| blob | e596f0713b4d6eadd0a4ade716b1063a91970dfa |
1 ;;; cc-engine.el --- core syntax guessing engine for CC mode -*- coding: utf-8 -*-
3 ;; Copyright (C) 1985, 1987, 1992-2014 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
2254 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2255 ;; string, comment, or macro.
2256 ;;
2257 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2258 ;; MAY NOT contain any positions within macros, since macros are frequently
2259 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2260 ;; We cannot rely on this mechanism whilst determining a cache pos since
2261 ;; this function is also called from outwith `c-parse-state'.
2266 pos npos high-pos lit macro-beg macro-end)
2267 ;; Trim the cache to take account of buffer changes.
2279 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2283 ;; Test for being in a literal. If so, go to after it.
2290 ;; Test for being in a macro. If so, go to after it.
2305 ;; Add one extra element above HERE so as to to avoid the previous
2306 ;; expensive calculation when the next call is close to the current
2307 ;; one. This is especially useful when inside a large macro.
2312 pos))))
2315 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2316 ;; string or comment. It may be in a macro.
2321 pos npos high-pos lit macro-beg macro-end)
2322 ;; Trim the cache to take account of buffer changes.
2334 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2338 ;; Test for being in a literal. If so, go to after it.
2351 pos))))
2354 ;; If position HERE is inside a literal, return (START . END), the
2355 ;; boundaries of the literal (which may be outside the accessible bit of the
2356 ;; buffer). Otherwise, return nil.
2357 ;;
2358 ;; This function is almost the same as `c-literal-limits'. Previously, it
2359 ;; differed in that it was a lower level function, and that it rigorously
2360 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2361 ;; virtually identical to this function.
2369 ;; Return the start of the literal containing POS, or POS itself.
2371 pos))
2374 ;; Return a position outside of a string/comment/macro at or before POS.
2375 ;; STATE is the parse-partial-sexp state at POS.
2379 pos)))
2384 res))))
2386 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2387 ;; Stuff to do with point-min, and coping with any literal there.
2390 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2391 ;; narrowing is likely to affect the parens that are visible before the point.
2397 ;; These two variables define the literal, if any, containing point-min.
2398 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2399 ;; literal. If there's no literal there, they're both nil.
2403 ;; This is the earliest buffer-pos from which scanning can be done. It is
2404 ;; either the end of the literal containing point-min, or point-min itself.
2405 ;; It becomes nil if the buffer is changed earlier than this point.
2407 ;; Return the lowest valid scanning pos. This will be the end of the
2408 ;; literal enclosing point-min, or point-min itself.
2420 ;; Determine the properties of any literal containing POINT-MIN, setting the
2421 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2422 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2424 lit)
2439 c-state-point-min-lit-start nil
2440 c-state-min-scan-pos p-min)))))
2443 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2444 ;; A variable which signals a brace dessert - helpful for reducing the number
2445 ;; of fruitless backward scans.
2448 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2449 ;; that defun has searched backwards for a brace pair and not found one. Its
2450 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2451 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2452 ;; nil when at top level) and FROM is where the backward search started. It
2453 ;; is reset to nil in `c-invalidate-state-cache'.
2456 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2457 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2458 ;; list of like structure.
2460 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2461 ;; (default `c-state-cache') (or nil).
2468 ;; Return the address of the latest brace/bracket/paren (whether left or
2469 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2476 ;; Return the position just after the latest brace/bracket/paren (whether
2477 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2485 ;; From the state-cache, determine the buffer position from which we might
2486 ;; scan forward to HERE to update this cache. This position will be just
2487 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2488 ;; return the earliest position in the accessible region which isn't within
2489 ;; a literal. If the visible portion of the buffer is entirely within a
2490 ;; literal, return NIL.
2492 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2508 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2509 ;; Variables which keep track of preprocessor constructs.
2518 ;; These are the limits of the macro containing point at the previous call of
2519 ;; `c-parse-state', or nil.
2521 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2522 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2524 ;; Return the position of the opening paren/brace/bracket before HERE- which
2525 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2526 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2527 ;;
2528 ;; ............................................
2529 ;; | |
2530 ;; ( [ ( .........#macro.. ) ( ) ] )
2531 ;; ^ ^ ^ ^
2532 ;; | | | |
2533 ;; return HERE- HERE+ TOP
2534 ;;
2535 ;; If there aren't enough opening paren/brace/brackets, return the position
2536 ;; of the outermost one found, or HERE- if there are none. If there are no
2537 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2538 ;; must not be inside literals. Only the accessible portion of the buffer
2539 ;; will be scanned.
2541 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2542 ;; `here'. Go round the next loop each time we pass over such a ")". These
2543 ;; probably match "("s before `here-'.
2555 ;; PART 2: Scan back before `here-' searching for the "("s
2556 ;; matching/mismatching the ")"s found above. We only need to direct the
2557 ;; caller to scan when we've encountered unmatched right parens.
2566 pos))
2569 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2570 ;; to minimize the amount of scanning. HERE is the pertinent position in
2571 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2572 ;; its head trimmed) is known to be good, or nil if there is no such
2573 ;; position.
2574 ;;
2575 ;; The return value is a list, one of the following:
2576 ;;
2577 ;; o - ('forward START-POINT) - scan forward from START-POINT,
2578 ;; which is not less than the highest position in `c-state-cache' below HERE,
2579 ;; which is after GOOD-POS.
2580 ;; o - ('backward nil) - scan backwards (from HERE).
2581 ;; o - ('back-and-forward START-POINT) - like 'forward, but when HERE is earlier
2582 ;; than GOOD-POS.
2583 ;; o - ('IN-LIT nil) - point is inside the literal containing point-min.
2585 strategy ; 'forward, 'backward, or 'IN-LIT.
2586 start-point)
2598 start-point cache-pos)))
2602 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2603 ;; Routines which change `c-state-cache' and associated values.
2605 ;; The region (more precisely, point-min) has changed since we
2606 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2608 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2609 ;; It would be possible to do a better job here and recalculate the top
2610 ;; only.
2614 c-state-cache-good-pos c-state-min-scan-pos
2615 c-state-brace-pair-desert nil))
2617 ;; point-min has MOVED FORWARD.
2619 ;; Is the new point-min inside a (different) literal?
2624 ;; Cut off a bit of the tail from `c-state-cache'.
2626 pa)
2634 c-state-cache-good-pos c-state-min-scan-pos)
2637 )))
2642 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2643 ;; of nesting (not necessarily immediately preceding), push a cons onto
2644 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2645 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2646 ;; UPPER-LIM.
2647 ;;
2648 ;; Return non-nil when this has been done.
2649 ;;
2650 ;; The situation it copes with is this transformation:
2651 ;;
2652 ;; OLD: { (.) {...........}
2653 ;; ^ ^
2654 ;; FROM HERE
2655 ;;
2656 ;; NEW: { {....} (.) {.........
2657 ;; ^ ^ ^
2658 ;; LOWER BRACE PAIR HERE or HERE
2659 ;;
2660 ;; This routine should be fast. Since it can get called a LOT, we maintain
2661 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2662 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2672 ;; Don't start scanning in the middle of a CPP construct unless
2673 ;; it contains HERE - these constructs, in Emacs, are "commented
2674 ;; out" with category properties.
2677 macro-start-or-from
2678 from))
2682 ;; If we're essentially repeating a fruitless search, just give up.
2688 ;; DESERT-LIM. Avoid repeated searching through the cached desert.
2694 ;; CACHE-LIM. This limit will be necessary when an opening
2695 ;; paren at `cache-pos' has just had its matching close paren
2696 ;; inserted into the buffer. `cache-pos' continues to be a
2697 ;; search bound, even though the algorithm below would skip
2698 ;; over the new paren pair.
2707 ;; The top limit is EOB to ensure that `bra' is inside the
2708 ;; accessible part of the buffer at the next scan operation.
2711 ;; In the next pair of nested loops, the inner one moves back past a
2712 ;; pair of (mis-)matching parens or brackets; the outer one moves
2713 ;; back over a sequence of unmatched close brace/paren/bracket each
2714 ;; time round.
2730 ; brace, ignore it.
2733 ;; We've found the desired brace-pair.
2745 ;; We haven't found a brace pair. Record this in the cache.
2748 bra
2753 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2754 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2755 ;; "push" "a" brace pair onto `c-state-cache'.
2756 ;;
2757 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2758 ;; otherwise push it normally.
2759 ;;
2760 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2761 ;; latter is inside a macro, not being a macro containing
2762 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2763 ;; base pair. This latter case is assumed to be rare.
2764 ;;
2765 ;; Note: POINT is not preserved in this routine.
2775 c-state-cache)))
2776 ;; N.B. This defsubst codes one method for the simple, normal case,
2777 ;; and a more sophisticated, slower way for the general case. Don't
2778 ;; eliminate this defsubst - it's a speed optimization.
2782 ;; Scan the buffer from FROM to HERE, adding elements into `c-state-cache'
2783 ;; for braces etc. Return a candidate for `c-state-cache-good-pos'.
2784 ;;
2785 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2786 ;; any. Typically, it is immediately after it. It must not be inside a
2787 ;; literal.
2793 here)))
2796 ; Is actually the pos. to scan for a (/{/[ from,
2797 ; which sometimes is after a silly )/}/].
2800 ; good-pos.
2808 ;; Each time round the following loop, we enter a successively deeper
2809 ;; level of brace/paren nesting. (Except sometimes we "continue at
2810 ;; the existing level".) `pa+1' is a pos inside an opening
2811 ;; brace/paren/bracket, usually just after it.
2814 ;; Each time round the next loop moves forward over an opening then
2815 ;; a closing brace/bracket/paren. This loop is white hot, so it
2816 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2817 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2818 ;; call of `scan-lists' signals an error, which happens when there
2819 ;; are no more b/b/p's to scan.
2830 ;; We've just entered a deeper nesting level.
2832 ;; Insert the brace pair (if present) and the single open
2833 ;; paren/brace/bracket into `c-state-cache' It cannot be
2834 ;; inside a macro, except one around point, because of what
2835 ;; `c-neutralize-syntax-in-CPP' has done.
2837 ;; Insert the opening brace/bracket/paren position.
2839 ;; Clear admin stuff for the next more nested part of the scan.
2843 ;; All open p/b/b's at this nesting level, if any, have probably
2844 ;; been closed by matching/mismatching ones. We're probably
2845 ;; finished - we just need to check for having found an
2846 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2847 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2850 ;; Record the final, innermost, brace-pair if there is one.
2853 ;; Determine a good pos
2861 here-bol))
2870 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2871 ;; not be in it when it is amended for position HERE. This may involve
2872 ;; replacing a CONS element for a brace pair containing HERE with its car.
2873 ;; Additionally, the "outermost" open-brace entry before HERE will be
2874 ;; converted to a cons if the matching close-brace is below HERE.
2875 ;;
2876 ;; START-POINT is a "maximal" "safe position" - there must be no open
2877 ;; parens/braces/brackets between START-POINT and HERE.
2878 ;;
2879 ;; As a second thing, calculate the result of parse-partial-sexp at
2880 ;; PPS-POINT, w.r.t. START-POINT. The motivation here is that
2881 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2882 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2883 ;; needs to be FAST).
2884 ;;
2885 ;; Return a list (GOOD-POS SCAN-BACK-POS CONS-SEPARATED PPS-STATE), where
2886 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2887 ;; to be good (we aim for this to be as high as possible);
2888 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2889 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2890 ;; position to scan backwards from. It is the position of the "{" of the
2891 ;; last element to be removed from `c-state-cache', when that elt is a
2892 ;; cons, otherwise nil.
2893 ;; o - CONS-SEPARATED is t when a cons element in `c-state-cache' has been
2894 ;; replaced by its car because HERE lies inside the brace pair represented
2895 ;; by the cons.
2896 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2906 ; start-point or nil
2913 ; HERE), or nil.
2919 pps-state ; Will be 9 or 10 elements long.
2920 pos
2921 upper-lim ; ,beyond which `c-state-cache' entries are removed
2922 scan-back-pos
2923 cons-separated
2924 pair-beg pps-point-state target-depth)
2926 ;; Remove entries beyond HERE. Also remove any entries inside
2927 ;; a macro, unless HERE is in the same macro.
2932 here
2938 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2939 ;; RBrace and indicate we'll need to search backwards for a previous
2940 ;; brace pair.
2946 cons-separated t))
2948 ;; The next loop jumps forward out of a nested level of parens each
2949 ;; time round; the corresponding elements in `c-state-cache' are
2950 ;; removed. `pos' is just after the brace-pair or the open paren at
2951 ;; (car c-state-cache). There can be no open parens/braces/brackets
2952 ;; between `start-point'/`start-point-actual-macro-start' and HERE,
2953 ;; due to the interface spec to this function.
2956 in-macro-start)))
2958 ; marked by a `category' text property.
2959 start-point))
2970 ;; Do nothing when we've merely reached pps-point.
2971 )
2973 ;; Scan!
2977 target-depth
2978 nil pps-state))
2988 ;; We've just moved out of the paren pair containing the brace-pair
2989 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2990 ;; and is potentially where the open brace of a cons in
2991 ;; c-state-cache will be.
2997 ; containing Lparen
3001 ; together with enclosed brace pair.
3002 ;; (t nil) ; Ignore an unmated Rparen.
3003 )))
3007 nil nil ; TARGETDEPTH, STOPBEFORE
3008 pps-state)))
3010 ;; If the last paren pair we moved out of was actually a brace pair,
3011 ;; insert it into `c-state-cache'.
3016 c-state-cache)))
3021 ;; Strip stale elements of `c-state-cache' by moving backwards through the
3022 ;; buffer, and inform the caller of the scenario detected.
3023 ;;
3024 ;; HERE is the position we're setting `c-state-cache' for.
3025 ;; CACHE-POS (a locally bound variable) is just after the latest recorded
3026 ;; position in `c-state-cache' before HERE, or a position at or near
3027 ;; point-min which isn't in a literal.
3028 ;;
3029 ;; This function must only be called only when (> `c-state-cache-good-pos'
3030 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
3031 ;; optimized to eliminate (or minimize) scanning between these two
3032 ;; positions.
3033 ;;
3034 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
3035 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
3036 ;; could become so after missing elements are inserted into
3037 ;; `c-state-cache'. This is JUST AFTER an opening or closing
3038 ;; brace/paren/bracket which is already in `c-state-cache' or just before
3039 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
3040 ;; before `here''s line, or the start of the literal containing it.
3041 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
3042 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
3043 ;; to scan backwards from.
3044 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
3045 ;; POS and HERE which aren't recorded in `c-state-cache'.
3046 ;;
3047 ;; The comments in this defun use "paren" to mean parenthesis or square
3048 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
3049 ;;
3050 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
3051 ;; | | | | | |
3052 ;; CP E here D C good
3055 pa ren ; positions of "(" and ")"
3056 dropped-cons ; whether the last element dropped from `c-state-cache'
3057 ; was a cons (representing a brace-pair)
3058 good-pos ; see above.
3059 lit ; (START . END) of a literal containing some point.
3060 here-lit-start here-lit-end ; bounds of literal containing `here'
3061 ; or `here' itself.
3066 ;; Remove completely irrelevant entries from `c-state-cache'.
3072 ;; At this stage, (> pos here);
3073 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3078 ;; CASE 1: The top of the cache is a brace pair which now encloses
3079 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3080 ;; knowledge of what's inside these braces, we have no alternative but
3081 ;; to direct the caller to scan the buffer from the opening brace.
3085 ; entry into a { entry, so the caller needs to
3086 ; search for a brace pair before the {.
3088 ;; `here' might be inside a literal. Check for this.
3093 ;; Has `here' just "newly entered" a macro?
3106 ;; `here' might be nested inside any depth of parens (or brackets but
3107 ;; not braces). Scan backwards to find the outermost such opening
3108 ;; paren, if there is one. This will be the scan position to return.
3115 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3116 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3117 ;; a brace pair preceding this, it will already be in `c-state-cache',
3118 ;; unless there was a brace pair after it, i.e. there'll only be one to
3119 ;; scan for if we've just deleted one.
3122 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3123 ;; Further forward scanning isn't needed, but we still need to find a
3124 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3134 ;; CASE 3: After a }/)/] before `here''s BOL.
3138 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3139 ;; literal containing it.
3144 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3145 ;; Externally visible routines.
3149 c-state-cache-good-pos 1
3150 c-state-nonlit-pos-cache nil
3151 c-state-nonlit-pos-cache-limit 1
3152 c-state-semi-nonlit-pos-cache nil
3153 c-state-semi-nonlit-pos-cache-limit 1
3154 c-state-brace-pair-desert nil
3155 c-state-point-min 1
3156 c-state-point-min-lit-type nil
3157 c-state-point-min-lit-start nil
3158 c-state-min-scan-pos 1
3159 c-state-old-cpp-beg nil
3160 c-state-old-cpp-end nil)
3163 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3164 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3165 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3166 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3167 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3168 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3169 ;; (defun c-state-dump ()
3170 ;; ;; For debugging.
3171 ;; ;(message
3172 ;; (concat
3173 ;; (c-sc-qde c-state-cache)
3174 ;; (c-sc-de c-state-cache-good-pos)
3175 ;; (c-sc-qde c-state-nonlit-pos-cache)
3176 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3177 ;; (c-sc-qde c-state-brace-pair-desert)
3178 ;; (c-sc-de c-state-point-min)
3179 ;; (c-sc-de c-state-point-min-lit-type)
3180 ;; (c-sc-de c-state-point-min-lit-start)
3181 ;; (c-sc-de c-state-min-scan-pos)
3182 ;; (c-sc-de c-state-old-cpp-beg)
3183 ;; (c-sc-de c-state-old-cpp-end)))
3184 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3187 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3188 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3189 ;; left in a consistent state.
3190 ;;
3191 ;; This is much like `c-whack-state-after', but it never changes a paren
3192 ;; pair element into an open paren element. Doing that would mean that the
3193 ;; new open paren wouldn't have the required preceding paren pair element.
3194 ;;
3195 ;; This function is called from c-after-change.
3197 ;; The caches of non-literals:
3198 ;; Note that we use "<=" for the possibility of the second char of a two-char
3199 ;; comment opener being typed; this would invalidate any cache position at
3200 ;; HERE.
3206 ;; `c-state-cache':
3207 ;; Case 1: if `here' is in a literal containing point-min, everything
3208 ;; becomes (or is already) nil.
3212 c-state-cache-good-pos nil
3213 c-state-min-scan-pos nil)
3215 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3216 ;; below `here'. To maintain its consistency, we may need to insert a new
3217 ;; brace pair.
3220 too-high-pa ; recorded {/(/[ next above here, or nil.
3221 dropped-cons ; was the last removed element a brace pair?
3222 pa)
3223 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3230 ;; Do we need to add in an earlier brace pair, having lopped one off?
3237 ;; The brace-pair desert marker:
3245 ;; Find and record all noteworthy parens between some good point earlier in
3246 ;; the file and point. That good point is at least the beginning of the
3247 ;; top-level construct we are in, or the beginning of the preceding
3248 ;; top-level construct if we aren't in one.
3249 ;;
3250 ;; The returned value is a list of the noteworthy parens with the last one
3251 ;; first. If an element in the list is an integer, it's the position of an
3252 ;; open paren (of any type) which has not been closed before the point. If
3253 ;; an element is a cons, it gives the position of a closed BRACE paren
3254 ;; pair[*]; the car is the start brace position and the cdr is the position
3255 ;; following the closing brace. Only the last closed brace paren pair
3256 ;; before each open paren and before the point is recorded, and thus the
3257 ;; state never contains two cons elements in succession. When a close brace
3258 ;; has no matching open brace (e.g., the matching brace is outside the
3259 ;; visible region), it is not represented in the returned value.
3260 ;;
3261 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3262 ;; This defun explicitly treats mismatching parens/braces/brackets as
3263 ;; matching. It is the open brace which makes it a "brace" pair.
3264 ;;
3265 ;; If POINT is within a macro, open parens and brace pairs within
3266 ;; THIS macro MIGHT be recorded. This depends on whether their
3267 ;; syntactic properties have been suppressed by
3268 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3269 ;;
3270 ;; Currently no characters which are given paren syntax with the
3271 ;; syntax-table property are recorded, i.e. angle bracket arglist
3272 ;; parens are never present here. Note that this might change.
