default handling in c-add-stmt-syntax
10799 ;; really fixes it better, but we do like this to keep
10800 ;; the indentation compatible with version 5.28 and
10801 ;; earlier. C.f. case 17H.
10808 ;; c-backward-to-block-anchor not necessary here; those
10809 ;; situations are handled in case 16E above.
10812 ;; CASE 16D: Only top level defun close left.
10818 paren-state))
10819 ))
10821 ;; CASE 19: line is an expression, not a statement, and is directly
10822 ;; contained by a template delimiter. Most likely, we are in a
10823 ;; template arglist within a statement. This case is based on CASE
10824 ;; 7. At some point in the future, we may wish to create more
10825 ;; syntactic symbols such as `template-intro',
10826 ;; `template-cont-nonempty', etc., and distinguish between them as we
10827 ;; do for `arglist-intro' etc. (2009-12-07).
10833 ; '<') before indent-point.
10841 paren-state))
10843 ;; CASE 17: Statement or defun catchall.
10846 ;; Back up statements until we find one that starts at boi.
10849 containing-sexp)))
10853 nil)
10858 ;; CASE 17B: continued statement
10862 containing-sexp paren-state))
10864 ;; CASE 17A: After a case/default label?
10872 'statement-case-open
10874 nil t containing-sexp paren-state))
10876 ;; CASE 17D: any old statement
10883 containing-sexp paren-state)
10887 ;; CASE 17I: Inside a substatement block.
10889 ;; The following tests are all based on containing-sexp.
10891 ;; From here on we have the next containing sexp in lim.
10896 lim paren-state)
10900 ;; CASE 17E: first statement in an in-expression block.
10901 ;; C.f. cases 4, 7B and 16A.
10904 nil))
10906 'defun-block-intro
10915 paren-state)
10921 ;; CASE 17F: first statement in an inline, or first
10922 ;; statement in a top-level defun. we can tell this is it
10923 ;; if there are no enclosing braces that haven't been
10924 ;; narrowed out by a class (i.e. don't use bod here).
10927 paren-state)))
10933 nil))))
10938 ;; CASE 17G: First statement in a function declared inside
10939 ;; a normal block. This can occur in Pike and with
10940 ;; e.g. the gcc extensions, but watch out for macros
10941 ;; followed by blocks. C.f. cases B.3 and 16F.
10947 ;; Turn off recognition of constructs that lacks
10948 ;; a type in this case, since that's more likely
10949 ;; to be a macro followed by a block.
10955 lim paren-state))
10957 ;; CASE 17H: First statement in a block.
10959 ;; If the block is preceded by a case/switch label on the
10960 ;; same line, we anchor at the first preceding label at
10961 ;; boi. The default handling in c-add-stmt-syntax is
10962 ;; really fixes it better, but we do like this to keep the
10963 ;; indentation compatible with version 5.28 and earlier.
10964 ;; C.f. case 16C.
10971 ;; c-backward-to-block-anchor not necessary here; those
10972 ;; situations are handled in case 17I above.
10974 lim paren-state))
10977 ))
10978 )
10980 ;; now we need to look at any modifiers
10984 ;; are we looking at a comment only line?
10989 ;; we might want to give additional offset to friends (in C++).
10994 ;; Set syntactic-relpos.
11000 nil)
11001 t))
11004 ;; Start of or a continuation of a preprocessor directive?
11019 ;; If inside a cpp expression, i.e. anywhere in a
11020 ;; cpp directive except a #define body, we only let
11021 ;; through the syntactic analysis that is internal
11022 ;; in the expression. That means the arglist
11023 ;; elements, if they are anchored inside the cpp
11024 ;; expression.
11033 ;; Inside a #define body and the syntactic analysis is
11034 ;; anchored on the start of the #define. In this case
11035 ;; we add cpp-define-intro to get the extra
11036 ;; indentation of the #define body.
11039 ;; return the syntax
11040 c-syntactic-context)))
11042 \f
11043 ;; Indentation calculation.
11046 ;; offset can be a number, a function, a variable, a list, or one of
11047 ;; the symbols + or -
11048 ;;
11049 ;; This function might do hidden buffer changes.
11068 langelem symbol))
11074 offset symbol)
11075 nil)
11089 Error evaluating offset %S for %s: \
11090 Cannot combine absolute offset %S with relative %S in `%s' method"
11096 Error evaluating offset %S for %s: \
11097 Cannot combine relative offset %S with absolute %S in `%s' method"
11102 res))
11120 Error evaluating offset %S for %s: \
11121 Cannot combine absolute offsets %S and %S in `add' method"
11125 res))
11134 res))))
11141 nil))))
11145 res
11147 offset symbol res)
11148 nil)))
11151 ;; Get offset from LANGELEM which is a list beginning with the
11152 ;; syntactic symbol and followed by any analysis data it provides.
11153 ;; That data may be zero or more elements, but if at least one is
11154 ;; given then the first is the anchor position (or nil). The symbol
11155 ;; is matched against `c-offsets-alist' and the offset calculated
11156 ;; from that is returned.
11157 ;;
11158 ;; This function might do hidden buffer changes.
11168 offset
11172 ))
11175 ;; This is a compatibility wrapper for `c-calc-offset' in case
11176 ;; someone is calling it directly. It takes an old style syntactic
11177 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
11178 ;; new list form.
11179 ;;
11180 ;; This function might do hidden buffer changes.
11187 ;; Calculate the syntactic indentation from a syntactic description
11188 ;; as returned by `c-guess-syntax'.
11189 ;;
11190 ;; Note that topmost-intro always has an anchor position at bol, for
11191 ;; historical reasons. It's often used together with other symbols
11192 ;; that has more sane positions. Since we always use the first
11193 ;; found anchor position, we rely on that these other symbols always
11194 ;; precede topmost-intro in the LANGELEMS list.
11195 ;;
11196 ;; This function might do hidden buffer changes.
11204 ;; Got an absolute column that overrides any indentation
11205 ;; we've collected so far, but not the relative
11206 ;; indentation we might get for the nested structures
11207 ;; further down the langelems list.
11211 ;; Got a relative change of the current calculated
11212 ;; indentation.
11215 ;; Use the anchor position from the first syntactic
11216 ;; element with one.
11226 indent)))
11228 \f
11231 ;;; Local Variables:
11232 ;;; indent-tabs-mode: t
11233 ;;; tab-width: 8
11234 ;;; End:
11235 ;;; cc-engine.el ends here