| blob | 82aee7bdbb937c9c9955537fa383f5a6fef69054 |
1 ;;; cc-engine.el --- core syntax guessing engine for CC mode
3 ;; Copyright (C) 1985, 1987, 1992-2012 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)))
150 ;; Silence the compiler.
153 \f
154 ;; Make declarations for all the `c-lang-defvar' variables in cc-langs.
167 \f
168 ;;; Internal state variables.
170 ;; Internal state of hungry delete key feature
174 ;; The electric flag (toggled by `c-toggle-electric-state').
175 ;; If t, electric actions (like automatic reindentation, and (if
176 ;; c-auto-newline is also set) auto newlining) will happen when an electric
177 ;; key like `{' is pressed (or an electric keyword like `else').
181 ;; Internal state of auto newline feature.
185 ;; Included in the mode line to indicate the active submodes.
186 ;; (defvar c-submode-indicators nil)
187 ;; (make-variable-buffer-local 'c-submode-indicators)
190 ;; Calculate the new state of PREVSTATE, t or nil, based on arg. If
191 ;; arg is nil or zero, toggle the state. If arg is negative, turn
192 ;; the state off, and if arg is positive, turn the state on
198 \f
199 ;; Basic handling of preprocessor directives.
201 ;; This is a dynamically bound cache used together with
202 ;; `c-query-macro-start' and `c-query-and-set-macro-start'. It only
203 ;; works as long as point doesn't cross a macro boundary.
212 c-macro-start))
220 c-macro-start))
222 ;; One element macro cache to cope with continual movement within very large
223 ;; CPP macros.
226 ;; Nil or cons of the bounds of the most recent CPP form probed by
227 ;; `c-beginning-of-macro', `c-end-of-macro' or `c-syntactic-end-of-macro'.
228 ;; The cdr will be nil if we know only the start of the CPP form.
231 ;; The starting position from where we determined `c-macro-cache'.
234 ;; non-nil iff `c-macro-cache' has both elements set AND the cdr is at a
235 ;; syntactic end of macro, not merely an apparent one.
238 ;; Called from a before-change function. If the change region is before or
239 ;; in the macro characterized by `c-macro-cache' etc., nullify it
240 ;; appropriately. BEG and END are the standard before-change-functions
241 ;; parameters. END isn't used.
246 c-macro-cache-start-pos nil
247 c-macro-cache-syntactic nil))
252 c-macro-cache-syntactic nil))))
255 "Go to the beginning of a preprocessor directive.
256 Leave point at the beginning of the directive and return t if in one,
257 otherwise return nil and leave point unchanged.
259 Note that this function might do hidden buffer changes. See the
260 comment at the start of cc-engine.el for more info."
272 t))
274 c-macro-cache-start-pos nil
275 c-macro-cache-syntactic nil)
287 c-macro-cache-start-pos here)
288 t)
290 nil))))))
293 "Go to the end of a preprocessor directive.
294 More accurately, move the point to the end of the closest following
295 line that doesn't end with a line continuation backslash - no check is
296 done that the point is inside a cpp directive to begin with.
298 Note that this function might do hidden buffer changes. See the
299 comment at the start of cc-engine.el for more info."
308 c-macro-cache-start-pos nil
309 c-macro-cache-syntactic nil))
315 t)))
320 ;; Go to the end of a CPP directive, or a "safe" pos just before.
321 ;;
322 ;; This is normally the end of the next non-escaped line. A "safe"
323 ;; position is one not within a string or comment. (The EOL on a line
324 ;; comment is NOT "safe").
325 ;;
326 ;; This function must only be called from the beginning of a CPP construct.
327 ;;
328 ;; Note that this function might do hidden buffer changes. See the comment
329 ;; at the start of cc-engine.el for more info.
332 s)
344 ;; Assuming point is at the "#" that introduces a preprocessor
345 ;; directive, it's moved forward to the end of the identifier which is
346 ;; "#define"d (or whatever c-opt-cpp-macro-define specifies). Non-nil
347 ;; is returned in this case, in all other cases nil is returned and
348 ;; point isn't moved.