3273 ;;
3274 ;; BUG: This function doesn't cope entirely well with unbalanced
3275 ;; parens in macros. (2008-12-11: this has probably been resolved
3276 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3277 ;; following case the brace before the macro isn't balanced with the
3278 ;; one after it:
3279 ;;
3280 ;; {
3281 ;; #define X {
3282 ;; }
3283 ;;
3284 ;; Note to maintainers: this function DOES get called with point
3285 ;; within comments and strings, so don't assume it doesn't!
3286 ;;
3287 ;; This function might do hidden buffer changes.
3290 open-paren-in-column-0-is-defun-start
3291 strategy ; 'forward, 'backward etc..
3292 ;; Candidate positions to start scanning from:
3293 cache-pos ; highest position below HERE already existing in
3294 ; cache (or 1).
3295 good-pos
3296 start-point ; (when scanning forward) a place below HERE where there
3297 ; are no open parens/braces between it and HERE.
3298 bopl-state
3299 res
3300 cons-separated
3302 ;; If POINT-MIN has changed, adjust the cache
3306 ;; Strategy?
3311 ;; SCAN!
3319 ; start-point)
3329 good-pos)))
3341 good-pos)))
3345 c-state-cache-good-pos nil))))
3347 c-state-cache)
3350 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3351 ;;
3352 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3353 ;; of all parens in preprocessor constructs, except for any such construct
3354 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3355 ;; worrying further about macros and template delimiters.
3360 c-state-old-cpp-beg
3367 ;; If PLACE is non-nil, return a marker marking it, otherwise nil.
3368 ;; We (re)use MARKER.
3374 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3375 ;; description of the functionality and return value.
3376 ;;
3377 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3378 ;; of all parens in preprocessor constructs, except for any such construct
3379 ;; containing point. We can then call `c-parse-state-1' without worrying
3380 ;; further about macros and template delimiters.
3387 here-cpp-beg)
3389 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3390 ;; subsystem.
3395 here-cpp-beg here-cpp-end
3401 c-state-old-cpp-end
3404 ;; Debug tool to catch cache inconsistencies. This is called from
3405 ;; 000tests.el.
3423 val))))
3425 c-state-cache-good-pos
3426 c-state-nonlit-pos-cache
3427 c-state-nonlit-pos-cache-limit
3428 c-state-semi-nonlit-pos-cache
3429 c-state-semi-nonlit-pos-cache-limit
3430 c-state-brace-pair-desert
3431 c-state-point-min
3432 c-state-point-min-lit-type
3433 c-state-point-min-lit-start
3434 c-state-min-scan-pos
3435 c-state-old-cpp-beg
3436 c-state-old-cpp-end
3437 c-parse-state-point))))
3455 conses-not-ok))
3472 ;; The cache can actually go further back due to the ad-hoc way
3473 ;; the first paren is found, so try to whack off a bit of its
3474 ;; start before complaining.
3475 ;; (save-excursion
3476 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3477 ;; (c-beginning-of-defun-1)
3478 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3479 ;; (c-beginning-of-defun-1))
3480 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3481 ;; (message (concat "c-parse-state inconsistency at %s: "
3482 ;; "using cache: %s, from scratch: %s")
3483 ;; here res1 res2)))
3486 here res1 res2)
3492 here res1)
3497 res2 ; res1 correct a cascading series of errors ASAP
3498 ))
3504 'c-debug-parse-state
3512 \f
3514 ;; Whack off any state information from PAREN-STATE which lies
3515 ;; before BUFPOS. Not destructive on PAREN-STATE.
3518 car)
3529 ;; Whack off any state information from PAREN-STATE which lies at or
3530 ;; after BUFPOS. Not destructive on PAREN-STATE.
3535 ;; just check the car, because in a balanced brace
3536 ;; expression, it must be impossible for the corresponding
3537 ;; close brace to be before point, but the open brace to
3538 ;; be after.
3540 nil ; whack it off
3542 ;; its possible that the open brace is before
3543 ;; bufpos, but the close brace is after. In that
3544 ;; case, convert this to a non-cons element. The
3545 ;; rest of the state is before bufpos, so we're
3546 ;; done.
3548 ;; we know that both the open and close braces are
3549 ;; before bufpos, so we also know that everything else
3550 ;; on state is before bufpos.
3553 nil ; whack it off
3554 ;; it's before bufpos, so everything else should too.
3557 nil)))
3560 ;; Return the bufpos of the innermost enclosing open paren before
3561 ;; bufpos, or nil if none was found.
3571 enclosingp))
3574 ;; Return the bufpos of the outermost enclosing open paren, or nil
3575 ;; if none was found.
3582 pos))
3585 ;; Return the closest "safe" position recorded on PAREN-STATE that
3586 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3587 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3588 ;; find the closest limit before a given limit that might be nil.
3589 ;;
3590 ;; A "safe" position is a position at or after a recorded open
3591 ;; paren, or after a recorded close paren. The returned position is
3592 ;; thus either the first position after a close brace, or the first
3593 ;; position after an enclosing paren, or at the enclosing paren in
3594 ;; case BUFPOS is immediately after it.
3604 ;; See below.
3607 ;; elem is the position at and not after the opening paren, so
3608 ;; we can go forward one more step unless it's equal to
3609 ;; bufpos. This is useful in some cases avoid an extra paren
3610 ;; level between the safe position and bufpos.
3615 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3616 ;; goes to the closest previous point that is known to be outside
3617 ;; any string literal or comment. `c-state-cache' is used if it has
3618 ;; a position in the vicinity.
3620 elem
3623 ;; Note: Similar code in `c-safe-position'. The
3624 ;; difference is that we accept a safe position at
3625 ;; the point and don't bother to go forward past open
3626 ;; parens.
3641 ;; The position is far back. Try `c-beginning-of-defun-1'
3642 ;; (although we can't be entirely sure it will go to a position
3643 ;; outside a comment or string in current emacsen). FIXME:
3644 ;; Consult `syntax-ppss' here.
3649 \f
3650 ;; Tools for scanning identifiers and other tokens.
3653 "Return non-nil if the point is on or directly after an identifier.
3654 Keywords are recognized and not considered identifiers. If an
3655 identifier is detected, the returned value is its starting position.
3656 If an identifier ends at the point and another begins at it \(can only
3657 happen in Pike) then the point for the preceding one is returned.
3659 Note that this function might do hidden buffer changes. See the
3660 comment at the start of cc-engine.el for more info."
3662 ;; FIXME: Shouldn't this function handle "operator" in C++?
3669 ;; Check for a normal (non-keyword) identifier.
3675 ;; Handle the `<operator> syntax in Pike.
3679 t
3686 ;; Handle the "operator +" syntax in C++.
3703 )))
3706 ;; If the point is at the end of a symbol then skip backward to the
3707 ;; beginning of it. Don't move otherwise. Return non-nil if point
3708 ;; moved.
3709 ;;
3710 ;; This function might do hidden buffer changes.
3713 ;; Handle the `<operator> syntax in Pike.
3719 t
3721 nil)))))
3724 ;; Move to the beginning of the current token. Do not move if not
3725 ;; in the middle of one. BACK-LIMIT may be used to bound the
3726 ;; backward search; if given it's assumed to be at the boundary
3727 ;; between two tokens. Return non-nil if the point is moved, nil
3728 ;; otherwise.
3729 ;;
3730 ;; This function might do hidden buffer changes.
3737 ;; `c-nonsymbol-token-regexp' should always match
3738 ;; since we've skipped backward over punctuation
3739 ;; or paren syntax, but consume one char in case
3740 ;; it doesn't so that we don't leave point before
3741 ;; some earlier incorrect token.
3748 ;; Move to the end of the current token. Do not move if not in the
3749 ;; middle of one. BACK-LIMIT may be used to bound the backward
3750 ;; search; if given it's assumed to be at the boundary between two
3751 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3752 ;;
3753 ;; This function might do hidden buffer changes.
3761 ;; `c-nonsymbol-token-regexp' should always match since
3762 ;; we've skipped backward over punctuation or paren
3763 ;; syntax, but move forward in case it doesn't so that
3764 ;; we don't leave point earlier than we started with.
3780 "Move forward by tokens.
3781 A token is defined as all symbols and identifiers which aren't
3783 treated properly). Point is always either left at the beginning of a
3784 token or not moved at all. COUNT specifies the number of tokens to
3785 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3786 moves to the next token beginning only if not already at one. If
3787 BALANCED is true, move over balanced parens, otherwise move into them.
3788 Also, if BALANCED is true, never move out of an enclosing paren.
3790 LIMIT sets the limit for the movement and defaults to the point limit.
3791 The case when LIMIT is set in the middle of a token, comment or macro
3792 is handled correctly, i.e. the point won't be left there.
3794 Return the number of tokens left to move \(positive or negative). If
3795 BALANCED is true, a move over a balanced paren counts as one. Note
3796 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3797 be returned. Thus, a return value of 0 guarantees that point is at
3798 the requested position and a return value less \(without signs) than
3799 COUNT guarantees that point is at the beginning of some token.
3801 Note that this function might do hidden buffer changes. See the
3802 comment at the start of cc-engine.el for more info."
3809 c-jump-syntax-balanced
3810 c-jump-syntax-unbalanced))
3815 ;; If count is zero we should jump if in the middle of a token.
3822 ;; Skip whitespace. Count this as a move if we did in
3823 ;; fact move.
3827 ;; Moved out of bounds. Make sure the returned count isn't zero.
3832 ;; Use `condition-case' to avoid having the limit tests
3833 ;; inside the loop.
3841 t)
3844 t)
3845 ;; `c-nonsymbol-token-regexp' above should always
3846 ;; match if there are correct tokens. Try to
3847 ;; widen to see if the limit was set in the
3848 ;; middle of one, else fall back to treating
3849 ;; the offending thing as a one character token.
3854 nil)
3857 t))))
3867 count)))
3870 "Move backward by tokens.
3871 See `c-forward-token-2' for details."
3879 c-jump-syntax-balanced
3880 c-jump-syntax-unbalanced))
3884 ;; The count is zero so try to skip to the beginning of the
3885 ;; current token.
3889 ;; The limit is inside the same token, so return 1.
3892 ;; We're not in the middle of a token. If there's
3893 ;; whitespace after the point then we must move backward,
3894 ;; so set count to 1 in that case.
3896 ;; If we're looking at a '#' that might start a cpp
3897 ;; directive then we have to do a more elaborate check.
3910 ;; Use `condition-case' to avoid having to check for buffer
3911 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3920 ;; This can be very inefficient if there's a long
3921 ;; sequence of operator tokens without any separation.
3922 ;; That doesn't happen in practice, anyway.
3932 count)))
3935 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3937 characters are jumped over character by character. This function is
3938 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3943 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3945 characters are jumped over character by character. This function is
3946 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3950 \f
3951 ;; Tools for doing searches restricted to syntactically relevant text.
3954 paren-level not-inside-token
3955 lookbehind-submatch)
3956 "Like `re-search-forward', but only report matches that are found
3957 in syntactically significant text. I.e. matches in comments, macros
3958 or string literals are ignored. The start point is assumed to be
3959 outside any comment, macro or string literal, or else the content of
3960 that region is taken as syntactically significant text.
3962 If PAREN-LEVEL is non-nil, an additional restriction is added to
3963 ignore matches in nested paren sexps. The search will also not go
3964 outside the current list sexp, which has the effect that if the point
3965 should be moved to BOUND when no match is found \(i.e. NOERROR is
3966 neither nil nor t), then it will be at the closing paren if the end of
3967 the current list sexp is encountered first.
3969 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3970 ignored. Things like multicharacter operators and special symbols
3972 constants.
3974 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3975 subexpression in REGEXP. The end of that submatch is used as the
3976 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
3977 isn't used or if that subexpression didn't match then the start
3979 subexpression is never tested before the starting position, so it
3983 subexpression can match the end of nonwhite syntactic whitespace,
3984 i.e. the end of comments or cpp directives. This since the function
3985 skips over such things before resuming the search. It's on the other
3987 matches syntactic whitespace.
3989 Bug: Unbalanced parens inside cpp directives are currently not handled
3990 correctly \(i.e. they don't get ignored as they should) when
3991 PAREN-LEVEL is set.
3993 Note that this function might do hidden buffer changes. See the
3994 comment at the start of cc-engine.el for more info."
3999 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
4002 tmp
4003 ;; Start position for the last search.
4004 search-pos
4005 ;; The `parse-partial-sexp' state between the start position
4006 ;; and the point.
4007 state
4008 ;; The current position after the last state update. The next
4009 ;; `parse-partial-sexp' continues from here.
4011 ;; The position at which to check the state and the state
4012 ;; there. This is separate from `state-pos' since we might
4013 ;; need to back up before doing the next search round.
4014 check-pos check-state
4015 ;; Last position known to end a token.
4017 ;; Set when a valid match is found.
4018 found)
4036 state-pos check-pos paren-level nil state))
4038 check-state state))
4040 ;; NOTE: If we got a look behind subexpression and get
4041 ;; an insignificant match in something that isn't
4042 ;; syntactic whitespace (i.e. strings or in nested
4043 ;; parentheses), then we can never skip more than a
4044 ;; single character from the match start position
4045 ;; (i.e. `state-pos' here) before continuing the
4046 ;; search. That since the look behind subexpression
4047 ;; might match the end of the insignificant region in
4048 ;; the next search.
4052 ;; Match inside a line comment. Skip to eol. Use
4053 ;; `re-search-forward' instead of `skip-chars-forward' to get
4054 ;; the right bound behavior.
4058 ;; Match inside a block comment. Skip to the '*/'.
4065 ;; Match in the middle of the opener of a block or line
4066 ;; comment.
4071 ;; The last `parse-partial-sexp' above might have
4072 ;; stopped short of the real check position if the end
4073 ;; of the current sexp was encountered in paren-level
4074 ;; mode. The checks above are always false in that
4075 ;; case, and since they can do better skipping in
4076 ;; lookbehind-submatch mode, we do them before
4077 ;; checking the paren level.
4081 ;; Check the paren level first since we're short of the
4082 ;; syntactic checking position if the end of the
4083 ;; current sexp was encountered by `parse-partial-sexp'.
4086 ;; Inside a nested paren sexp.
4088 ;; See the NOTE above.
4090 ;; Skip out of the paren quickly.
4094 ;; Have exited the current paren sexp.
4097 ;; The last `parse-partial-sexp' call above
4098 ;; has left us just after the closing paren
4099 ;; in this case, so we can modify the bound
4100 ;; to leave the point at the right position
4101 ;; upon return.
4103 nil)
4107 ;; Match inside a string.
4110 ;; See the NOTE above.
4112 ;; Skip to the end of the string before continuing.
4121 continue)))
4126 ;; Match inside a macro. Skip to the end of it.
4134 (< check-pos
4140 ;; Inside a token.
4142 ;; See the NOTE above.
4147 ;; A real match.
4149 nil)))
4151 ;; Should loop to search again, but take care to avoid
4152 ;; looping on the same spot.
4156 nil
4159 t))))
4165 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4172 ;; Search failed. Set point as appropriate.
4176 nil)))
4181 ;; Return the start of the literal point is in, or nil.
4182 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4183 ;; bound in the caller.
4185 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4186 ;; if it's outside comments and strings.
4189 ;; Pick a safe position as close to the point as possible.
4190 ;;
4191 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4192 ;; position.
4205 ;; Cache positions along the way to use if we have to back up more. We
4206 ;; cache every closing paren on the same level. If the paren cache is
4207 ;; relevant in this region then we're typically already on the same
4208 ;; level as the target position. Note that we might cache positions
4209 ;; after opening parens in case safe-pos is in a nested list. That's
4210 ;; both uncommon and harmless.
4218 ;; If the state contains the start of the containing sexp we cache that
4219 ;; position too, so that parse-partial-sexp in the next run has a bigger
4220 ;; chance of starting at the same level as the target position and thus
4221 ;; will get more good safe positions into the list.
4227 ;; Inside string or comment. Continue search at the
4228 ;; beginning of it.
4232 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4233 i.e. don't stop at positions inside syntactic whitespace or string
4234 literals. Preprocessor directives are also ignored, with the exception
4235 of the one that the point starts within, if any. If LIMIT is given,
4236 it's assumed to be at a syntactically relevant position.
4238 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4239 sexps, and the search will also not go outside the current paren sexp.
4240 However, if LIMIT or the buffer limit is reached inside a nested paren
4241 then the point will be left at the limit.
4243 Non-nil is returned if the point moved, nil otherwise.
4245 Note that this function might do hidden buffer changes. See the
4246 comment at the start of cc-engine.el for more info."
4249 state-2
4250 ;; A list of syntactically relevant positions in descending
4251 ;; order. It's used to avoid scanning repeatedly over
4252 ;; potentially large regions with `parse-partial-sexp' to verify
4253 ;; each position. Used in `c-ssb-lit-begin'
4254 safe-pos-list
4255 ;; The result from `c-beginning-of-macro' at the start position or the
4256 ;; start position itself if it isn't within a macro. Evaluated on
4257 ;; demand.
4258 start-macro-beg
4259 ;; The earliest position after the current one with the same paren
4260 ;; level. Used only when `paren-level' is set.
4261 lit-beg
4266 ;; The next loop "tries" to find the end point each time round,
4267 ;; loops when it hasn't succeeded.
4273 ;; Don't stop inside a literal.
4276 t)))
4288 ;; Not at the right level.
4291 ;; We stop above if we go out of a paren.
4292 ;; Now check whether it precedes or is
4293 ;; nested in the starting sexp.
4297 pps-end-pos paren-level-pos
4298 nil nil state-2))
4301 ;; We've stopped short of the starting position
4302 ;; so the hit was inside a nested list. Go up
4303 ;; until we are at the right level.
4312 nil)
4313 t))
4316 nil))
4318 ;; The hit was outside the list at the start
4319 ;; position. Go to the start of the list and exit.
4321 nil))
4324 ;; Inside a macro.
4332 t
4334 ;; It's inside the same macro we started in so it's
4335 ;; a relevant match.
4337 nil))))))
4341 ;; Skip syntactic ws afterwards so that we don't stop at the
4342 ;; end of a comment if `skip-chars' is something like "^/".
4346 ;; We might want to extend this with more useful return values in
4347 ;; the future.
4350 ;; The following is an alternative implementation of
4351 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4352 ;; track of the syntactic context. It turned out to be generally
4353 ;; slower than the one above which uses forward checks from earlier
4354 ;; safe positions.
4355 ;;
4356 ;;(defconst c-ssb-stop-re
4357 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4358 ;; ;; stop at to avoid going into comments and literals.
4359 ;; (concat
4360 ;; ;; Match comment end syntax and string literal syntax. Also match
4361 ;; ;; '/' for block comment endings (not covered by comment end
4362 ;; ;; syntax).
4363 ;; "\\s>\\|/\\|\\s\""
4364 ;; (if (memq 'gen-string-delim c-emacs-features)
4365 ;; "\\|\\s|"
4366 ;; "")
4367 ;; (if (memq 'gen-comment-delim c-emacs-features)
4368 ;; "\\|\\s!"
4369 ;; "")))
4370 ;;
4371 ;;(defconst c-ssb-stop-paren-re
4372 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4373 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4374 ;;
4375 ;;(defconst c-ssb-sexp-end-re
4376 ;; ;; Regexp matching the ending syntax of a complex sexp.
4377 ;; (concat c-string-limit-regexp "\\|\\s)"))
4378 ;;
4379 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4380 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4381 ;;i.e. don't stop at positions inside syntactic whitespace or string
4382 ;;literals. Preprocessor directives are also ignored. However, if the
4383 ;;point is within a comment, string literal or preprocessor directory to
4384 ;;begin with, its contents is treated as syntactically relevant chars.
4385 ;;If LIMIT is given, it limits the backward search and the point will be
4386 ;;left there if no earlier position is found.
4387 ;;
4388 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4389 ;;sexps, and the search will also not go outside the current paren sexp.
4390 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4391 ;;then the point will be left at the limit.
4392 ;;
4393 ;;Non-nil is returned if the point moved, nil otherwise.
4394 ;;
4395 ;;Note that this function might do hidden buffer changes. See the
4396 ;;comment at the start of cc-engine.el for more info."
4397 ;;
4398 ;; (save-restriction
4399 ;; (when limit
4400 ;; (narrow-to-region limit (point-max)))
4401 ;;
4402 ;; (let ((start (point)))
4403 ;; (catch 'done
4404 ;; (while (let ((last-pos (point))
4405 ;; (stop-pos (progn
4406 ;; (skip-chars-backward skip-chars)
4407 ;; (point))))
4408 ;;
4409 ;; ;; Skip back over the same region as
4410 ;; ;; `skip-chars-backward' above, but keep to
4411 ;; ;; syntactically relevant positions.
4412 ;; (goto-char last-pos)
4413 ;; (while (and
4414 ;; ;; `re-search-backward' with a single char regexp
4415 ;; ;; should be fast.
4416 ;; (re-search-backward
4417 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4418 ;; stop-pos 'move)
4419 ;;
4420 ;; (progn
4421 ;; (cond
4422 ;; ((looking-at "\\s(")
4423 ;; ;; `paren-level' is set and we've found the
4424 ;; ;; start of the containing paren.
4425 ;; (forward-char)
4426 ;; (throw 'done t))
4427 ;;
4428 ;; ((looking-at c-ssb-sexp-end-re)
4429 ;; ;; We're at the end of a string literal or paren
4430 ;; ;; sexp (if `paren-level' is set).
4431 ;; (forward-char)
4432 ;; (condition-case nil
4433 ;; (c-backward-sexp)
4434 ;; (error
4435 ;; (goto-char limit)
4436 ;; (throw 'done t))))
4437 ;;
4438 ;; (t
4439 ;; (forward-char)
4440 ;; ;; At the end of some syntactic ws or possibly
4441 ;; ;; after a plain '/' operator.
4442 ;; (let ((pos (point)))
4443 ;; (c-backward-syntactic-ws)
4444 ;; (if (= pos (point))
4445 ;; ;; Was a plain '/' operator. Go past it.
4446 ;; (backward-char)))))
4447 ;;
4448 ;; (> (point) stop-pos))))
4449 ;;
4450 ;; ;; Now the point is either at `stop-pos' or at some
4451 ;; ;; position further back if `stop-pos' was at a
4452 ;; ;; syntactically irrelevant place.
4453 ;;
4454 ;; ;; Skip additional syntactic ws so that we don't stop
4455 ;; ;; at the end of a comment if `skip-chars' is
4456 ;; ;; something like "^/".
4457 ;; (c-backward-syntactic-ws)
4458 ;;
4459 ;; (< (point) stop-pos))))
4460 ;;
4461 ;; ;; We might want to extend this with more useful return values
4462 ;; ;; in the future.
4463 ;; (/= (point) start))))
4465 \f
4466 ;; Tools for handling comments and string literals.
4469 "Return the type of literal point is in, if any.
4470 The return value is `c' if in a C-style comment, `c++' if in a C++
4471 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4472 is non-nil and in a preprocessor line, or nil if somewhere else.
4473 Optional LIM is used as the backward limit of the search. If omitted,
4474 or nil, `c-beginning-of-defun' is used.
4476 The last point calculated is cached if the cache is enabled, i.e. if
4477 `c-in-literal-cache' is bound to a two element vector.
4479 Note that this function might do hidden buffer changes. See the
4480 comment at the start of cc-engine.el for more info."
4491 "Return a cons of the beginning and end positions of the comment or
4492 string surrounding point (including both delimiters), or nil if point
4494 to start parsing from. If NEAR is non-nil, then the limits of any
4496 spaces and tabs between point and the literal. The search for such a
4497 literal is done first in forward direction. If NOT-IN-DELIMITER is
4498 non-nil, the case when point is inside a starting delimiter won't be
4499 recognized. This only has effect for comments which have starting
4500 delimiters with more than one character.
4502 Note that this function might do hidden buffer changes. See the
4503 comment at the start of cc-engine.el for more info."
4519 ;; Search forward for a literal.
4530 ;; Search backward.
4541 ;; Block comment. Due to the nature of line
4542 ;; comments, they will always be covered by the
4543 ;; normal case above.
4546 ;; If LIM is bogus, beg will be bogus.
4550 ))))
4552 ;; In case external callers use this; it did have a docstring.
4556 "If the argument is a cons of two buffer positions (such as returned by
4557 `c-literal-limits'), and that range contains a C++ style line comment,
4558 then an extended range is returned that contains all adjacent line
4559 comments (i.e. all comments that starts in the same column with no
4560 empty lines or non-whitespace characters between them). Otherwise the
4561 argument is returned.
4563 Note that this function might do hidden buffer changes. See the
4564 comment at the start of cc-engine.el for more info."
4575 ;; Got to take care in the backward direction to handle
4576 ;; comments which are preceded by code.
4590 range)
4594 "Convenience function that given the result of `c-literal-limits',
4595 returns nil or the type of literal that the range surrounds, one
4596 of the symbols 'c, 'c++ or 'string. It's much faster than using
4597 `c-in-literal' and is intended to be used when you need both the
4598 type of a literal and its limits.
4600 Note that this function might do hidden buffer changes. See the
4601 comment at the start of cc-engine.el for more info."
4612 range))
4615 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4616 ;; This doesn't preserve point.
4625 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4626 ;; (default point). This is done by going back further in the buffer then
4627 ;; searching forward for literals. The position found won't be in a
4628 ;; literal. We start searching for the sought position TRY-SIZE (default
4629 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4630 ;; :-)
4638 stack elt size
4641 ;; Move forward one literal each time round this loop.
4642 ;; Move forward to the start of a comment or string.
4644 pos
4645 start
4646 nil ; target-depth
4647 nil ; stop-before
4648 s ; state
4651 ;; Gather details of the non-literal-bit - starting pos and size.
4655 pos))
4659 ;; Move forward to the end of the comment/string.
4663 start
4664 nil ; target-depth
4665 nil ; stop-before
4666 s ; state
4670 ;; Now try and find enough non-literal characters recorded on the stack.
4671 ;; Go back one recorded literal each time round this loop.
4673 stack)
4678 ;; Have we found enough yet?
4688 ;; Return a buffer position about HOW-FAR non-literal characters forward
4689 ;; from START-POS (default point), which must not be inside a literal.
4696 ;; Scan over counted characters.
4698 pos
4700 nil ; target-depth
4701 nil ; stop-before
4702 s ; state
4706 ;; Scan over literal characters.
4709 pos
4711 nil ; target-depth
4712 nil ; stop-before
4713 s ; state
4718 \f
4719 ;; `c-find-decl-spots' and accompanying stuff.
4721 ;; Variables used in `c-find-decl-spots' to cache the search done for
4722 ;; the first declaration in the last call. When that function starts,
4723 ;; it needs to back up over syntactic whitespace to look at the last
4724 ;; token before the region being searched. That can sometimes cause
4725 ;; moves back and forth over a quite large region of comments and
4726 ;; macros, which would be repeated for each changed character when
4727 ;; we're called during fontification, since font-lock refontifies the
4728 ;; current line for each change. Thus it's worthwhile to cache the
4729 ;; first match.
4730 ;;
4731 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4732 ;; the syntactic whitespace less or equal to some start position.
4733 ;; There's no cached value if it's nil.
4734 ;;
4735 ;; `c-find-decl-match-pos' is the match position if
4736 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4737 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4748 ; (defface c-debug-decl-spot-face
4749 ; '((t (:background "Turquoise")))
4750 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4751 ; (defface c-debug-decl-sws-face
4752 ; '((t (:background "Khaki")))
4753 ; "Debug face to mark the syntactic whitespace between the declaration
4754 ; spots and the preceding token end.")
4770 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4771 ;; but it contains lots of free variables that refer to things
4772 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4773 ;; if there is a match, otherwise at `cfd-limit'.
4774 ;;
4775 ;; The macro moves point forward to the next putative start of a declaration
4776 ;; or cfd-limit. This decl start is the next token after a "declaration
4777 ;; prefix". The declaration prefix is the earlier of `cfd-prop-match' and
4778 ;; `cfd-re-match'. `cfd-match-pos' is set to the decl prefix.
4779 ;;
4780 ;; This macro might do hidden buffer changes.
4783 ;; Find the next property match position if we haven't got one already.
4794 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4795 ;; got one already.
4801 ;; Each time round, the next `while' moves forward over a pseudo match
4802 ;; of `c-decl-prefix-or-start-re' which is either inside a literal, or
4803 ;; is a ":" not preceded by "public", etc.. `cfd-re-match' and
4804 ;; `cfd-re-match-end' get set.
4811 ;; No match. Finish up and exit the loop.
4813 nil)
4816 ;; Matched the end of a token preceding a decl spot.
4820 ;; Matched a token that start a decl spot.
4823 c-literal-faces)
4824 ;; Pseudo match inside a comment or string literal. Skip out
4825 ;; of comments and string literals.
4837 ;; Found a ":" which isn't part of "public:", etc.
4838 t)
4841 ;; If our match was at the decl start, we have to back up over the
4842 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4843 ;; any decl spots in the syntactic ws.
4848 ;; Choose whichever match is closer to the start.
4851 cfd-re-match nil)
4853 cfd-prop-match nil))
4858 ;; Skip forward past comments only so we don't skip macros.
4860 ;; Set the position to continue at. We can avoid going over
4861 ;; the comments skipped above a second time, but it's possible
4862 ;; that the comment skipping has taken us past `cfd-prop-match'
4863 ;; since the property might be used inside comments.
4869 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4870 ;; label from the point to CFD-LIMIT.
4871 ;;
4872 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4873 ;; arguments: The first is the end position of the token preceding the spot,
4874 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4875 ;; the match is inside a macro. Point should be moved forward by at least
4876 ;; one token.
4877 ;;
4878 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4879 ;; it should return non-nil to ensure that the next search will find them.
4880 ;;
4881 ;; Such a spot is:
4882 ;; o The first token after bob.
4883 ;; o The first token after the end of submatch 1 in
4884 ;; `c-decl-prefix-or-start-re' when that submatch matches. This
4885 ;; submatch is typically a (L or R) brace or paren, a ;, or a ,.
4886 ;; o The start of each `c-decl-prefix-or-start-re' match when
4887 ;; submatch 1 doesn't match. This is, for example, the keyword
4888 ;; "class" in Pike.
4889 ;; o The start of a previously recognized declaration; "recognized"
4890 ;; means that the last char of the previous token has a `c-type'
4891 ;; text property with the value `c-decl-end'; this only holds
4892 ;; when `c-type-decl-end-used' is set.
4893 ;;
4894 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4895 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4896 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4897 ;;
4898 ;; If the match is inside a macro then the buffer is narrowed to the
4899 ;; end of it, so that CFD-FUN can investigate the following tokens
4900 ;; without matching something that begins inside a macro and ends
4901 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4902 ;; CFD-FACE-CHECKLIST checks exist.
4903 ;;
4904 ;; The spots are visited approximately in order from top to bottom.
4905 ;; It's however the positions where `c-decl-prefix-or-start-re'
4906 ;; matches and where `c-decl-end' properties are found that are in
4907 ;; order. Since the spots often are at the following token, they
4908 ;; might be visited out of order insofar as more spots are reported
4909 ;; later on within the syntactic whitespace between the match
4910 ;; positions and their spots.
4911 ;;
4912 ;; It's assumed that comments and strings are fontified in the
4913 ;; searched range.
4914 ;;
4915 ;; This is mainly used in fontification, and so has an elaborate
4916 ;; cache to handle repeated calls from the same start position; see
4917 ;; the variables above.
4918 ;;
4919 ;; All variables in this function begin with `cfd-' to avoid name
4920 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4921 ;;
4922 ;; This function might do hidden buffer changes.
4926 ;; The end of the token preceding the decl spot last found
4927 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4928 ;; no match.
4929 cfd-re-match
4930 ;; The end position of the last `c-decl-prefix-or-start-re'
4931 ;; match. If this is greater than `cfd-continue-pos', the
4932 ;; next regexp search is started here instead.
4934 ;; The end of the last `c-decl-end' found by
4935 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4936 ;; match. If searching for the property isn't needed then we
4937 ;; disable it by setting it to `cfd-limit' directly.
4939 ;; The end of the token preceding the decl spot last found by
4940 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4941 ;; bob. `cfd-limit' if there's no match. In other words,
4942 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4944 ;; The position to continue searching at.
4945 cfd-continue-pos
4946 ;; The position of the last "real" token we've stopped at.
4947 ;; This can be greater than `cfd-continue-pos' when we get
4948 ;; hits inside macros or at `c-decl-end' positions inside
4949 ;; comments.
4951 ;; The end position of the last entered macro.
4954 ;; Initialize by finding a syntactically relevant start position
4955 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4956 ;; search unless we're at bob.
4959 ;; Must back up a bit since we look for the end of the previous
4960 ;; statement or declaration, which is earlier than the first
4961 ;; returned match.
4963 ;; This `cond' moves back over any literals or macros. It has special
4964 ;; handling for when the region being searched is entirely within a
4965 ;; macro. It sets `cfd-continue-pos' (unless we've reached
4966 ;; `cfd-limit').
4968 ;; First we need to move to a syntactically relevant position.
4969 ;; Begin by backing out of comment or string literals.
4970 ;;
4971 ;; This arm of the cond actually triggers if we're in a literal,
4972 ;; and cfd-limit is at most at BONL.
4974 ;; This arm of the `and' moves backwards out of a literal when
4975 ;; the face at point is a literal face. In this case, its value
4976 ;; is always non-nil.
4978 ;; Try to use the faces to back up to the start of the
4979 ;; literal. FIXME: What if the point is on a declaration
4980 ;; inside a comment?
4986 ;; XEmacs doesn't fontify the quotes surrounding string
4987 ;; literals.
4996 ;; Don't trust the literal to contain only literal faces
4997 ;; (the font lock package might not have fontified the
4998 ;; start of it at all, for instance) so check that we have
4999 ;; arrived at something that looks like a start or else
5000 ;; resort to `c-literal-limits'.
5007 ;; The start is in a literal. If the limit is in the same
5008 ;; one we don't have to find a syntactic position etc. We
5009 ;; only check that if the limit is at or before bonl to save
5010 ;; time; it covers the by far most common case when font-lock
5011 ;; refontifies the current line only.
5022 ;; Completely inside a literal. Set up variables to trig the
5023 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
5024 ;; find a suitable start position.
5027 ;; Check if the region might be completely inside a macro, to
5028 ;; optimize that like the completely-inside-literal above.
5035 ;; (Maybe) completely inside a macro. Only need to trig the
5036 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
5037 ;; set things up.
5039 start-in-macro t))
5041 ;; The default arm of the `cond' moves back over any macro we're in
5042 ;; and over any syntactic WS. It sets `c-find-decl-syntactic-pos'.
5044 ;; Back out of any macro so we don't miss any declaration
5045 ;; that could follow after it.
5049 ;; Now we're at a proper syntactically relevant position so we
5050 ;; can use the cache. But first clear it if it applied
5051 ;; further down.
5056 ;; Don't have to do this if the cache is relevant here,
5057 ;; typically if the same line is refontified again. If
5058 ;; we're just some syntactic whitespace further down we can
5059 ;; still use the cache to limit the skipping.
5062 ;; If we hit `c-find-decl-syntactic-pos' and
5063 ;; `c-find-decl-match-pos' is set then we install the cached
5064 ;; values. If we hit `c-find-decl-syntactic-pos' and
5065 ;; `c-find-decl-match-pos' is nil then we know there's no decl
5066 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
5067 ;; and so we can continue the search from this point. If we
5068 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
5069 ;; the right spot to begin searching anyway.
5071 c-find-decl-match-pos)
5073 cfd-continue-pos syntactic-pos)
5078 ;; Always consider bob a match to get the first
5079 ;; declaration in the file. Do this separately instead of
5080 ;; letting `c-decl-prefix-or-start-re' match bob, so that
5081 ;; regexp always can consume at least one character to
5082 ;; ensure that we won't get stuck in an infinite loop.
5087 ;; Do an initial search now. In the bob case above it's
5088 ;; only done to search for a `c-decl-end' spot.
5094 ;; Advance `cfd-continue-pos' if it's before the start position.
5095 ;; The closest continue position that might have effect at or
5096 ;; after the start depends on what we started in. This also
5097 ;; finds a suitable start position in the special cases when the
5098 ;; region is completely within a literal or macro.
5103 ;; If we're in a macro then it's the closest preceding token
5104 ;; in the macro. Check this before `start-in-literal',
5105 ;; since if we're inside a literal in a macro, the preceding
5106 ;; token is earlier than any `c-decl-end' spot inside the
5107 ;; literal (comment).
5109 ;; The only syntactic ws in macros are comments.
5115 ;; If we're in a comment it can only be the closest
5116 ;; preceding `c-decl-end' position within that comment, if
5117 ;; any. Go back to the beginning of such a property so that
5118 ;; `c-find-decl-prefix-search' will find the end of it.
5119 ;; (Can't stop at the end and install it directly on
5120 ;; `cfd-prop-match' since that variable might be cleared
5121 ;; after `cfd-fun' below.)
5122 ;;
5123 ;; Note that if the literal is a string then the property
5124 ;; search will simply skip to the beginning of it right
5125 ;; away.
5137 ;; Didn't find any property inside the comment, so we can
5138 ;; skip it entirely. (This won't skip past a string, but
5139 ;; that'll be handled quickly by the next
5140 ;; `c-find-decl-prefix-search' anyway.)
5146 ;; If we started in normal code, the only match that might
5147 ;; apply before the start is what we already got in
5148 ;; `cfd-match-pos' so we can continue at the start position.
5149 ;; (Note that we don't get here if the first match is below
5150 ;; it.)
5153 ;; Delete found matches if they are before our new continue
5154 ;; position, so that `c-find-decl-prefix-search' won't back up
5155 ;; to them later on.
5163 ;; This is the normal case and we got a proper syntactic
5164 ;; position. If there's a match then it's always outside
5165 ;; macros and comments, so advance to the next token and set
5166 ;; `cfd-token-pos'. The loop below will later go back using
5167 ;; `cfd-continue-pos' to fix declarations inside the
5168 ;; syntactic ws.
5176 ;; Have one of the special cases when the region is completely
5177 ;; within a literal or macro. `cfd-continue-pos' is set to a
5178 ;; good start position for the search, so do it.
5181 ;; Now loop, one decl spot per iteration. We already have the first
5182 ;; match in `cfd-match-pos'.
5184 ;; Go forward over "false matches", one per iteration.
5189 ;; Kludge to filter out matches on the "<" that
5190 ;; aren't open parens, for the sake of languages
5191 ;; that got `c-recognize-<>-arglists' set.
5196 ;; If `cfd-continue-pos' is less or equal to
5197 ;; `cfd-token-pos', we've got a hit inside a macro
5198 ;; that's in the syntactic whitespace before the last
5199 ;; "real" declaration we've checked. If they're equal
5200 ;; we've arrived at the declaration a second time, so
5201 ;; there's nothing to do.
5205 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5206 ;; we're still searching for declarations embedded in
5207 ;; the syntactic whitespace. In that case we need
5208 ;; only to skip comments and not macros, since they
5209 ;; can't be nested, and that's already been done in
5210 ;; `c-find-decl-prefix-search'.
5215 ;; Continue if the following token fails the
5216 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5223 t)))
5234 ;; Narrow to the end of the macro if we got a hit inside
5235 ;; one, to avoid recognizing things that start inside the
5236 ;; macro and end outside it.
5238 ;; Not in the same macro as in the previous round.
5249 t
5252 t)
5253 ;; The matched token was the last thing in the macro,
5254 ;; so the whole match is bogus.
5263 ;; Restore limits if we did macro narrowing above.
5270 ; cfd-match-pos, etc.
5272 \f
5273 ;; A cache for found types.
5275 ;; Buffer local variable that contains an obarray with the types we've
5276 ;; found. If a declaration is recognized somewhere we record the
5277 ;; fully qualified identifier in it to recognize it as a type
5278 ;; elsewhere in the file too. This is not accurate since we do not
5279 ;; bother with the scoping rules of the languages, but in practice the
5280 ;; same name is seldom used as both a type and something else in a
5281 ;; file, and we only use this as a last resort in ambiguous cases (see
5282 ;; `c-forward-decl-or-cast-1').
5283 ;;
5284 ;; Not every type need be in this cache. However, things which have
5285 ;; ceased to be types must be removed from it.
5286 ;;
5287 ;; Template types in C++ are added here too but with the template
5288 ;; arglist replaced with "<>" in references or "<" for the one in the
5289 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5290 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5291 ;; template specs can be fairly sized programs in themselves) and
5292 ;; improves the hit ratio (it's a type regardless of the template
5293 ;; args; it's just not the same type, but we're only interested in
5294 ;; recognizing types, not telling distinct types apart). Note that
5295 ;; template types in references are added here too; from the example
5296 ;; above there will also be an entry "Foo<".
5301 ;; Clears `c-found-types'.