349 ;;
350 ;; This function might do hidden buffer changes.
356 ;; Assuming point is at the "#" that introduces a preprocessor
357 ;; directive, it's moved forward to the start of the definition body
358 ;; if it's a "#define" (or whatever c-opt-cpp-macro-define
359 ;; specifies). Non-nil is returned in this case, in all other cases
360 ;; nil is returned and point isn't moved.
361 ;;
362 ;; This function might do hidden buffer changes.
368 \f
369 ;;; Basic utility functions.
372 ;; Return the given region as a string where all syntactic
373 ;; whitespace is removed or, where necessary, replaced with a single
374 ;; space. If PAREN-LEVEL is given then all parens in the region are
375 ;; collapsed to "()", "[]" etc.
376 ;;
377 ;; This function might do hidden buffer changes.
421 ;; Shift the indentation of the current line with the specified
422 ;; amount (positive inwards). The buffer is modified only if
423 ;; SHIFT-AMT isn't equal to zero.
426 tmp-char-inserted)
428 nil
429 ;; If we're on an empty line inside a macro, we take the point
430 ;; to be at the current indentation and shift it to the
431 ;; appropriate column. This way we don't treat the extra
432 ;; whitespace out to the line continuation as indentation.
448 ;; If initial point was within line's indentation and we're not on
449 ;; a line with a line continuation in a macro, position after the
450 ;; indentation. Else stay at same point in text.
458 ;; Return non-nil if the string KEYWORD is a known keyword. More
459 ;; precisely, the value is the symbol for the keyword in
460 ;; `c-keywords-obarray'.
464 ;; Return non-nil if the symbol KEYWORD-SYM, as returned by
465 ;; `c-keyword-sym', is a member of LANG-CONSTANT, which is the name
466 ;; of a language constant that ends with "-kwds". If KEYWORD-SYM is
467 ;; nil then the result is nil.
470 ;; String syntax chars, suitable for skip-syntax-(forward|backward).
475 ;; Regexp matching string limit syntax.
480 ;; Regexp matching WS followed by string limit syntax.
484 ;; Holds formatted error strings for the few cases where parse errors
485 ;; are reported.
492 c-parsing-error)
494 ;; Faces given to comments and string literals. This is used in some
495 ;; situations to speed up recognition; it isn't mandatory that font
496 ;; locking is in use. This variable is extended with the face in
497 ;; `c-doc-face-name' when fontification is activated in cc-fonts.el.
501 ;; New in Emacs 22.
505 ;; Put a c-type property with the given value at POS.
509 ;; Remove all occurrences of the c-type property that has the given
510 ;; value in the region between FROM and TO. VALUE is assumed to not
511 ;; be nil.
512 ;;
513 ;; Note: This assumes that c-type is put on single chars only; it's
514 ;; very inefficient if matching properties cover large regions.
523 \f
524 ;; Some debug tools to visualize various special positions. This
525 ;; debug code isn't as portable as the rest of CC Mode.
561 \f
562 ;; `c-beginning-of-statement-1' and accompanying stuff.
564 ;; KLUDGE ALERT: c-maybe-labelp is used to pass information between
565 ;; c-crosses-statement-barrier-p and c-beginning-of-statement-1. A
566 ;; better way should be implemented, but this will at least shut up
567 ;; the byte compiler.
570 ;; New awk-compatible version of c-beginning-of-statement-1, ACM 2002/6/22
572 ;; Macros used internally in c-beginning-of-statement-1 for the
573 ;; automaton actions.
576 stack)))
581 t
582 ,do-if-done
587 ;; Throw nil if stack is empty, else throw non-nil.
611 noerror comma-delim)
612 "Move to the start of the current statement or declaration, or to
613 the previous one if already at the beginning of one. Only
614 statements/declarations on the same level are considered, i.e. don't
615 move into or out of sexps (not even normal expression parentheses).
617 If point is already at the earliest statement within braces or parens,
618 this function doesn't move back into any whitespace preceding it; it
619 returns 'same in this case.
623 is within the continuation. If starting at such a token, move to the
624 corresponding statement start. If at the beginning of a statement,
625 move to the closest containing statement if there is any. This might
626 also stop at a continuation clause.