5305 ;; Add the given region as a type in `c-found-types'. If the region
5306 ;; doesn't match an existing type but there is a type which is equal
5307 ;; to the given one except that the last character is missing, then
5308 ;; the shorter type is removed. That's done to avoid adding all
5309 ;; prefixes of a type as it's being entered and font locked. This
5310 ;; doesn't cover cases like when characters are removed from a type
5311 ;; or added in the middle. We'd need the position of point when the
5312 ;; font locking is invoked to solve this well.
5313 ;;
5314 ;; This function might do hidden buffer changes.
5321 ;; Remove the "NAME" from c-found-types, if present.
5325 ;; Return non-nil if the given region contains a type in
5326 ;; `c-found-types'.
5327 ;;
5328 ;; This function might do hidden buffer changes.
5330 c-found-types))
5333 ;; Return all the types in `c-found-types' as a sorted list of
5334 ;; strings.
5338 type-list)))
5339 c-found-types)
5342 ;; Shut up the byte compiler.
5346 ;; An after change function which, in conjunction with the info in
5347 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5348 ;; from `c-found-types', should this type have become stale. For
5349 ;; example, this happens to "foo" when "foo \n bar();" becomes
5350 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5351 ;; the fontification.
5352 ;;
5353 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5354 ;; type?
5374 ;; Changing the amount of (already existing) whitespace - don't do anything.
5379 ;; The syntactic relationship which defined a "found type" has been
5380 ;; destroyed.
5383 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5384 )))
5386 \f
5387 ;; Setting and removing syntax properties on < and > in languages (C++
5388 ;; and Java) where they can be template/generic delimiters as well as
5389 ;; their normal meaning of "less/greater than".
5391 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5392 ;; be delimiters, they are marked as such with the category properties
5393 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5395 ;; STRATEGY:
5396 ;;
5397 ;; It is impossible to determine with certainty whether a <..> pair in
5398 ;; C++ is two comparison operators or is template delimiters, unless
5399 ;; one duplicates a lot of a C++ compiler. For example, the following
5400 ;; code fragment:
5401 ;;
5402 ;; foo (a < b, c > d) ;
5403 ;;
5404 ;; could be a function call with two integer parameters (each a
5405 ;; relational expression), or it could be a constructor for class foo
5406 ;; taking one parameter d of templated type "a < b, c >". They are
5407 ;; somewhat easier to distinguish in Java.
5408 ;;
5409 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5410 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5411 ;; individually when their context so indicated. This gave rise to
5412 ;; intractable problems when one of a matching pair was deleted, or
5413 ;; pulled into a literal.]
5414 ;;
5415 ;; At each buffer change, the syntax-table properties are removed in a
5416 ;; before-change function and reapplied, when needed, in an
5417 ;; after-change function. It is far more important that the
5418 ;; properties get removed when they they are spurious than that they
5419 ;; be present when wanted.
5420 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5422 ;; POS (default point) is at a < character. If it is marked with
5423 ;; open paren syntax-table text property, remove the property,
5424 ;; together with the close paren property on the matching > (if
5425 ;; any).
5440 ;; POS (default point) is at a > character. If it is marked with
5441 ;; close paren syntax-table property, remove the property, together
5442 ;; with the open paren property on the matching < (if any).
5457 ;; POS (default point) is at a < or > character. If it has an
5458 ;; open/close paren syntax-table property, remove this property both
5459 ;; from the current character and its partner (which will also be
5460 ;; thusly marked).
5470 ;; POS (default point) is at a < character. If it is both marked
5471 ;; with open/close paren syntax-table property, and has a matching >
5472 ;; (also marked) which is after LIM, remove the property both from
5473 ;; the current > and its partner. Return t when this happens, nil
5474 ;; when it doesn't.
5488 t)))
5491 ;; POS (default point) is at a > character. If it is both marked
5492 ;; with open/close paren syntax-table property, and has a matching <
5493 ;; (also marked) which is before LIM, remove the property both from
5494 ;; the current < and its partner. Return t when this happens, nil
5495 ;; when it doesn't.
5509 t)))
5511 ;; Set by c-common-init in cc-mode.el.
5516 ;; Unmark certain pairs of "< .... >" which are currently marked as
5517 ;; template/generic delimiters. (This marking is via syntax-table
5518 ;; text properties).
5519 ;;
5520 ;; These pairs are those which are in the current "statement" (i.e.,
5521 ;; the region between the {, }, or ; before BEG and the one after
5522 ;; END), and which enclose any part of the interval (BEG END).
5523 ;;
5524 ;; Note that in C++ (?and Java), template/generic parens cannot
5525 ;; enclose a brace or semicolon, so we use these as bounds on the
5526 ;; region we must work on.
5527 ;;
5528 ;; This function is called from before-change-functions (via
5529 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5530 ;; and point is undefined, both at entry and exit.
5531 ;;
5532 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5533 ;; 2010-01-29.
5537 new-beg new-end need-new-beg need-new-end)
5538 ;; Locate the barrier before the changed region
5543 ;; Remove the syntax-table properties from each pertinent <...> pair.
5544 ;; Firsly, the ones with the < before beg and > after beg.
5549 ;; Locate the barrier after END.
5554 ;; Remove syntax-table properties from the remaining pertinent <...>
5555 ;; pairs, those with a > after end and < before end.
5560 ;; Extend the fontification region, if needed.
5570 ;; This is called from `after-change-functions' when
5571 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5572 ;; chars with paren syntax become part of another operator like "<<"
5573 ;; or ">=".
5574 ;;
5575 ;; This function might do hidden buffer changes.
5608 \f
5609 ;; Handling of small scale constructs like types and names.
5611 ;; Dynamically bound variable that instructs `c-forward-type' to also
5612 ;; treat possible types (i.e. those that it normally returns 'maybe or
5613 ;; 'found for) as actual types (and always return 'found for them).
5614 ;; This means that it records them in `c-record-type-identifiers' if
5615 ;; that is set, and that it adds them to `c-found-types'.
5618 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5619 ;; mark up successfully parsed arglists with paren syntax properties on
5620 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5621 ;; `c-type' property of each argument separating comma.
5622 ;;
5623 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5624 ;; all arglists for side effects (i.e. recording types), otherwise it
5625 ;; exploits any existing paren syntax properties to quickly jump to the
5626 ;; end of already parsed arglists.
5627 ;;
5628 ;; Marking up the arglists is not the default since doing that correctly
5629 ;; depends on a proper value for `c-restricted-<>-arglists'.
5632 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5633 ;; not accept arglists that contain binary operators.
5634 ;;
5635 ;; This is primarily used to handle C++ template arglists. C++
5636 ;; disambiguates them by checking whether the preceding name is a
5637 ;; template or not. We can't do that, so we assume it is a template
5638 ;; if it can be parsed as one. That usually works well since
5639 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5640 ;; in almost all cases would be pointless.
5641 ;;
5642 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5643 ;; should let the comma separate the function arguments instead. And
5644 ;; in a context where the value of the expression is taken, e.g. in
5645 ;; "if (a < b || c > d)", it's probably not a template.
5648 ;; Dynamically bound variables that instructs
5649 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5650 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5651 ;; `c-forward-label' to record the ranges of all the type and
5652 ;; reference identifiers they encounter. They will build lists on
5653 ;; these variables where each element is a cons of the buffer
5654 ;; positions surrounding each identifier. This recording is only
5655 ;; activated when `c-record-type-identifiers' is non-nil.
5656 ;;
5657 ;; All known types that can't be identifiers are recorded, and also
5658 ;; other possible types if `c-promote-possible-types' is set.
5659 ;; Recording is however disabled inside angle bracket arglists that
5660 ;; are encountered inside names and other angle bracket arglists.
5661 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5662 ;; instead.
5663 ;;
5664 ;; Only the names in C++ template style references (e.g. "tmpl" in
5665 ;; "tmpl<a,b>::foo") are recorded as references, other references
5666 ;; aren't handled here.
5667 ;;
5668 ;; `c-forward-label' records the label identifier(s) on
5669 ;; `c-record-ref-identifiers'.
5673 ;; This variable will receive a cons cell of the range of the last
5674 ;; single identifier symbol stepped over by `c-forward-name' if it's
5675 ;; successful. This is the range that should be put on one of the
5676 ;; record lists above by the caller. It's assigned nil if there's no
5677 ;; such symbol in the name.
5682 ;; Always true.
5692 ;; Always true.
5700 ;; Dynamically bound variable that instructs `c-forward-type' to
5701 ;; record the ranges of types that only are found. Behaves otherwise
5702 ;; like `c-record-type-identifiers'.
5706 ;; Used internally in `c-forward-keyword-clause' to move forward
5707 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5708 ;; possibly is prefixed by keywords and their associated clauses.
5709 ;; Try with a type/name first to not trip up on those that begin
5710 ;; with a keyword. Return t if a known or found type is moved
5711 ;; over. The point is clobbered if nil is returned. If range
5712 ;; recording is enabled, the identifier is recorded on as a type
5713 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5714 ;;
5715 ;; This macro might do hidden buffer changes.
5720 nil
5727 t)))
5730 ;; Used internally in `c-forward-keyword-clause' to move forward
5731 ;; over a comma separated list of types or names using
5732 ;; `c-forward-keyword-prefixed-id'.
5733 ;;
5734 ;; This macro might do hidden buffer changes.
5745 ;; Submatch MATCH in the current match data is assumed to surround a
5746 ;; token. If it's a keyword, move over it and any immediately
5747 ;; following clauses associated with it, stopping at the start of
5748 ;; the next token. t is returned in that case, otherwise the point
5749 ;; stays and nil is returned. The kind of clauses that are
5750 ;; recognized are those specified by `c-type-list-kwds',
5751 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5752 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5753 ;; and `c-<>-arglist-kwds'.
5754 ;;
5755 ;; This function records identifier ranges on
5756 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5757 ;; `c-record-type-identifiers' is non-nil.
5758 ;;
5759 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5760 ;; apply directly after the keyword, the type list is moved over
5761 ;; only when there is no unaccounted token before it (i.e. a token
5762 ;; that isn't moved over due to some other keyword list). The
5763 ;; identifier ranges in the list are still recorded if that should
5764 ;; be done, though.
5765 ;;
5766 ;; This function might do hidden buffer changes.
5769 ;; The call to `c-forward-<>-arglist' below is made after
5770 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5771 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5772 ;; should therefore be nil.
5784 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5789 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5794 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5801 ;; Use `c-forward-type' on every identifier we can find
5802 ;; inside the paren, to record the types.
5827 ;; If we are at the colon already, we move over the type
5828 ;; list after it.
5834 ;; Not at the colon, so stop here. But the identifier
5835 ;; ranges in the type list later on should still be
5836 ;; recorded.
5839 ;; If a keyword matched both one of the types above and
5840 ;; this one, we match `c-colon-type-list-re' after the
5841 ;; clause matched above.
5848 ;; There's a type after the `c-colon-type-list-re' match
5849 ;; after a keyword in `c-colon-type-list-kwds'.
5853 t)))
5855 ;; cc-mode requires cc-fonts.
5859 ;; The point is assumed to be at a "<". Try to treat it as the open
5860 ;; paren of an angle bracket arglist and move forward to the
5861 ;; corresponding ">". If successful, the point is left after the
5862 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5863 ;; returned. If ALL-TYPES is t then all encountered arguments in
5864 ;; the arglist that might be types are treated as found types.
5865 ;;
5866 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5867 ;; function handles text properties on the angle brackets and argument
5868 ;; separating commas.
5869 ;;
5870 ;; `c-restricted-<>-arglists' controls how lenient the template
5871 ;; arglist recognition should be.
5872 ;;
5873 ;; This function records identifier ranges on
5874 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5875 ;; `c-record-type-identifiers' is non-nil.
5876 ;;
5877 ;; This function might do hidden buffer changes.
5880 ;; If `c-record-type-identifiers' is set then activate
5881 ;; recording of any found types that constitute an argument in
5882 ;; the arglist.
5890 ;; `nconc' doesn't mind that the tail of
5891 ;; `c-record-found-types' is t.
5894 t)
5897 nil)))
5900 ;; Recursive part of `c-forward-<>-arglist'.
5901 ;;
5902 ;; This function might do hidden buffer changes.
5904 ;; Cover this so that any recorded found type ranges are
5905 ;; automatically lost if it turns out to not be an angle
5906 ;; bracket arglist. It's propagated through the return value
5907 ;; on successful completion.
5909 ;; List that collects the positions after the argument
5910 ;; separating ',' in the arglist.
5911 arg-start-pos)
5912 ;; If the '<' has paren open syntax then we've marked it as an angle
5913 ;; bracket arglist before, so skip to the end.
5921 t
5922 ;; Got unmatched paren angle brackets. We don't clear the paren
5923 ;; syntax properties and retry, on the basis that it's very
5924 ;; unlikely that paren angle brackets become operators by code
5925 ;; manipulation. It's far more likely that it doesn't match due
5926 ;; to narrowing or some temporary change.
5928 nil))
5933 ;; go forward one non-alphanumeric character (group) per iteration of
5934 ;; this loop.
5941 ;; All encountered identifiers are types, so set the
5942 ;; promote flag and parse the type.
5965 ;; Note: These regexps exploit the match order in \| so
5966 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5968 ;; Stop on ',', '|', '&', '+' and '-' to catch
5969 ;; common binary operators that could be between
5970 ;; two comparison expressions "a<b" and "c>d".
5972 nil t t))
5976 ;; Either an operator starting with '>' or the end of
5977 ;; the angle bracket arglist.
5984 ;; The angle bracket arglist is finished.
5996 ;; Either an operator starting with '<' or a nested arglist.
6000 ;; The '<' begins a multi-char operator.
6004 ;; We're at a nested <.....>
6010 ;; There's always an identifier before an angle
6011 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
6012 ;; or `c-<>-arglist-kwds'.
6028 ;; It was an angle bracket arglist.
6031 ;; Record the identifier before the template as a type
6032 ;; or reference depending on whether the arglist is last
6033 ;; in a qualified identifier.
6043 ;; At a "less than" operator.
6046 )))
6053 ;; Just another argument. Record the position. The
6054 ;; type check stuff that made us stop at it is at
6055 ;; the top of the loop.
6059 ;; Got a character that can't be in an angle bracket
6060 ;; arglist argument. Abort using `throw', since
6061 ;; it's useless to try to find a surrounding arglist
6062 ;; if we're nested.
6068 ;; The point is assumed to be directly after a ">". Try to treat it
6069 ;; as the close paren of an angle bracket arglist and move back to
6070 ;; the corresponding "<". If successful, the point is left at
6071 ;; the "<" and t is returned, otherwise the point isn't moved and
6072 ;; nil is returned. ALL-TYPES is passed on to
6073 ;; `c-forward-<>-arglist'.
6074 ;;
6075 ;; If the optional LIMIT is given, it bounds the backward search.
6076 ;; It's then assumed to be at a syntactically relevant position.
6077 ;;
6078 ;; This is a wrapper around `c-forward-<>-arglist'. See that
6079 ;; function for more details.
6088 t
6089 ;; See corresponding note in `c-forward-<>-arglist'.
6091 nil)
6098 t
6099 ;; Stopped at bob or a char that isn't allowed in an
6100 ;; arglist, so we've failed.
6102 nil)
6107 ;; If we moved then the "<" was part of some
6108 ;; multicharacter token.
6109 t
6115 ;; Matched the arglist. Break the while.
6117 nil)
6119 ;; We started from a non-paren ">" inside an
6120 ;; arglist.
6122 nil)
6124 ;; Matched a shorter arglist. Can be a nested
6125 ;; one so continue looking.
6127 t))
6128 t))))))
6133 ;; Move forward over a complete name if at the beginning of one,
6134 ;; stopping at the next following token. A keyword, as such,
6135 ;; doesn't count as a name. If the point is not at something that
6136 ;; is recognized as a name then it stays put.
6137 ;;
6138 ;; A name could be something as simple as "foo" in C or something as
6139 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6140 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6141 ;; int>::*volatile const" in C++ (this function is actually little
6142 ;; more than a `looking-at' call in all modes except those that,
6143 ;; like C++, have `c-recognize-<>-arglists' set).
6144 ;;
6145 ;; Return
6146 ;; o - nil if no name is found;
6147 ;; o - 'template if it's an identifier ending with an angle bracket
6148 ;; arglist;
6149 ;; o - 'operator of it's an operator identifier;
6150 ;; o - t if it's some other kind of name.
6151 ;;
6152 ;; This function records identifier ranges on
6153 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6154 ;; `c-record-type-identifiers' is non-nil.
6155 ;;
6156 ;; This function might do hidden buffer changes.
6159 ;; Turn off `c-promote-possible-types' here since we might
6160 ;; call `c-forward-<>-arglist' and we don't want it to promote
6161 ;; every suspect thing in the arglist to a type. We're
6162 ;; typically called from `c-forward-type' in this case, and
6163 ;; the caller only wants the top level type that it finds to
6164 ;; be promoted.
6165 c-promote-possible-types)
6171 ;; Check for keyword. We go to the last symbol in
6172 ;; `c-identifier-key' first.
6185 ;; "template" is only valid inside an
6186 ;; identifier if preceded by "::".
6191 t))
6193 ;; Handle a C++ operator or template identifier.
6197 ;; Got "... ::template".
6201 res subres))))
6204 ;; Got a cast operator.
6208 ;; Now we should match a sequence of either
6209 ;; '*', '&' or a name followed by ":: *",
6210 ;; where each can be followed by a sequence
6211 ;; of `c-opt-type-modifier-key'.
6214 t)
6224 t))))
6232 ;; Got some other operator.
6240 nil)
6242 ;; `id-start' is equal to `id-end' if we've jumped over
6243 ;; an identifier that doesn't end with a symbol token.
6244 ;; That can occur e.g. for Java import directives on the
6245 ;; form "foo.bar.*".
6252 res t)))
6262 ;; Got a concatenated identifier. This handles the
6263 ;; cases with tricky syntactic whitespace that aren't
6264 ;; covered in `c-identifier-key'.
6267 t)
6271 ;; Maybe an angle bracket arglist.
6279 c-last-identifier-range nil)
6284 ;; Continue if there's an identifier concatenation
6285 ;; operator after the template argument.
6291 t)
6296 nil)))
6297 )))))
6300 res))
6303 ;; Move forward over a type spec if at the beginning of one,
6304 ;; stopping at the next following token. The keyword "typedef"
6305 ;; isn't part of a type spec here.
6306 ;;
6307 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6308 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6309 ;; The current (2009-03-10) intention is to convert all uses of
6310 ;; `c-forward-type' to call with this parameter set, then to
6311 ;; eliminate it.
6312 ;;
6313 ;; Return
6314 ;; o - t if it's a known type that can't be a name or other
6315 ;; expression;
6316 ;; o - 'known if it's an otherwise known type (according to
6317 ;; `*-font-lock-extra-types');
6318 ;; o - 'prefix if it's a known prefix of a type;
6319 ;; o - 'found if it's a type that matches one in `c-found-types';
6320 ;; o - 'maybe if it's an identifier that might be a type;
6321 ;; o - 'decltype if it's a decltype(variable) declaration; - or
6322 ;; o - nil if it can't be a type (the point isn't moved then).
6323 ;;
6324 ;; The point is assumed to be at the beginning of a token.
6325 ;;
6326 ;; Note that this function doesn't skip past the brace definition
6327 ;; that might be considered part of the type, e.g.
6328 ;; "enum {a, b, c} foo".
6329 ;;
6330 ;; This function records identifier ranges on
6331 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6332 ;; `c-record-type-identifiers' is non-nil.
6333 ;;
6334 ;; This function might do hidden buffer changes.
6342 ;; Skip leading type modifiers. If any are found we know it's a
6343 ;; prefix of a type.
6362 ; "typedef".
6370 ;; In many languages the name can be used without the
6371 ;; prefix, so we add it to `c-found-types'.
6374 c-last-identifier-range)
6396 id-range c-last-identifier-range))))
6408 id-end)
6410 ;; Looking at a primitive or known type identifier. We've
6411 ;; checked for a name first so that we don't go here if the
6412 ;; known type match only is a prefix of another name.
6423 ;; There might be more keywords for the type.
6451 ;; A normal identifier.
6461 ;; It's an identifier that has been used as
6462 ;; a type somewhere else.
6463 'found
6464 ;; It's an identifier that might be a type.
6467 ;; A template is a type.
6471 ;; Otherwise it's an operator identifier, which is not a type.
6476 ;; Skip trailing type modifiers. If any are found we know it's
6477 ;; a type.
6483 ;; Step over any type suffix operator. Do not let the existence
6484 ;; of these alter the classification of the found type, since
6485 ;; these operators typically are allowed in normal expressions
6486 ;; too.
6493 ;; Look for a trailing operator that concatenates the type
6494 ;; with a following one, and if so step past that one through
6495 ;; a recursive call. Note that we don't record concatenated
6496 ;; types in `c-found-types' - it's the component types that
6497 ;; are recorded when appropriate.
6500 c-promote-possible-types))
6501 ;; If we can't promote then set `c-record-found-types' so that
6502 ;; we can merge in the types from the second part afterwards if
6503 ;; it turns out to be a known type there.
6506 subres)
6515 ;; If either operand certainly is a type then both are, but we
6516 ;; don't let the existence of the operator itself promote two
6517 ;; uncertain types to a certain one.
6536 ;; Merge in the ranges of any types found by the second
6537 ;; `c-forward-type'.
6539 ;; `nconc' doesn't mind that the tail of
6540 ;; `c-record-found-types' is t.
6542 c-record-type-identifiers))))
6550 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6552 res))
6555 ;; Used for Java code only at the moment. Assumes point is on the
6556 ;; @, moves forward an annotation. returns nil if there is no
6557 ;; annotation at point.
6564 t)))
6567 ;; Pull the next open brace from PS (which has the form of paren-state),
6568 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
6576 ;; Test whether we are in a C++ member initializer list, and if so, go back
6577 ;; to the introducing ":", returning the position of the opening paren of
6578 ;; the function's arglist. Otherwise return nil, leaving point unchanged.
6581 res)
6619 res))
6621 \f
6622 ;; Handling of large scale constructs like statements and declarations.
6624 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6625 ;; defsubst or perhaps even a defun, but it contains lots of free
6626 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6628 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6629 ;; of types when parsing a declaration, which means that it
6630 ;; sometimes consumes the identifier in the declaration as a type.
6631 ;; This is used to "backtrack" and make the last type be treated as
6632 ;; an identifier instead.
6635 ;; These identifiers are bound only in the inner let.
6637 identifier-start type-start
6638 got-parens nil
6639 got-identifier t
6640 got-suffix t
6641 got-suffix-after-parens id-start
6645 t
6646 backup-at-type))
6648 id-start backup-id-start)
6650 id-start start-pos))
6652 ;; When these flags already are set we've found specifiers that
6653 ;; unconditionally signal these attributes - backtracking doesn't
6654 ;; change that. So keep them set in that case.
6661 ;; This identifier is bound only in the inner let.
6665 ;; Move forward over a declaration or a cast if at the start of one.
6666 ;; The point is assumed to be at the start of some token. Nil is
6667 ;; returned if no declaration or cast is recognized, and the point
6668 ;; is clobbered in that case.
6669 ;;
6670 ;; If a declaration is parsed:
6671 ;;
6672 ;; The point is left at the first token after the first complete
6673 ;; declarator, if there is one. The return value is a cons where
6674 ;; the car is the position of the first token in the declarator. (See
6675 ;; below for the cdr.)
6676 ;; Some examples:
6677 ;;
6678 ;; void foo (int a, char *b) stuff ...
6679 ;; car ^ ^ point
6680 ;; float (*a)[], b;
6681 ;; car ^ ^ point
6682 ;; unsigned int a = c_style_initializer, b;
6683 ;; car ^ ^ point
6684 ;; unsigned int a (cplusplus_style_initializer), b;
6685 ;; car ^ ^ point (might change)
6686 ;; class Foo : public Bar {}
6687 ;; car ^ ^ point
6688 ;; class PikeClass (int a, string b) stuff ...
6689 ;; car ^ ^ point
6690 ;; enum bool;
6691 ;; car ^ ^ point
6692 ;; enum bool flag;
6693 ;; car ^ ^ point
6694 ;; void cplusplus_function (int x) throw (Bad);
6695 ;; car ^ ^ point
6696 ;; Foo::Foo (int b) : Base (b) {}
6697 ;; car ^ ^ point
6698 ;;
6699 ;; auto foo = 5;
6700 ;; car ^ ^ point
6701 ;; auto cplusplus_11 (int a, char *b) -> decltype (bar):
6702 ;; car ^ ^ point
6703 ;;
6704 ;;
6705 ;;
6706 ;; The cdr of the return value is non-nil when a
6707 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6708 ;; Specifically it is a dotted pair (A . B) where B is t when a
6709 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6710 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6711 ;; specifier is present. I.e., (some of) the declared
6712 ;; identifier(s) are types.
6713 ;;
6714 ;; If a cast is parsed:
6715 ;;
6716 ;; The point is left at the first token after the closing paren of
6717 ;; the cast. The return value is `cast'. Note that the start
6718 ;; position must be at the first token inside the cast parenthesis
6719 ;; to recognize it.
6720 ;;
6721 ;; PRECEDING-TOKEN-END is the first position after the preceding
6722 ;; token, i.e. on the other side of the syntactic ws from the point.
6723 ;; Use a value less than or equal to (point-min) if the point is at
6724 ;; the first token in (the visible part of) the buffer.
6725 ;;
6726 ;; CONTEXT is a symbol that describes the context at the point:
6727 ;; 'decl In a comma-separated declaration context (typically
6728 ;; inside a function declaration arglist).
6729 ;; '<> In an angle bracket arglist.
6730 ;; 'arglist Some other type of arglist.
6731 ;; nil Some other context or unknown context. Includes
6732 ;; within the parens of an if, for, ... construct.
6733 ;;
6734 ;; LAST-CAST-END is the first token after the closing paren of a
6735 ;; preceding cast, or nil if none is known. If
6736 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6737 ;; the position after the closest preceding call where a cast was
6738 ;; matched. In that case it's used to discover chains of casts like
6739 ;; "(a) (b) c".
6740 ;;
6741 ;; This function records identifier ranges on
6742 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6743 ;; `c-record-type-identifiers' is non-nil.
6744 ;;
6745 ;; This function might do hidden buffer changes.
6748 ;; first type, i.e. after any leading specifiers. It might
6749 ;; also point at the beginning of the preceding syntactic
6750 ;; whitespace.
6752 ;; Set to the result of `c-forward-type'.
6753 at-type
6754 ;; The position of the first token in what we currently
6755 ;; believe is the type in the declaration or cast, after any
6756 ;; specifiers and their associated clauses.
6757 type-start
6758 ;; The position of the first token in what we currently
6759 ;; believe is the declarator for the first identifier. Set
6760 ;; when the type is found, and moved forward over any
6761 ;; `c-decl-hangon-kwds' and their associated clauses that
6762 ;; occurs after the type.
6763 id-start
6764 ;; These store `at-type', `type-start' and `id-start' of the
6765 ;; identifier before the one in those variables. The previous
6766 ;; identifier might turn out to be the real type in a
6767 ;; declaration if the last one has to be the declarator in it.
6768 ;; If `backup-at-type' is nil then the other variables have
6769 ;; undefined values.
6770 backup-at-type backup-type-start backup-id-start
6771 ;; This stores `kwd-sym' of the symbol before the current one.
6772 ;; This is needed to distinguish the C++11 version of "auto" from
6773 ;; the pre C++11 meaning.
6774 backup-kwd-sym
6775 ;; Set if we've found a specifier (apart from "typedef") that makes
6776 ;; the defined identifier(s) types.
6777 at-type-decl
6778 ;; Set if we've a "typedef" keyword.
6779 at-typedef
6780 ;; Set if we've found a specifier that can start a declaration
6781 ;; where there's no type.
6782 maybe-typeless
6783 ;; Save the value of kwd-sym between loops of the "Check for a
6784 ;; type" loop. Needed to distinguish a C++11 "auto" from a pre
6785 ;; C++11 one.
6786 prev-kwd-sym
6787 ;; If a specifier is found that also can be a type prefix,
6788 ;; these flags are set instead of those above. If we need to
6789 ;; back up an identifier, they are copied to the real flag
6790 ;; variables. Thus they only take effect if we fail to
6791 ;; interpret it as a type.
6792 backup-at-type-decl backup-maybe-typeless
6793 ;; Whether we've found a declaration or a cast. We might know
6794 ;; this before we've found the type in it. It's 'ids if we've
6795 ;; found two consecutive identifiers (usually a sure sign, but
6796 ;; we should allow that in labels too), and t if we've found a
6797 ;; specifier keyword (a 100% sure sign).
6798 at-decl-or-cast
6799 ;; Set when we need to back up to parse this as a declaration
6800 ;; but not as a cast.
6801 backup-if-not-cast
6802 ;; For casts, the return position.
6803 cast-end
6804 ;; Have we got a new-style C++11 "auto"?
6805 new-style-auto
6806 ;; Save `c-record-type-identifiers' and
6807 ;; `c-record-ref-identifiers' since ranges are recorded
6808 ;; speculatively and should be thrown away if it turns out
6809 ;; that it isn't a declaration or cast.
6816 ;; Check for a type. Unknown symbols are treated as possible
6817 ;; types, but they could also be specifiers disguised through
6818 ;; macros like __INLINE__, so we recognize both types and known
6819 ;; specifiers after them too.
6823 ;; Look for a specifier keyword clause.
6836 ;; Found a known or possible type or a prefix of a known type.
6845 ;; Got two identifiers with nothing but whitespace
6846 ;; between them. That can only happen in declarations.
6850 ;; If the previous identifier is a found type we
6851 ;; record it as a real one; it might be some sort of
6852 ;; alias for a prefix like "unsigned".
6859 backup-type-start type-start
6860 backup-id-start id-start
6861 backup-kwd-sym kwd-sym
6862 at-type found-type
6863 type-start start
6865 ;; The previous ambiguous specifier/type turned out
6866 ;; to be a type since we've parsed another one after
6867 ;; it, so clear these backup flags.
6868 backup-at-type-decl nil
6869 backup-maybe-typeless nil))
6873 ;; Handle known specifier keywords and
6874 ;; `c-decl-hangon-kwds' which can occur after known
6875 ;; types.
6878 ;; It's a hang-on keyword that can occur anywhere.
6882 ;; Move the identifier start position if
6883 ;; we've passed a type.
6885 ;; Otherwise treat this as a specifier and
6886 ;; move the fallback position.
6890 ;; It's an ordinary specifier so we know that
6891 ;; anything before this can't be the type.
6893 start-pos kwd-clause-end)
6896 ;; It's ambiguous whether this keyword is a
6897 ;; specifier or a type prefix, so set the backup
6898 ;; flags. (It's assumed that `c-forward-type'
6899 ;; moved further than `c-forward-keyword-clause'.)
6907 ;; This test only happens after we've scanned a type.
6908 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6913 ;; Haven't matched a type so it's an unambiguous
6914 ;; specifier keyword and we know we're in a
6915 ;; declaration.
6921 ;; If the type isn't known we continue so that we'll jump
6922 ;; over all specifiers and type identifiers. The reason
6923 ;; to do this for a known type prefix is to make things
6924 ;; like "unsigned INT16" work.
6929 ;; If a known type was found, we still need to skip over any
6930 ;; hangon keyword clauses after it. Otherwise it has already
6931 ;; been done in the loop above.
6937 ;; A prefix type is itself a primitive type when it's not
6938 ;; followed by another type.
6942 ;; Got no type but set things up to continue anyway to handle
6943 ;; the various cases when a declaration doesn't start with a
6944 ;; type.
6949 ;; If it's C++ then check if the last "type" ends on the form
6950 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
6951 ;; (con|de)structor.
6962 ;; Cheating in the handling of syntactic ws below.
6969 name))
6970 ;; It is a (con|de)structor name. In that case the
6971 ;; declaration is typeless so zap out any preceding
6972 ;; identifier(s) that we might have taken as types.
6975 backup-at-type nil
6976 id-start type-start))))))
6978 ;; Check for and step over a type decl expression after the thing
6979 ;; that is or might be a type. This can't be skipped since we
6980 ;; need the correct end position of the declarator for
6981 ;; `max-type-decl-end-*'.
6983 ;; True if there's a non-open-paren match of
6984 ;; `c-type-decl-prefix-key'.
6985 got-prefix
6986 ;; True if the declarator is surrounded by a parenthesis pair.
6987 got-parens
6988 ;; True if there is an identifier in the declarator.
6989 got-identifier
6990 ;; True if there's a non-close-paren match of
6991 ;; `c-type-decl-suffix-key'.
6992 got-suffix
6993 ;; True if there's a prefix match outside the outermost
6994 ;; paren pair that surrounds the declarator.
6995 got-prefix-before-parens
6996 ;; True if there's a suffix match outside the outermost
6997 ;; paren pair that surrounds the declarator. The value is
6998 ;; the position of the first suffix match.
6999 got-suffix-after-parens
7000 ;; True if we've parsed the type decl to a token that is
7001 ;; known to end declarations in this context.
7002 at-decl-end
7003 ;; The earlier values of `at-type' and `type-start' if we've
7004 ;; shifted the type backwards.
7005 identifier-type identifier-start
7006 ;; If `c-parse-and-markup-<>-arglists' is set we need to
7007 ;; turn it off during the name skipping below to avoid
7008 ;; getting `c-type' properties that might be bogus. That
7009 ;; can happen since we don't know if
7010 ;; `c-restricted-<>-arglists' will be correct inside the
7011 ;; arglist paren that gets entered.
7012 c-parse-and-markup-<>-arglists
7013 ;; Start of the identifier for which `got-identifier' was set.
7014 name-start)
7018 ;; Skip over type decl prefix operators. (Note similar code in
7019 ;; `c-font-lock-declarators'.)
7029 ;; If the third submatch matches in C++ then
7030 ;; we're looking at an identifier that's a
7031 ;; prefix only if it specifies a member pointer.
7036 ;; We only check for a trailing "::" and
7037 ;; let the "*" that should follow be
7038 ;; matched in the next round.
7040 ;; It turned out to be the real identifier,
7041 ;; so stop.
7042 nil))
7043 t))
7057 ;; Skip over an identifier.
7064 ;; Skip over type decl suffix operators.
7071 t)
7075 t)
7081 ;; No suffix matched. We might have matched the
7082 ;; identifier as a type and the open paren of a
7083 ;; function arglist as a type decl prefix. In that
7084 ;; case we should "backtrack": Reinterpret the last
7085 ;; type as the identifier, move out of the arglist and
7086 ;; continue searching for suffix operators.
7087 ;;
7088 ;; Do this even if there's no preceding type, to cope
7089 ;; with old style function declarations in K&R C,
7090 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
7091 ;; style declarations. That isn't applicable in an
7092 ;; arglist context, though.
7097 maybe-typeless
7098 backup-maybe-typeless
7105 t))
7114 at-type))
7115 ;; Have found no identifier but `c-typeless-decl-kwds' has
7116 ;; matched so we know we're inside a declaration. The
7117 ;; preceding type must be the identifier instead.
7120 ;; Prepare the "-> type;" for fontification later on.
7130 at-decl-or-cast
7133 ;; CASE 1
7135 ;; Encountered something inside parens that isn't matched by
7136 ;; the `c-type-decl-*' regexps, so it's not a type decl
7137 ;; expression. Try to skip out to the same paren depth to
7138 ;; not confuse the cast check below.
7140 ;; If we've found a specifier keyword then it's a
7141 ;; declaration regardless.
7149 ;; Now we've collected info about various characteristics of
7150 ;; the construct we're looking at. Below follows a decision
7151 ;; tree based on that. It's ordered to check more certain
7152 ;; signs before less certain ones.
7157 ;; CASE 2
7160 ;; Got another identifier directly after the type, so it's a
7161 ;; declaration.
7166 ;; (not got-suffix-after-parens)
7168 maybe-typeless
7169 backup-maybe-typeless
7174 ;; Got a declaration of the form "foo bar (gnu);" or "bar
7175 ;; (gnu);" where we've recognized "bar" as the type and "gnu"
7176 ;; as the declarator. In this case it's however more likely
7177 ;; that "bar" is the declarator and "gnu" a function argument
7178 ;; or initializer (if `c-recognize-paren-inits' is set),
7179 ;; since the parens around "gnu" would be superfluous if it's
7180 ;; a declarator. Shift the type one step backward.
7183 ;; Found no identifier.
7188 ;; CASE 3
7190 ;; Got a plain list of identifiers. If a colon follows it's
7191 ;; a valid label, or maybe a bitfield. Otherwise the last
7192 ;; one probably is the declared identifier and we should
7193 ;; back up to the previous type, providing it isn't a cast.
7197 ;; If we've found a specifier keyword then it's a
7198 ;; declaration regardless.
7209 ;; CASE 4
7213 ;; Got a plain list of identifiers followed by some suffix.
7214 ;; If this isn't a cast then the last identifier probably is
7215 ;; the declared one and we should back up to the previous
7216 ;; type.
7220 ;; CASE 5
7222 ;; If the type is known we know that there can't be any
7223 ;; identifier somewhere else, and it's only in declarations in
7224 ;; e.g. function prototypes and in casts that the identifier may
7225 ;; be left out.
7229 ;; Only got a single identifier (parsed as a type so far).
7230 ;; CASE 6
7232 ;; Check that the identifier isn't at the start of an
7233 ;; expression.
7234 at-decl-end
7237 ;; Inside an arglist that contains declarations. If K&R
7238 ;; style declarations and parenthesis style initializers
7239 ;; aren't allowed then the single identifier must be a
7240 ;; type, else we require that it's known or found
7241 ;; (primitive types are handled above).
7246 ;; Inside a template arglist. Accept known and found
7247 ;; types; other identifiers could just as well be
7248 ;; constants in C++.
7251 ;; CASE 7
7252 ;; Can't be a valid declaration or cast, but if we've found a
7253 ;; specifier it can't be anything else either, so treat it as
7254 ;; an invalid/unfinished declaration or cast.
7262 maybe-typeless
7263 backup-maybe-typeless
7267 c-after-suffixed-type-maybe-decl-key))))))
7268 ;; Got an empty paren pair and a preceding type that probably
7269 ;; really is the identifier. Shift the type backwards to make
7270 ;; the last one the identifier. This is analogous to the
7271 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7272 ;; above.
7273 ;;
7274 ;; Exception: In addition to the conditions in that
7275 ;; "backtracking" code, do not shift backward if we're not
7276 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7277 ;; Since there's no preceding type, the shift would mean that
7278 ;; the declaration is typeless. But if the regexp doesn't match
7279 ;; then we will simply fall through in the tests below and not
7280 ;; recognize it at all, so it's better to try it as an abstract
7281 ;; declarator instead.
7284 ;; Still no identifier.
7285 ;; CASE 8
7287 ;; Require `got-prefix' together with either `got-parens' or
7288 ;; `got-suffix' to recognize it as an abstract declarator:
7289 ;; `got-parens' only is probably an empty function call.
7290 ;; `got-suffix' only can build an ordinary expression together
7291 ;; with the preceding identifier which we've taken as a type.
7292 ;; We could actually accept on `got-prefix' only, but that can
7293 ;; easily occur temporarily while writing an expression so we
7294 ;; avoid that case anyway. We could do a better job if we knew
7295 ;; the point when the fontification was invoked.
7298 ;; CASE 9
7302 got-suffix-after-parens
7304 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7305 ;; normal function call after all (or perhaps a C++ style object
7306 ;; instantiation expression).
7309 ;; CASE 10
7311 ;; By now we've located the type in the declaration that we know
7312 ;; we're in.
7315 ;; CASE 11
7323 ;; Shift the type backward in the case that there's a
7324 ;; single identifier inside parens. That can only
7325 ;; occur in K&R style function declarations so it's
7326 ;; more likely that it really is a function call.
7327 ;; Therefore we only do this after
7328 ;; `c-after-suffixed-type-decl-key' has matched.
7330 got-suffix-after-parens))
7331 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7334 ;; CASE 12
7337 ;; It's a declaration if a known type precedes it and it can't be a
7338 ;; function call.
7341 ;; If we get here we can't tell if this is a type decl or a normal
7342 ;; expression by looking at it alone. (That's under the assumption
7343 ;; that normal expressions always can look like type decl expressions,
7344 ;; which isn't really true but the cases where it doesn't hold are so
7345 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7346 ;; the effort to look for them.)
7348 ;;; 2008-04-16: commented out the next form, to allow the function to recognize
7349 ;;; "foo (int bar)" in CC (an implicit type (in class foo) without a semicolon)
7350 ;;; as a(n almost complete) declaration, enabling it to be fontified.
7351 ;; CASE 13
7352 ;; (unless (or at-decl-end (looking-at "=[^=]"))
7353 ;; If this is a declaration it should end here or its initializer(*)
7354 ;; should start here, so check for allowed separation tokens. Note
7355 ;; that this rule doesn't work e.g. with a K&R arglist after a
7356 ;; function header.
7357 ;;
7358 ;; *) Don't check for C++ style initializers using parens
7359 ;; since those already have been matched as suffixes.
7360 ;;
7361 ;; If `at-decl-or-cast' is then we've found some other sign that
7362 ;; it's a declaration or cast, so then it's probably an
7363 ;; invalid/unfinished one.
7364 ;; (throw 'at-decl-or-cast at-decl-or-cast))
7366 ;; Below are tests that only should be applied when we're certain to
7367 ;; not have parsed halfway through an expression.
7369 ;; CASE 14
7371 ;; The expression starts with a known type so treat it as a
7372 ;; declaration.
7375 ;; CASE 15
7377 ;; In C++ we check if the identifier is a known type, since
7378 ;; (con|de)structors use the class name as identifier.
7379 ;; We've always shifted over the identifier as a type and
7380 ;; then backed up again in this case.
7381 identifier-type
7384 ;; `at-type' probably won't be 'found for
7385 ;; destructors since the "~" is then part of the
7386 ;; type name being checked against the list of
7387 ;; known types, so do a check without that
7388 ;; operator.
7393 c-identifier-syntax-table
7397 ;; We have already parsed the type earlier,
7398 ;; so it'd be possible to cache the end
7399 ;; position instead of redoing it here, but
7400 ;; then we'd need to keep track of another
7401 ;; position everywhere.
7409 ;; CASE 16
7411 at-type
7415 ;; Got something like "foo * bar;". Since we're not inside an
7416 ;; arglist it would be a meaningless expression because the
7417 ;; result isn't used. We therefore choose to recognize it as
7418 ;; a declaration. Do not allow a suffix since it could then
7419 ;; be a function call.
7422 ;; CASE 17
7427 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7428 ;; be an odd expression or it could be a declaration. Treat
7429 ;; it as a declaration if "a" has been used as a type
7430 ;; somewhere else (if it's a known type we won't get here).
7433 ;; CASE 18
7439 ;; Got a type followed by an abstract declarator. If `got-prefix'
7440 ;; is set it's something like "a *" without anything after it. If
7441 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7442 ;; or similar, which we accept only if the context rules out
7443 ;; expressions.
7446 ;; If we had a complete symbol table here (which rules out
7447 ;; `c-found-types') we should return t due to the disambiguation rule
7448 ;; (in at least C++) that anything that can be parsed as a declaration
7449 ;; is a declaration. Now we're being more defensive and prefer to
7450 ;; highlight things like "foo (bar);" as a declaration only if we're
7451 ;; inside an arglist that contains declarations.
7452 ;; CASE 19
7455 ;; The point is now after the type decl expression.
7458 ;; Check for a cast.
7461 c-cast-parens
7463 ;; Should be the first type/identifier in a cast paren.
7467 ;; The closing paren should follow.
7472 ;; There should be a primary expression after it.
7481 ;; Check if the expression begins with a prefix keyword.
7484 ;; Expression begins with an ambiguous operator. Treat
7485 ;; it as a cast if it's a type decl or if we've
7486 ;; recognized the type somewhere else.
7489 ;; Unless it's a keyword, it's the beginning of a primary
7490 ;; expression.
7492 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7493 ;; that it matched a whole one so that we don't e.g. confuse
7494 ;; the operator '-' with '->'. It's ok if it matches further,
7495 ;; though, since it e.g. can match the float '.5' while the
7496 ;; operator regexp only matches '.'.
7500 ;; There should either be a cast before it or something that isn't an
7501 ;; identifier or close paren.
7509 ;; It's a symbol. Accept it only if it's one of the
7510 ;; keywords that can precede an expression (without
7511 ;; surrounding parens).
7514 ;; Check that it isn't a close paren (block close is ok,
7515 ;; though).
7517 ;; Check that it isn't a nonsymbol identifier.
7520 ;; Handle the cast.
7530 ;; We're at a declaration. Highlight the type and the following
7531 ;; declarators.
7537 ;; Make sure to propagate the `c-decl-arg-start' property to
7538 ;; the next argument if it's set in this one, to cope with
7539 ;; interactive refontification.
7542 ;; Record the type's coordinates in `c-record-type-identifiers' for
7543 ;; later fontification.
7545 ;; There seems no reason to exclude a token from
7546 ;; fontification just because it's "a known type that can't
7547 ;; be a name or other expression". 2013-09-18.
7548 )
7559 ;; False alarm. Restore the recorded ranges.
7561 c-record-ref-identifiers save-rec-ref-ids)
7562 nil))))
7565 ;; Assuming that point is at the beginning of a token, check if it starts a
7566 ;; label and if so move over it and return non-nil (t in default situations,
7567 ;; specific symbols (see below) for interesting situations), otherwise don't
7568 ;; move and return nil. "Label" here means "most things with a colon".
7569 ;;
7570 ;; More precisely, a "label" is regarded as one of:
7571 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7572 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7573 ;; bare "case", should the colon be missing. We return t;
7574 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7575 ;; return t;
7576 ;; (iv) One of QT's "extended" C++ variants of
7577 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7578 ;; Returns the symbol `qt-2kwds-colon'.
7579 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7580 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7581 ;; colon). Currently (2006-03), this applies only to Objective C's
7582 ;; keywords "@private", "@protected", and "@public". Returns t.
7583 ;;
7584 ;; One of the things which will NOT be recognized as a label is a bit-field
7585 ;; element of a struct, something like "int foo:5".
7586 ;;
7587 ;; The end of the label is taken to be just after the colon, or the end of
7588 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7589 ;; after the end on return. The terminating char gets marked with
7590 ;; `c-decl-end' to improve recognition of the following declaration or
7591 ;; statement.
7592 ;;
7593 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7594 ;; label, if any, has already been marked up like that.
7595 ;;
7596 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7597 ;; after the preceding token, i.e. on the other side of the
7598 ;; syntactic ws from the point. Use a value less than or equal to
7599 ;; (point-min) if the point is at the first token in (the visible
7600 ;; part of) the buffer.
7601 ;;
7602 ;; The optional LIMIT limits the forward scan for the colon.
7603 ;;
7604 ;; This function records the ranges of the label symbols on
7605 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7606 ;; non-nil.
7607 ;;
7608 ;; This function might do hidden buffer changes.
7611 label-end
7612 qt-symbol-idx
7613 macro-start ; if we're in one.
7614 label-type
7615 kwd)
7617 ;; "case" or "default" (Doesn't apply to AWK).
7620 ;; Record only the keyword itself for fontification, since in
7621 ;; case labels the following is a constant expression and not
7622 ;; a label.
7626 ;; Find the label end.
7630 ;; Stop on chars that aren't allowed in expressions,
7631 ;; and on operator chars that would be meaningless
7632 ;; there. FIXME: This doesn't cope with ?: operators.
7640 t)
7642 ;; It's an unfinished label. We consider the keyword enough
7643 ;; to recognize it as a label, so that it gets fontified.
7644 ;; Leave the point at the end of it, but don't put any
7645 ;; `c-decl-end' marker.
7647 t))))
7649 ;; @private, @protected, @public, in Objective C, or similar.
7652 ;; For a `c-opt-extra-label-key' match, we record the whole
7653 ;; thing for fontification. That's to get the leading '@' in
7654 ;; Objective-C protection labels fontified.
7661 ;; All other cases of labels.
7664 ;; A colon label must have something before the colon.
7667 ;; Check that we're not after a token that can't precede a label.
7669 ;; Trivially succeeds when there's no preceding token.
7670 ;; Succeeds when we're at a virtual semicolon.
7679 ;; Check if we're after a label, if we're after a closing
7680 ;; paren that belong to statement, and with
7681 ;; `c-label-prefix-re'. It's done in different order
7682 ;; depending on `assume-markup' since the checks have
7683 ;; different expensiveness.
7710 ;; Might be after another label, so check it recursively.
7714 ;; Essentially the same as the
7715 ;; `c-syntactic-re-search-forward' regexp below.
7721 ;; Note: the following should work instead of the
7722 ;; narrow-to-region above. Investigate why not,
7723 ;; sometime. ACM, 2006-03-31.
7724 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7725 ;; macro-start t)
7727 ;; If the caller turned on recording for us,
7728 ;; it shouldn't apply when we check the
7729 ;; preceding label.
7730 c-record-type-identifiers)
7731 ;; A label can't start at a cpp directive. Check for
7732 ;; this, since c-forward-syntactic-ws would foul up on it.
7737 ;; Point is still at the beginning of the possible label construct.
7738 ;;
7739 ;; Check that the next nonsymbol token is ":", or that we're in one
7740 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7741 ;; arguments. FIXME: Should build this regexp from the language
7742 ;; constants.
7744 ;; public: protected: private:
7754 ;; QT double keyword like "protected slots:" or goto target.
7773 'qt-1kwd-colon
7785 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7795 t)
7798 ;; Record the identifiers in the label for fontification, unless
7799 ;; it begins with `c-label-kwds' in which case the following
7800 ;; identifiers are part of a (constant) expression that
7801 ;; shouldn't be fontified.
7813 ;; Not a label.
7815 label-type))
7818 ;; Assuming the point is at the beginning of a token, try to move
7819 ;; forward to the end of the Objective-C directive that starts
7820 ;; there. Return t if a directive was fully recognized, otherwise
7821 ;; the point is moved as far as one could be successfully parsed and
7822 ;; nil is returned.
7823 ;;
7824 ;; This function records identifier ranges on
7825 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7826 ;; `c-record-type-identifiers' is non-nil.
7827 ;;
7828 ;; This function might do hidden buffer changes.
7831 start-char
7833 lim
7834 ;; Turn off recognition of angle bracket arglists while parsing
7835 ;; types here since the protocol reference list might then be
7836 ;; considered part of the preceding name or superclass-name.
7847 t)
7856 ;; Handle the name of the class itself.
7858 ;; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7859 ;; at EOB.
7866 ;; Look for ": superclass-name" or "( category-name )".
7877 ;; Look for a protocol reference list.
7883 t))))
7889 t)
7892 nil)))
7895 ;; Go to the first non-whitespace after the colon that starts a
7896 ;; multiple inheritance introduction. Optional LIM is the farthest
7897 ;; back we should search.
7898 ;;
7899 ;; This function might do hidden buffer changes.
7907 ;; Return nil if we aren't in a method definition, otherwise the
7908 ;; position of the initial [+-].
7909 ;;
7910 ;; This function might do hidden buffer changes.
7916 ))
7918 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7920 ;; Return non-nil if point is within a gcc \"asm\" block.
7921 ;;
7922 ;; This should be called with point inside an argument list.
7923 ;;
7924 ;; Only one level of enclosing parentheses is considered, so for
7925 ;; instance `nil' is returned when in a function call within an asm
7926 ;; operand.
7927 ;;
7928 ;; This function might do hidden buffer changes.
7940 a function.
7942 More precisely, being at the top-level means that point is either
7943 outside any enclosing block (such as a function definition), or
7944 directly inside a class, namespace or other block that contains
7945 another declaration level.
7947 If point is not at the top-level (e.g. it is inside a method
7948 definition), then nil is returned. Otherwise, if point is at a
7949 top-level not enclosed within a class definition, t is returned.
7950 Otherwise, a 2-vector is returned where the zeroth element is the
7951 buffer position of the start of the class declaration, and the first
7952 element is the buffer position of the enclosing class's opening
7953 brace.
7955 Note that this function might do hidden buffer changes. See the
7956 comment at the start of cc-engine.el for more info."
7962 ;; Return non-nil if the point is in the region after the argument
7963 ;; list of a function and its opening brace (or semicolon in case it
7964 ;; got no body). If there are K&R style argument declarations in
7965 ;; that region, the point has to be inside the first one for this
7966 ;; function to recognize it.
7967 ;;
7968 ;; If successful, the point is moved to the first token after the
7969 ;; function header (see `c-forward-decl-or-cast-1' for details) and
7970 ;; the position of the opening paren of the function arglist is
7971 ;; returned.
7972 ;;
7973 ;; The point is clobbered if not successful.
7974 ;;
7975 ;; LIM is used as bound for backward buffer searches.
7976 ;;
7977 ;; This function might do hidden buffer changes.
7990 ;; There should not be a '=' or ',' between beg and the
7991 ;; start of the declaration since that means we were in the
7992 ;; "expression part" of the declaration.
7997 ;; Check that there's an arglist paren in the
7998 ;; declaration.
8001 ;; The declarator is a paren expression, so skip past it
8002 ;; so that we don't get stuck on that instead of the
8003 ;; function arglist.
8007 ;; Don't trip up on "operator ()".
8014 ;; Return the position of the first argument declaration if point is
8015 ;; inside a K&R style argument declaration list, nil otherwise.
8016 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
8017 ;; position that bounds the backward search for the argument list.
8018 ;;
8019 ;; Point must be within a possible K&R region, e.g. just before a top-level
8020 ;; "{". It must be outside of parens and brackets. The test can return
8021 ;; false positives otherwise.
8022 ;;
8023 ;; This function might do hidden buffer changes.
8027 ;; If we're in a macro, our search range is restricted to it. Narrow to
8028 ;; the searchable range.
8032 before-lparen after-rparen
8034 ; we give up
8037 ;; Search backwards for the defun's argument list. We give up if we
8038 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
8039 ;; a knr region) or BOB.
8040 ;;
8041 ;; The criterion for a paren structure being the arg list is:
8042 ;; o - there is non-WS stuff after it but before any "{"; AND
8043 ;; o - the token after it isn't a ";" AND
8044 ;; o - it is preceded by either an identifier (the function name) or
8045 ;; a macro expansion like "DEFUN (...)"; AND
8046 ;; o - its content is a non-empty comma-separated list of identifiers
8047 ;; (an empty arg list won't have a knr region).
8048 ;;
8049 ;; The following snippet illustrates these rules:
8050 ;; int foo (bar, baz, yuk)
8051 ;; int bar [] ;
8052 ;; int (*baz) (my_type) ;
8053 ;; int (*) (void) (*yuk) (void) ;
8054 ;; {
8065 ; parens/brackets.
8069 ;; We're inside a paren. Could it be our argument list....?
8075 ;; FIXME!!! What about macros between the parens? 2007/01/20
8078 ;; It can't be the arg list if next token is ; or {
8083 ;; Is the thing preceding the list an identifier (the
8084 ;; function name), or a macro expansion?
8094 ;; Have we got a non-empty list of comma-separated
8095 ;; identifiers?
8109 ;; ...Yes. We've identified the function's argument list.
8115 ;; ...No. The current parens aren't the function's arg list.
8122 ;; skip forward over conditional at point, including any predicate
8123 ;; statements in parentheses. No error checking is performed.
8124 ;;
8125 ;; This function might do hidden buffer changes.
8127 ;; else if()
8132 ;; do, else, try, finally
8137 ;; for, if, while, switch, catch, synchronized, foreach
8141 ;; If looking at the token after a conditional then return the
8142 ;; position of its start, otherwise return nil.
8143 ;;
8144 ;; This function might do hidden buffer changes.
8155 ;; If the point is after an operator identifier that isn't handled
8156 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
8157 ;; position of the start of that identifier is returned. nil is
8158 ;; returned otherwise. The point may be anywhere in the syntactic
8159 ;; whitespace after the last token of the operator identifier.
8160 ;;
8161 ;; This function might do hidden buffer changes.
8173 ;; Assuming point is at a brace that opens a statement block of some
8174 ;; kind, move to the proper anchor point for that block. It might
8175 ;; need to be adjusted further by c-add-stmt-syntax, but the
8176 ;; position at return is suitable as start position for that
8177 ;; function.
8178 ;;
8179 ;; This function might do hidden buffer changes.
8186 ;; Assuming point is at a brace that opens the block of a top level
8187 ;; declaration of some kind, move to the proper anchor point for
8188 ;; that block.
8189 ;;
8190 ;; This function might do hidden buffer changes.
8195 ;; Search forward for the end of the "header" of the current
8196 ;; declaration. That's the position where the definition body
8197 ;; starts, or the first variable initializer, or the ending
8198 ;; semicolon. I.e. search forward for the closest following
8199 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
8200 ;; _after_ the first found token, or at point-max if none is found.
8201 ;;
8202 ;; This function might do hidden buffer changes.
8207 ;; In C++ we need to take special care to handle operator
8208 ;; tokens and those pesky template brackets.
8213 ;; Handle operator identifiers, i.e. ignore any
8214 ;; operator token preceded by "operator".
8221 t
8223 nil)))))
8232 ;; Go to the beginning of the current declaration, or the beginning
8233 ;; of the previous one if already at the start of it. Point won't
8234 ;; be moved out of any surrounding paren. Return a cons cell of the
8235 ;; form (MOVE . KNR-POS). MOVE is like the return value from
8236 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
8237 ;; style argument declarations (and they are to be recognized) then
8238 ;; KNR-POS is set to the start of the first such argument
8239 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
8240 ;; position that bounds the backward search.
8241 ;;
8242 ;; NB: Cases where the declaration continues after the block, as in
8243 ;; "struct foo { ... } bar;", are currently recognized as two
8244 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
8245 ;;
8246 ;; This function might do hidden buffer changes.
8252 ;; `c-beginning-of-statement-1' stops at a block start, but we
8253 ;; want to continue if the block doesn't begin a top level
8254 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8255 ;; or an open paren.
8257 ;; Go back one "statement" each time round the loop until we're just
8258 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8259 ;; an ObjC method. This will move over a multiple declaration whose
8260 ;; components are comma separated.
8262 ;; Must check with c-opt-method-key in ObjC mode.
8272 ;; Check that we don't move from the first thing in a
8273 ;; macro to its header.
8279 move tentative-move))
8284 ;; Handle K&R argdecls. Back up after the "statement" jumped
8285 ;; over by `c-beginning-of-statement-1', unless it was the
8286 ;; function body, in which case we're sitting on the opening
8287 ;; brace now. Then test if we're in a K&R argdecl region and
8288 ;; that we started at the other side of the first argdecl in
8289 ;; it.
8299 'previous
8301 knr-argdecl-start))
8304 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8305 ;; statement, so the result will be 'previous if we've moved over any.
8306 ;; So change our result back to 'same if necessary.
8307 ;;
8308 ;; If they were brace list initializers we might not have moved over a
8309 ;; declaration boundary though, so change it to 'same if we've moved
8310 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8311 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8312 ;; potentially can search over a large amount of text.). Take special
8313 ;; pains not to get mislead by C++'s "operator=", and the like.
8316 c++-template-syntax-table
8322 ; no more, or get to one which isn't an "operator ="
8325 c-overloadable-operators-regexp
8326 c-opt-op-identifier-prefix
8341 ;; Assuming point is at the start of a declaration (as detected by
8342 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8343 ;; `c-beginning-of-decl-1', this function handles the case when a
8344 ;; block is followed by identifiers in e.g. struct declarations in C
8345 ;; or C++. If a proper end was found then t is returned, otherwise
8346 ;; point is moved as far as possible within the current sexp and nil
8347 ;; is returned. This function doesn't handle macros; use
8348 ;; `c-end-of-macro' instead in those cases.
8349 ;;
8350 ;; This function might do hidden buffer changes.
8353 c++-template-syntax-table
8361 ;; Stopped at the ';' in a K&R argdecl section which is
8362 ;; detected using the same criteria as in
8363 ;; `c-beginning-of-decl-1'. Move to the following block
8364 ;; start.
8368 ;; Encountered a block in the declaration. Jump over it.
8379 ;; Check for `c-opt-block-decls-with-vars-key'
8380 ;; before the first paren.
8383 c-opt-block-decls-with-vars-key
8385 lim t t t)
8388 ;; Check that the first following paren is
8389 ;; the block.
8391 lim t t t)
8393 ;; The declaration doesn't have any of the
8394 ;; `c-opt-block-decls-with-vars' keywords in the
8395 ;; beginning, so it ends here at the end of the block.
8403 nil)))
8406 ;; Assuming the point is at an open brace, check if it starts a
8407 ;; block that contains another declaration level, i.e. that isn't a
8408 ;; statement block or a brace list, and if so return non-nil.
8409 ;;
8410 ;; If the check is successful, the return value is the start of the
8411 ;; keyword that tells what kind of construct it is, i.e. typically
8412 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8413 ;; the point will be at the start of the construct, before any
8414 ;; leading specifiers, otherwise it's at the returned position.
8415 ;;
8416 ;; The point is clobbered if the check is unsuccessful.
8417 ;;
8418 ;; CONTAINING-SEXP is the position of the open of the surrounding
8419 ;; paren, or nil if none.
8420 ;;
8421 ;; The optional LIMIT limits the backward search for the start of
8422 ;; the construct. It's assumed to be at a syntactically relevant
8423 ;; position.
8424 ;;
8425 ;; If any template arglists are found in the searched region before
8426 ;; the open brace, they get marked with paren syntax.
8427 ;;
8428 ;; This function might do hidden buffer changes.
8435 ;; Could be at the end of a template arglist.
8446 ;; Note: Can't get bogus hits inside template arglists below since they
8447 ;; have gotten paren syntax above.
8449 ;; If `goto-start' is set we begin by searching for the
8450 ;; first possible position of a leading specifier list.
8451 ;; The `c-decl-block-key' search continues from there since
8452 ;; we know it can't match earlier.
8455 open-brace t t)
8457 t)
8458 t)
8464 ;; Exclude cases where we matched what would ordinarily
8465 ;; be a block declaration keyword, except where it's not
8466 ;; legal because it's part of a "compound keyword" like
8467 ;; "enum class". Of course, if c-after-brace-list-key
8468 ;; is nil, we can skip the test.
8478 ;; Found a keyword that can't be a type?
8481 ;; Can be a type too, in which case it's the return type of a
8482 ;; function (under the assumption that no declaration level
8483 ;; block construct starts with a type).
8486 ;; Jumped over a type, but it could be a declaration keyword
8487 ;; followed by the declared identifier that we've jumped over
8488 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8489 ;; then we should be at the declarator now, so check for a
8490 ;; valid declarator start.
8491 ;;
8492 ;; Note: This doesn't cope with the case when a declared
8493 ;; identifier is followed by e.g. '(' in a language where '('
8494 ;; also might be part of a declarator expression. Currently
8495 ;; there's no such language.
8499 ;; In Pike a list of modifiers may be followed by a brace
8500 ;; to make them apply to many identifiers. Note that the
8501 ;; match data will be empty on return in this case.
8507 ;; Use this variant to avoid yet another special regexp.
8511 t)))
8513 ;; Got a match.
8516 ;; Back up over any preceding specifiers and their clauses
8517 ;; by going forward from `first-specifier-pos', which is the
8518 ;; earliest possible position where the specifier list can
8519 ;; start.
8525 ;; Accept any plain symbol token on the ground that
8526 ;; it's a specifier masked through a macro (just
8527 ;; like `c-forward-decl-or-cast-1' skip forward over
8528 ;; such tokens).
8529 ;;
8530 ;; Could be more restrictive wrt invalid keywords,
8531 ;; but that'd only occur in invalid code so there's
8532 ;; no use spending effort on it.
8538 ;; Can't parse a declaration preamble and is still
8539 ;; before `kwd-start'. That means `first-specifier-pos'
8540 ;; was in some earlier construct. Search again.
8544 ;; Got no preamble before the block declaration keyword.
8550 kwd-start)))
8553 ;; Check if the closest containing paren sexp is a declaration
8554 ;; block, returning a 2 element vector in that case. Aref 0
8555 ;; contains the bufpos at boi of the class key line, and aref 1
8556 ;; contains the bufpos of the open brace. This function is an
8557 ;; obsolete wrapper for `c-looking-at-decl-block'.
8558 ;;
8559 ;; This function might do hidden buffer changes.
8567 nil))
8572 ;; Return the buffer position of the most enclosing decl-block brace (in the
8573 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8574 ;; none was found.
8583 open-brace))
8586 ;; Return the position of the L-brace if point is inside a brace list
8587 ;; initialization of an array, etc. This is an approximate function,
8588 ;; designed for speed over accuracy. It will not find every bracelist, but
8589 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8590 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8591 ;; is everywhere else.
8600 b-pos)))
8603 ;; We're at the token after what might be a type prefixed with a colon. Try
8604 ;; moving backward over this type and the colon. On success, return t and
8605 ;; leave point before colon; on failure, leave point unchanged. Will clobber
8606 ;; match data.
8623 nil)
8625 t)
8627 t)
8631 t)))
8634 ;; We're at a "{". Move back to the enum-like keyword that starts this
8635 ;; declaration and return t, otherwise don't move and return nil.
8637 up-sexp-pos before-identifier)
8653 t)
8659 t)
8665 ;; return the buffer position of the beginning of the brace list
8666 ;; statement if we're inside a brace list, otherwise return nil.
8667 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8668 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8669 ;; braces
8670 ;;
8671 ;; N.B.: This algorithm can potentially get confused by cpp macros
8672 ;; placed in inconvenient locations. It's a trade-off we make for
8673 ;; speed.
8674 ;;
8675 ;; This function might do hidden buffer changes.
8677 ;; This will pick up brace list declarations.
8681 ;; this will pick up array/aggregate init lists, even if they are nested.
8684 ;; Pike can have class definitions anywhere, so we must
8685 ;; check for the class key here.
8687 c-decl-block-key))
8688 bufpos braceassignp lim next-containing macro-start)
8690 containing-sexp)
8701 ;; We're in an in-expression block of some kind. Do not
8702 ;; check nesting. We deliberately set the limit to the
8703 ;; containing sexp, so that c-looking-at-inexpr-block
8704 ;; doesn't check for an identifier before it.
8706 ;; see if the open brace is preceded by = or [...] in
8707 ;; this statement, but watch out for operator=
8710 ;; Checks to do only on the first sexp before the brace.
8713 ;; In Java, an initialization brace list may follow
8714 ;; directly after "new Foo[]", so check for a "new"
8715 ;; earlier.
8721 ;; Carry on looking if this is an
8722 ;; identifier (may contain "." in Java)
8723 ;; or another "[]" sexp.
8726 ;; Checks to do on all sexps before the brace, up to the
8727 ;; beginning of the statement.
8735 ;; We've seen a =, but must check earlier tokens so
8736 ;; that it isn't something that should be ignored.
8742 ;; Check for operator =
8745 nil)
8746 ;; Check for `<opchar>= in Pike.
8749 ;; Special case for Pikes
8750 ;; `[]=, since '[' is not in
8751 ;; the punctuation class.
8754 nil)
8756 ;; make sure we're not in a C++ template
8757 ;; argument assignment
8769 ))))
8770 nil)
8777 ;; We've hit the beginning of the aggregate list.
8782 ;; Brace lists can't contain a semicolon, so we're done.
8787 ;; We've a macro whose expansion starts with the '{'.
8788 ;; Heuristically, if we have a ';' in it we've not got a
8789 ;; brace list, otherwise we have.
8796 containing-sexp nil)
8799 ;; Go up one level
8801 lim nil
8802 next-containing nil)))))
8804 bufpos))
8805 ))
8808 ;; If we're looking at the start of a pike-style list, i.e., `({Â })',
8809 ;; `([Â ])', `(<Â >)', etc., a cons of a cons of its starting and ending
8810 ;; positions and its entry in c-special-brace-lists is returned, nil
8811 ;; otherwise. The ending position is nil if the list is still open.
8812 ;; LIM is the limit for forward search. The point may either be at
8813 ;; the `(' or at the following paren character. Tries to check the
8814 ;; matching closer, but assumes it's correct if no balanced paren is
8815 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8816 ;; a special brace list).
8817 ;;
8818 ;; This function might do hidden buffer changes.
8823 inner-beg end type)
8838 nil)))))
8848 ;; Check balancing of the inner paren
8849 ;; below.
8852 t)
8853 ;; If the inner char isn't a paren then
8854 ;; we can't check balancing, so just
8855 ;; check the char before the outer
8856 ;; closing paren.
8871 ;; Return non-nil if between two statements or declarations, assuming
8872 ;; point is not inside a literal or comment.
8873 ;;
8874 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8875 ;; are recommended instead.
8876 ;;
8877 ;; This function might do hidden buffer changes.
8882 ;; Return non-nil if we're looking at the beginning of a block
8883 ;; inside an expression. The value returned is actually a cons of
8884 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8885 ;; position of the beginning of the construct.
8886 ;;
8887 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8888 ;; position of the closest containing list. If it's nil, the
8889 ;; containing paren isn't used to decide whether we're inside an
8890 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8891 ;; needs to be farther back.
8892 ;;
8893 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8894 ;; brace block might be done. It should only be used when the
8895 ;; construct can be assumed to be complete, i.e. when the original
8896 ;; starting position was further down than that.
8897 ;;
8898 ;; This function might do hidden buffer changes.
8903 ;; Look at the character after point only as a last resort
8904 ;; when we can't disambiguate.
8924 ;; If the class definition is at the start of
8925 ;; a statement, we don't consider it an
8926 ;; in-expression class.
8934 ;; Also, in Pike we treat it as an
8935 ;; in-expression class if it's used in an
8936 ;; object clone expression.
8952 nil)
8960 ;; Accept several square bracket sexps for
8961 ;; Java array initializations.
8969 block-follows
8970 containing-sexp
8979 nil
8982 res))))
8985 ;; Returns non-nil if we're looking at the end of an in-expression
8986 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
8987 ;; PAREN-STATE is the paren state relevant at the current position.
8988 ;;
8989 ;; This function might do hidden buffer changes.
8991 ;; We currently only recognize a block.
8994 containing-sexp)
9003 paren-state)
9004 containing-sexp)))))
9007 ;; Is there a "virtual semicolon" at POS or point?
9008 ;; (See cc-defs.el for full details of "virtual semicolons".)
9009 ;;
9010 ;; This is true when point is at the last non syntactic WS position on the
9011 ;; line, there is a macro call last on the line, and this particular macro's
9012 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
9013 ;; semicolon.
9021 c-macro-with-semi-re
9024 ;; Check we've got nothing after this except comments and empty lines
9025 ;; joined by escaped EOLs.
9038 nil)
9046 ;; Check for one of the listed macros being before point.
9050 t))
9058 \f
9059 ;; `c-guess-basic-syntax' and the functions that precedes it below
9060 ;; implements the main decision tree for determining the syntactic
9061 ;; analysis of the current line of code.
9063 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
9064 ;; auto newline analysis.
9068 ;; BRACEPOS is the position of a brace in a construct like "namespace
9069 ;; Bar {". Return the anchor point in this construct; this is the
9070 ;; earliest symbol on the brace's line which isn't earlier than
9071 ;; "namespace".
9072 ;;
9073 ;; Currently (2007-08-17), "like namespace" means "matches
9074 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
9075 ;; or anything like that.
9085 ;; A simple function to prepend a new syntax element to
9086 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
9087 ;; should always be dynamically bound but since we read it first
9088 ;; we'll fail properly anyway if this function is misused.
9090 c-syntactic-context)))
9093 ;; Like `c-add-syntax' but appends to the end of the syntax list.
9094 ;; (Normally not necessary.)
9099 syntax-extra-args
9100 stop-at-boi-only
9101 containing-sexp
9102 paren-state)
9103 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
9104 ;; needed with further syntax elements of the types `substatement',
9105 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
9106 ;; `defun-block-intro'.
9107 ;;
9108 ;; Do the generic processing to anchor the given syntax symbol on
9109 ;; the preceding statement: Skip over any labels and containing
9110 ;; statements on the same line, and then search backward until we
9111 ;; find a statement or block start that begins at boi without a
9112 ;; label or comment.
9113 ;;
9114 ;; Point is assumed to be at the prospective anchor point for the
9115 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
9116 ;; skip past open parens and containing statements. Most of the added
9117 ;; syntax elements will get the same anchor point - the exception is
9118 ;; for an anchor in a construct like "namespace"[*] - this is as early
9119 ;; as possible in the construct but on the same line as the {.
9120 ;;
9121 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
9122 ;;
9123 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
9124 ;; syntax symbol. They are appended after the anchor point.
9125 ;;
9126 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
9127 ;; if the current statement starts there.
9128 ;;
9129 ;; Note: It's not a problem if PAREN-STATE "overshoots"
9130 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
9131 ;;
9132 ;; This function might do hidden buffer changes.
9135 ;; This is by far the most common case, so let's give it special
9136 ;; treatment.
9141 ;; Set when we're on a label, so that we don't stop there.
9142 ;; FIXME: To be complete we should check if we're on a label
9143 ;; now at the start.
9144 on-label)
9146 ;; Use point as the anchor point for "namespace", "extern", etc.
9150 syntax-extra-args)
9152 ;; Loop while we have to back out of containing blocks.
9157 ;; Loop while we have to back up statements.
9159 on-label
9162 ;; Skip past any comments that stands between the
9163 ;; statement start and boi.
9171 ;; Skip to the beginning of this statement or backward
9172 ;; another one.
9180 ;; If we didn't move we're at the start of a block and
9181 ;; have to continue outside it.
9190 ;; Special case to avoid deeper and deeper indentation
9191 ;; of "else if" clauses.
9192 )
9197 ;; If stop-at-boi-only is nil, we shouldn't back up
9198 ;; over previous or containing statements to try to
9199 ;; reach boi, so go back to the last position and
9200 ;; exit.
9207 ;; If we've moved into another statement then we
9208 ;; should no longer try to stop in the middle of a
9209 ;; line.
9212 ;; Record this as a substatement if we skipped up one
9213 ;; level.
9216 )))
9218 containing-sexp)
9220 ;; Now we have to go out of this block.
9223 ;; Don't stop in the middle of a special brace list opener
9224 ;; like "({".
9236 ;; Analyze the construct in front of the block we've stepped out
9237 ;; from and add the right syntactic element for it.
9240 step-type)
9243 ;; Stepped out of a parenthesis block, so we're in an
9244 ;; expression now.
9256 ;; Stepped out of an in-expression statement. This
9257 ;; syntactic element won't get an anchor pos.
9260 ;; A parenthesis normally belongs to an arglist.
9268 on-label nil))
9270 ;; Stepped out of a brace block.
9282 containing-sexp)))
9284 'defun-block-intro
9286 nil))
9290 c-other-decl-block-key-in-symbols-alist))
9297 ;; Always done if the open brace was at boi. The
9298 ;; c-beginning-of-statement-1 call above is necessary
9299 ;; anyway, to decide the type of block-intro to add.
9302 ))
9304 ;; Fill in the current point as the anchor for all the symbols
9305 ;; added above.
9314 )))
9317 containing-decl-open
9318 containing-decl-start
9319 containing-decl-kwd
9320 paren-state)
9321 ;; The inclass and class-close syntactic symbols are added in
9322 ;; several places and some work is needed to fix everything.
9323 ;; Therefore it's collected here.
9324 ;;
9325 ;; This function might do hidden buffer changes.
9330 containing-decl-open)
9332 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9333 ;; here, but we have to do like this for compatibility.
9343 char-after-ip
9344 beg-of-same-or-containing-stmt
9345 containing-sexp
9346 paren-state)
9347 ;; This function contains the decision tree reached through both
9348 ;; cases 18 and 10. It's a continued statement or top level
9349 ;; construct of some kind.
9350 ;;
9351 ;; This function might do hidden buffer changes.
9358 ;; (CASE A removed.)
9359 ;; CASE B: open braces for class or brace-lists
9366 ;; CASE B.1: class-open
9373 ;; CASE B.2: brace-list-open
9378 indent-point t t t)))
9379 ;; The most semantically accurate symbol here is
9380 ;; brace-list-open, but we normally report it simply as a
9381 ;; statement-cont. The reason is that one normally adjusts
9382 ;; brace-list-open for brace lists as top-level constructs,
9383 ;; and brace lists inside statements is a completely different
9384 ;; context. C.f. case 5A.3.
9387 ;; Turn off the dwim above when we're
9388 ;; analyzing the nature of the brace
9389 ;; for the auto newline feature.
9390 'brace-list-open
9392 nil nil
9393 containing-sexp paren-state))
9395 ;; CASE B.3: The body of a function declared inside a normal
9396 ;; block. Can occur e.g. in Pike and when using gcc
9397 ;; extensions, but watch out for macros followed by blocks.
9398 ;; C.f. cases E, 16F and 17G.
9404 ;; Turn off recognition of constructs that lacks a
9405 ;; type in this case, since that's more likely to be
9406 ;; a macro followed by a block.
9409 containing-sexp paren-state))
9411 ;; CASE B.4: Continued statement with block open. The most
9412 ;; accurate analysis is perhaps `statement-cont' together with
9413 ;; `block-open' but we play DWIM and use `substatement-open'
9414 ;; instead. The rationale is that this typically is a macro
9415 ;; followed by a block which makes it very similar to a
9416 ;; statement with a substatement block.
9419 containing-sexp paren-state))
9420 ))
9422 ;; CASE C: iostream insertion or extraction operator
9426 ;; If there is no preceding streamop in the statement
9427 ;; then indent this line as a normal statement-cont.
9431 t))))
9433 ;; CASE E: In the "K&R region" of a function declared inside a
9434 ;; normal block. C.f. case B.3.
9436 ;; Check that the next token is a '{'. This works as
9437 ;; long as no language that allows nested function
9438 ;; definitions allows stuff like member init lists, K&R
9439 ;; declarations or throws clauses there.
9440 ;;
9441 ;; Note that we do a forward search for something ahead
9442 ;; of the indentation line here. That's not good since
9443 ;; the user might not have typed it yet. Unfortunately
9444 ;; it's exceedingly tricky to recognize a function
9445 ;; prototype in a code block without resorting to this.
9453 ;; Turn off recognition of constructs that lacks a
9454 ;; type in this case, since that's more likely to be
9455 ;; a macro followed by a block.
9458 containing-sexp paren-state))
9460 ;;CASE F: continued statement and the only preceding items are
9461 ;;annotations.
9469 t)
9476 ;; CASE G: a template list continuation?
9477 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9495 ;; In a nested template arglist.
9501 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9502 ;; template aware.
9505 ;; CASE D: continued statement.
9509 containing-sexp paren-state))
9510 )))
9512 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9513 ;; 2005/11/29).
9514 ;;;###autoload
9516 "Return the syntactic context of the current line."
9522 open-paren-in-column-0-is-defun-start
9523 ;; A whole ugly bunch of various temporary variables. Have
9524 ;; to declare them here since it's not possible to declare
9525 ;; a variable with only the scope of a cond test and the
9526 ;; following result clauses, and most of this function is a
9527 ;; single gigantic cond. :P
9528 literal char-before-ip before-ws-ip char-after-ip macro-start
9529 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9530 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9531 containing-<
9532 ;; The following record some positions for the containing
9533 ;; declaration block if we're directly within one:
9534 ;; `containing-decl-open' is the position of the open
9535 ;; brace. `containing-decl-start' is the start of the
9536 ;; declaration. `containing-decl-kwd' is the keyword
9537 ;; symbol of the keyword that tells what kind of block it
9538 ;; is.
9539 containing-decl-open
9540 containing-decl-start
9541 containing-decl-kwd
9542 ;; The open paren of the closest surrounding sexp or nil if
9543 ;; there is none.
9544 containing-sexp
9545 ;; The position after the closest preceding brace sexp
9546 ;; (nested sexps are ignored), or the position after
9547 ;; `containing-sexp' if there is none, or (point-min) if
9548 ;; `containing-sexp' is nil.
9549 lim
9550 ;; The paren state outside `containing-sexp', or at
9551 ;; `indent-point' if `containing-sexp' is nil.
9553 ;; There's always at most one syntactic element which got
9554 ;; an anchor pos. It's stored in syntactic-relpos.
9555 syntactic-relpos
9558 ;; Check if we're directly inside an enclosing declaration
9559 ;; level block.
9568 containing-sexp)
9569 t)))
9572 containing-sexp nil)
9577 ;; Init some position variables.
9586 ;; Ignore balanced paren. The next entry
9587 ;; can't be another one.
9590 ;; If there is no surrounding open paren then
9591 ;; put the last balanced pair back on paren-state.
9593 containing-sexp nil)))
9597 ;; If we're in a parenthesis list then ',' delimits the
9598 ;; "statements" rather than being an operator (with the
9599 ;; exception of the "for" clause). This difference is
9600 ;; typically only noticeable when statements are used in macro
9601 ;; arglists.
9605 ;; cache char before and after indent point, and move point to
9606 ;; the most likely position to perform the majority of tests
9615 ;; are we in a literal?
9618 ;; now figure out syntactic qualities of the current line
9621 ;; CASE 1: in a string.
9625 ;; CASE 2: in a C or C++ style comment.
9627 ;; This is a kludge for XEmacs where we use
9628 ;; `buffer-syntactic-context', which doesn't correctly
9629 ;; recognize "\*/" to end a block comment.
9630 ;; `parse-partial-sexp' which is used by
9631 ;; `c-literal-limits' will however do that in most
9632 ;; versions, which results in that we get nil from
9633 ;; `c-literal-limits' even when `c-in-literal' claims
9634 ;; we're inside a comment.
9638 ;; CASE 3: in a cpp preprocessor macro continuation.
9648 ;; If at the beginning of the body of a #define
9649 ;; directive then analyze as cpp-define-intro
9650 ;; only. Go on with the syntactic analysis
9651 ;; otherwise. in-macro-expr is set if we're in a
9652 ;; cpp expression, i.e. before the #define body
9653 ;; or anywhere in a non-#define directive.
9660 nil)))))
9664 ;; CASE 11: an else clause?
9668 containing-sexp paren-state))
9670 ;; CASE 12: while closure of a do/while construct?
9678 containing-sexp paren-state))
9680 ;; CASE 13: A catch or finally clause? This case is simpler
9681 ;; than if-else and do-while, because a block is required
9682 ;; after every try, catch and finally.
9690 t)
9694 t)
9697 t))
9702 containing-sexp paren-state))
9704 ;; CASE 18: A substatement we can recognize by keyword.
9714 ;; Ought to cache the result from the
9715 ;; c-beginning-of-statement-1 calls here.
9725 ;; Step up to the containing statement if we
9726 ;; stayed in the same one.
9734 ;; There was no containing statement after all.
9736 placeholder))
9738 ;; Require a parenthesis after these keywords.
9739 ;; Necessary to catch e.g. synchronized in Java,
9740 ;; which can be used both as statement and
9741 ;; modifier.
9747 ;; CASE 18A: Simple substatement.
9753 containing-sexp paren-state))
9759 containing-sexp paren-state))
9762 containing-sexp paren-state))))
9764 ;; CASE 18B: Some other substatement. This is shared
9765 ;; with case 10.
9767 char-after-ip
9768 placeholder
9769 lim
9770 paren-state)))
9772 ;; CASE 14: A case or default label
9778 containing-sexp))
9781 ;; Got a bogus label at the top level. In lack of better
9782 ;; alternatives, anchor it on (point-min).
9785 ;; CASE 15: any other label
9792 containing-decl-open
9793 containing-decl-start
9794 containing-decl-kwd
9795 paren-state))
9796 ;; Append access-label with the same anchor point as
9797 ;; inclass gets.
9803 containing-sexp))
9808 ;; If the surrounding statement is a switch then
9809 ;; let's analyze all labels as switch labels, so
9810 ;; that they get lined up consistently.
9811 'case-label
9817 ;; A label on the top level. Treat it as a class
9818 ;; context. (point-min) is the closest we get to the
9819 ;; class open brace.
9822 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9823 ;; 17E.
9826 containing-sexp
9827 ;; Have to turn on the heuristics after
9828 ;; the point even though it doesn't work
9829 ;; very well. C.f. test case class-16.pike.
9830 t))
9835 ;; It's a statement block or an anonymous class.
9837 ;; It's a Pike lambda. Check whether we are between the
9838 ;; lambda keyword and the argument list or at the defun
9839 ;; opener.
9841 'inline-open
9847 paren-state)
9851 ;; CASE 5: Line is inside a declaration level block or at top level.
9855 ;; CASE 5A: we are looking at a defun, brace list, class,
9856 ;; or inline-inclass method opening brace
9863 ;; CASE 5A.1: Non-class declaration block open.
9878 ;; Special case for extern-lang-open.
9879 'extern-lang-open
9881 nil t containing-sexp paren-state))
9883 ;; CASE 5A.2: we are looking at a class opening brace
9892 ;; CASE 5A.3: brace list open
9917 ))))
9921 ;; We're in Java and have found that the open brace
9922 ;; belongs to a "new Foo[]" initialization list,
9923 ;; which means the brace list is part of an
9924 ;; expression and not a top level definition. We
9925 ;; therefore treat it as any topmost continuation
9926 ;; even though the semantically correct symbol still
9927 ;; is brace-list-open, on the same grounds as in
9928 ;; case B.2.
9934 ;; CASE 5A.4: inline defun open
9940 containing-decl-open
9941 containing-decl-start
9942 containing-decl-kwd
9943 paren-state))
9945 ;; CASE 5A.5: ordinary defun open
9953 ;; Bogus to use bol here, but it's the legacy. (Resolved,
9954 ;; 2007-11-09)
9955 ))))
9957 ;; CASE 5R: Member init list. (Used to be part of CASE 5B.1)
9958 ;; Note there is no limit on the backward search here, since member
9959 ;; init lists can, in practice, be very large.
9965 ;; There is no preceding member init clause.
9966 ;; Indent relative to the beginning of indentation
9967 ;; for the topmost-intro line that contains the
9968 ;; prototype's open paren.
9971 ;; Indent relative to the first member init clause.
9976 ;; CASE 5B: After a function header but before the body (or
9977 ;; the ending semicolon if there's no body).
9984 ;; CASE 5B.1: Member init list.
9986 ;; There is no preceding member init clause.
9987 ;; Indent relative to the beginning of indentation
9988 ;; for the topmost-intro line that contains the
9989 ;; prototype's open paren.
9993 ;; CASE 5B.2: K&R arg decl intro
10000 containing-decl-open
10001 containing-decl-start
10002 containing-decl-kwd
10003 paren-state)))
10005 ;; CASE 5B.4: Nether region after a C++ or Java func
10006 ;; decl, which could include a `throws' declaration.
10010 )))
10012 ;; CASE 5C: inheritance line. could be first inheritance
10013 ;; line, or continuation of a multiple inheritance
10021 ;; watch out for scope operator
10026 )))
10035 ;; for Java
10040 cont done)
10045 done t))
10049 )
10051 injava-inher)
10054 ))
10057 ;; CASE 5C.1: non-hanging colon on an inher intro
10061 ;; don't add inclass symbol since relative point already
10062 ;; contains any class offset
10063 )
10065 ;; CASE 5C.2: hanging colon on an inher intro
10071 containing-decl-open
10072 containing-decl-start
10073 containing-decl-kwd
10074 paren-state)))
10076 ;; CASE 5C.3: in a Java implements/extends
10090 )))
10092 ;; CASE 5C.4: a continued inheritance line
10096 ;; don't add inclass symbol since relative point already
10097 ;; contains any class offset
10098 )))
10100 ;; CASE 5P: AWK pattern or function or continuation
10101 ;; thereof.
10107 'topmost-intro-cont
10109 nil nil
10110 containing-sexp paren-state))
10112 ;; CASE 5D: this could be a top-level initialization, a
10113 ;; member init list continuation, or a template argument
10114 ;; list continuation.
10116 ;; Note: We use the fact that lim is always after any
10117 ;; preceding brace sexp.
10127 t))
10132 t)
10138 nil)
10139 t))
10141 ;; NB: No c-after-special-operator-id stuff in this
10142 ;; clause - we assume only C++ needs it.
10147 ;; CASE 5D.3: perhaps a template list continuation?
10162 ;; In a nested template arglist.
10168 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
10169 ;; template aware.
10172 ;; CASE 5D.4: perhaps a multiple inheritance line?
10184 t)
10189 ;; CASE 5D.5: Continuation of the "expression part" of a
10190 ;; top level construct. Or, perhaps, an unrecognized construct.
10202 ;; A preceding comma at the top level means that a
10203 ;; new variable declaration starts here. Use
10204 ;; topmost-intro-cont for it, for consistency with
10205 ;; the first variable declaration. C.f. case 5N.
10208 nil nil containing-sexp paren-state))
10209 ))
10211 ;; CASE 5F: Close of a non-class declaration level block.
10215 ;; This is inconsistent: Should use `containing-decl-open'
10216 ;; here if it's at boi, like in case 5J.
10220 ;; Special case for compatibility with the
10221 ;; extern-lang syntactic symbols.
10222 'extern-lang-close
10225 nil t
10227 paren-state))
10229 ;; CASE 5G: we are looking at the brace which closes the
10230 ;; enclosing nested class decl
10235 containing-decl-open
10236 containing-decl-start
10237 containing-decl-kwd
10238 paren-state))
10240 ;; CASE 5H: we could be looking at subsequent knr-argdecls
10247 ;; Do an extra check to avoid tripping up on
10248 ;; statements that occur in invalid contexts
10249 ;; (e.g. in macro bodies where we don't really
10250 ;; know the context of what we're looking at).
10257 ;; CASE 5I: ObjC method definition.
10262 ;; Handle the case when it's the first (non-comment)
10263 ;; thing in the buffer. Can't look for a 'same return
10264 ;; value from cbos1 since ObjC directives currently
10265 ;; aren't recognized fully, so that we get 'same
10266 ;; instead of 'previous if it moved over a preceding
10267 ;; directive.
10271 ;; CASE 5N: At a variable declaration that follows a class
10272 ;; definition or some other block declaration that doesn't
10273 ;; end at the closing '}'. C.f. case 5D.5.
10282 ;; Speed up the backward search a bit.
10287 ;; The '}' is unbalanced.
10288 nil
10293 containing-sexp paren-state))
10295 ;; NOTE: The point is at the end of the previous token here.
10297 ;; CASE 5J: we are at the topmost level, make
10298 ;; sure we skip back past any access specifiers
10300 ;; A macro continuation line is never at top level.
10322 nil)))))
10323 ;; For historic reasons we anchor at bol of the last
10324 ;; line of the previous declaration. That's clearly
10325 ;; highly bogus and useless, and it makes our lives hard
10326 ;; to remain compatible. :P
10337 ;; Special case for compatibility with the
10338 ;; extern-lang syntactic symbols.
10339 'inextern-lang
10342 nil t
10344 paren-state))
10346 containing-decl-open
10347 containing-decl-start
10348 containing-decl-kwd
10349 paren-state)))
10351 macro-start
10356 ;; CASE 5K: we are at an ObjC method definition
10357 ;; continuation line.
10366 ;; CASE 5L: we are at the first argument of a template
10367 ;; arglist that begins on the previous line.
10375 ;; CASE 5Q: we are at a statement within a macro.
10380 ;;CASE 5N: We are at a topmost continuation line and the only
10381 ;;preceding items are annotations.
10388 t)
10394 ;; CASE 5M: we are at a topmost continuation line
10405 ))
10407 ;; (CASE 6 has been removed.)
10409 ;; CASE 7: line is an expression, not a statement. Most
10410 ;; likely we are either in a function prototype or a function
10411 ;; call argument list
10419 ;; CASE 7A: we are looking at the arglist closing paren.
10420 ;; C.f. case 7F.
10432 paren-state))
10434 ;; CASE 7B: Looking at the opening brace of an
10435 ;; in-expression block or brace list. C.f. cases 4, 16A
10436 ;; and 17E.
10440 paren-state))
10444 placeholder
10447 paren-state)
10448 containing-sexp)))
10449 ;; placeholder is nil if it's a block directly in
10450 ;; a function arglist. That makes us skip out of
10451 ;; this case.
10452 )))
10457 paren-state)
10461 ;; CASE 7C: we are looking at the first argument in an empty
10462 ;; argument list. Use arglist-close if we're actually
10463 ;; looking at a close paren or bracket.
10475 paren-state))
10477 ;; CASE 7D: we are inside a conditional test clause. treat
10478 ;; these things as statements
10488 ))
10490 ;; CASE 7E: maybe a continued ObjC method call. This is the
10491 ;; case when we are inside a [] bracketed exp, and what
10492 ;; precede the opening bracket is not an identifier.
10500 )))
10502 ;; CASE 7F: we are looking at an arglist continuation line,
10503 ;; but the preceding argument is on the same line as the
10504 ;; opening paren. This case includes multi-line
10505 ;; mathematical paren groupings, but we could be on a
10506 ;; for-list continuation line. C.f. case 7A.
10523 paren-state))
10525 ;; CASE 7G: we are looking at just a normal arglist
10526 ;; continuation line
10529 ))
10531 ;; CASE 8: func-local multi-inheritance line
10541 ;; CASE 8A: non-hanging colon on an inher intro
10546 ;; CASE 8B: hanging colon on an inher intro
10550 ;; CASE 8C: a continued inheritance line
10554 )))
10556 ;; CASE 9: we are inside a brace-list
10566 ;; CASE 9A: In the middle of a special brace list opener.
10586 ;; CASE 9B: brace-list-close brace
10588 ;; Check special brace list closer.
10595 ;; We were between the special close char and the `)'.
10598 ;; We were before the special close char.
10602 ;; Normal brace list check.
10613 ;; Prepare for the rest of the cases below by going to the
10614 ;; token following the opening brace
10627 ;; CASE 9C: we're looking at the first line in a brace-list
10638 ;; CASE 9D: this is just a later brace-list-entry or
10639 ;; brace-entry-open
10648 ))
10649 ))))
10651 ;; CASE 10: A continued statement or top level construct.
10661 ;; This is shared with case 18.
10663 char-after-ip
10664 placeholder
10665 containing-sexp
10666 paren-state))
10668 ;; CASE 16: block close brace, possibly closing the defun or
10669 ;; the class
10671 ;; From here on we have the next containing sexp in lim.
10676 ;; CASE 16E: Closing a statement block? This catches
10677 ;; cases where it's preceded by a statement keyword,
10678 ;; which works even when used in an "invalid" context,
10679 ;; e.g. a macro argument.
10684 ;; CASE 16A: closing a lambda defun or an in-expression
10685 ;; block? C.f. cases 4, 7B and 17E.
10688 nil))
10690 'inline-close
10700 paren-state)
10704 ;; CASE 16B: does this close an inline or a function in
10705 ;; a non-class declaration level block?
10712 nil))
10722 ;; CASE 16F: Can be a defun-close of a function declared
10723 ;; in a statement block, e.g. in Pike or when using gcc
10724 ;; extensions, but watch out for macros followed by
10725 ;; blocks. Let it through to be handled below.
10726 ;; C.f. cases B.3 and 17G.
10732 ;; Turn off recognition of constructs that
10733 ;; lacks a type in this case, since that's more
10734 ;; likely to be a macro followed by a block.
10741 ;; CASE 16C: If there is an enclosing brace then this is
10742 ;; a block close since defun closes inside declaration
10743 ;; level blocks have been handled above.
10745 ;; If the block is preceded by a case/switch label on
10746 ;; the same line, we anchor at the first preceding label
10747 ;; at boi. The default handling in c-add-stmt-syntax
10748 ;; really fixes it better, but we do like this to keep
10749 ;; the indentation compatible with version 5.28 and
10750 ;; earlier. C.f. case 17H.
10757 ;; c-backward-to-block-anchor not necessary here; those
10758 ;; situations are handled in case 16E above.
10761 ;; CASE 16D: Only top level defun close left.
10767 paren-state))
10768 ))
10770 ;; CASE 19: line is an expression, not a statement, and is directly
10771 ;; contained by a template delimiter. Most likely, we are in a
10772 ;; template arglist within a statement. This case is based on CASE
10773 ;; 7. At some point in the future, we may wish to create more
10774 ;; syntactic symbols such as `template-intro',
10775 ;; `template-cont-nonempty', etc., and distinguish between them as we
10776 ;; do for `arglist-intro' etc. (2009-12-07).
10782 ; '<') before indent-point.
10790 paren-state))
10792 ;; CASE 17: Statement or defun catchall.
10795 ;; Back up statements until we find one that starts at boi.
10798 containing-sexp)))
10802 nil)
10807 ;; CASE 17B: continued statement
10811 containing-sexp paren-state))
10813 ;; CASE 17A: After a case/default label?
10821 'statement-case-open
10823 nil t containing-sexp paren-state))
10825 ;; CASE 17D: any old statement
10832 containing-sexp paren-state)
10836 ;; CASE 17I: Inside a substatement block.
10838 ;; The following tests are all based on containing-sexp.
10840 ;; From here on we have the next containing sexp in lim.
10845 lim paren-state)
10849 ;; CASE 17E: first statement in an in-expression block.
10850 ;; C.f. cases 4, 7B and 16A.
10853 nil))
10855 'defun-block-intro
10864 paren-state)
10870 ;; CASE 17F: first statement in an inline, or first
10871 ;; statement in a top-level defun. we can tell this is it
10872 ;; if there are no enclosing braces that haven't been
10873 ;; narrowed out by a class (i.e. don't use bod here).
10876 paren-state)))
10882 nil))))
10887 ;; CASE 17G: First statement in a function declared inside
10888 ;; a normal block. This can occur in Pike and with
10889 ;; e.g. the gcc extensions, but watch out for macros
10890 ;; followed by blocks. C.f. cases B.3 and 16F.
10896 ;; Turn off recognition of constructs that lacks
10897 ;; a type in this case, since that's more likely
10898 ;; to be a macro followed by a block.
10904 lim paren-state))
10906 ;; CASE 17H: First statement in a block.
10908 ;; If the block is preceded by a case/switch label on the
10909 ;; same line, we anchor at the first preceding label at
10910 ;; boi. The default handling in c-add-stmt-syntax is
10911 ;; really fixes it better, but we do like this to keep the
10912 ;; indentation compatible with version 5.28 and earlier.
10913 ;; C.f. case 16C.
10920 ;; c-backward-to-block-anchor not necessary here; those
10921 ;; situations are handled in case 17I above.
10923 lim paren-state))
10926 ))
10927 )
10929 ;; now we need to look at any modifiers
10933 ;; are we looking at a comment only line?
10938 ;; we might want to give additional offset to friends (in C++).
10943 ;; Set syntactic-relpos.
10949 nil)
10950 t))
10953 ;; Start of or a continuation of a preprocessor directive?
10968 ;; If inside a cpp expression, i.e. anywhere in a
10969 ;; cpp directive except a #define body, we only let
10970 ;; through the syntactic analysis that is internal
10971 ;; in the expression. That means the arglist
10972 ;; elements, if they are anchored inside the cpp
10973 ;; expression.
10982 ;; Inside a #define body and the syntactic analysis is
10983 ;; anchored on the start of the #define. In this case
10984 ;; we add cpp-define-intro to get the extra
10985 ;; indentation of the #define body.
10988 ;; return the syntax
10989 c-syntactic-context)))
10991 \f
10992 ;; Indentation calculation.
10995 ;; offset can be a number, a function, a variable, a list, or one of
10996 ;; the symbols + or -
10997 ;;
10998 ;; This function might do hidden buffer changes.
11017 langelem symbol))
11023 offset symbol)
11024 nil)
11038 Error evaluating offset %S for %s: \
11039 Cannot combine absolute offset %S with relative %S in `%s' method"
11045 Error evaluating offset %S for %s: \
11046 Cannot combine relative offset %S with absolute %S in `%s' method"
11051 res))
11069 Error evaluating offset %S for %s: \
11070 Cannot combine absolute offsets %S and %S in `add' method"
11074 res))
11083 res))))
11090 nil))))
11094 res
11096 offset symbol res)
11097 nil)))
11100 ;; Get offset from LANGELEM which is a list beginning with the
11101 ;; syntactic symbol and followed by any analysis data it provides.
11102 ;; That data may be zero or more elements, but if at least one is
11103 ;; given then the first is the anchor position (or nil). The symbol
11104 ;; is matched against `c-offsets-alist' and the offset calculated
11105 ;; from that is returned.
11106 ;;
11107 ;; This function might do hidden buffer changes.
11117 offset
11121 ))
11124 ;; This is a compatibility wrapper for `c-calc-offset' in case
11125 ;; someone is calling it directly. It takes an old style syntactic
11126 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
11127 ;; new list form.
11128 ;;
11129 ;; This function might do hidden buffer changes.
11136 ;; Calculate the syntactic indentation from a syntactic description
11137 ;; as returned by `c-guess-syntax'.
11138 ;;
11139 ;; Note that topmost-intro always has an anchor position at bol, for
11140 ;; historical reasons. It's often used together with other symbols
11141 ;; that has more sane positions. Since we always use the first
11142 ;; found anchor position, we rely on that these other symbols always
11143 ;; precede topmost-intro in the LANGELEMS list.
11144 ;;
11145 ;; This function might do hidden buffer changes.
11153 ;; Got an absolute column that overrides any indentation
11154 ;; we've collected so far, but not the relative
11155 ;; indentation we might get for the nested structures
11156 ;; further down the langelems list.
11160 ;; Got a relative change of the current calculated
11161 ;; indentation.
11164 ;; Use the anchor position from the first syntactic
11165 ;; element with one.
11175 indent)))
11177 \f
11180 ;;; cc-engine.el ends here