628 Labels are treated as part of the following statements if
629 IGNORE-LABELS is non-nil. (FIXME: Doesn't work if we stop at a known
630 statement start keyword.) Otherwise, each label is treated as a
631 separate statement.
633 Macros are ignored \(i.e. skipped over) unless point is within one, in
634 which case the content of the macro is treated as normal code. Aside
635 from any normal statement starts found in it, stop at the first token
638 macros before leaving them.
640 Return:
642 'same if stopped at the beginning of the current statement;
643 'up if stepped to a containing statement;
644 'previous if stepped to a preceding statement;
645 'beginning if stepped from a statement continuation clause to
646 its start clause; or
647 'macro if stepped to a macro start.
648 Note that 'same and not 'label is returned if stopped at the same
649 label without crossing the colon character.
651 LIM may be given to limit the search. If the search hits the limit,
652 point will be left at the closest following token, or at the start
653 position if that is less ('same is returned in this case).
655 NOERROR turns off error logging to `c-parsing-error'.
657 Normally only ';' and virtual semicolons are considered to delimit
658 statements, but if COMMA-DELIM is non-nil then ',' is treated
659 as a delimiter too.
661 Note that this function might do hidden buffer changes. See the
662 comment at the start of cc-engine.el for more info."
664 ;; The bulk of this function is a pushdown automaton that looks at statement
665 ;; boundaries and the tokens (such as "while") in c-opt-block-stmt-key. Its
666 ;; purpose is to keep track of nested statements, ensuring that such
667 ;; statements are skipped over in their entirety (somewhat akin to what C-M-p
668 ;; does with nested braces/brackets/parentheses).
669 ;;
670 ;; Note: The position of a boundary is the following token.
671 ;;
672 ;; Beginning with the current token (the one following point), move back one
673 ;; sexp at a time (where a sexp is, more or less, either a token or the
674 ;; entire contents of a brace/bracket/paren pair). Each time a statement
675 ;; boundary is crossed or a "while"-like token is found, update the state of
676 ;; the PDA. Stop at the beginning of a statement when the stack (holding
677 ;; nested statement info) is empty and the position has been moved.
678 ;;
679 ;; The following variables constitute the PDA:
680 ;;
681 ;; sym: This is either the "while"-like token (e.g. 'for) we've just
682 ;; scanned back over, 'boundary if we've just gone back over a
683 ;; statement boundary, or nil otherwise.
684 ;; state: takes one of the values (nil else else-boundary while
685 ;; while-boundary catch catch-boundary).
686 ;; nil means "no "while"-like token yet scanned".
687 ;; 'else, for example, means "just gone back over an else".
688 ;; 'else-boundary means "just gone back over a statement boundary
689 ;; immediately after having gone back over an else".
690 ;; saved-pos: A vector of either saved positions (tok ptok pptok, etc.) or
691 ;; of error reporting information.
692 ;; stack: The stack onto which the PDA pushes its state. Each entry
693 ;; consists of a saved value of state and saved-pos. An entry is
694 ;; pushed when we move back over a "continuation" token (e.g. else)
695 ;; and popped when we encounter the corresponding opening token
696 ;; (e.g. if).
697 ;;
698 ;;
699 ;; The following diagram briefly outlines the PDA.
700 ;;
701 ;; Common state:
702 ;; "else": Push state, goto state `else'.
703 ;; "while": Push state, goto state `while'.
704 ;; "catch" or "finally": Push state, goto state `catch'.
705 ;; boundary: Pop state.
706 ;; other: Do nothing special.
707 ;;
708 ;; State `else':
709 ;; boundary: Goto state `else-boundary'.
710 ;; other: Error, pop state, retry token.
711 ;;
712 ;; State `else-boundary':
713 ;; "if": Pop state.
714 ;; boundary: Error, pop state.
715 ;; other: See common state.
716 ;;
717 ;; State `while':
718 ;; boundary: Save position, goto state `while-boundary'.
719 ;; other: Pop state, retry token.
720 ;;
721 ;; State `while-boundary':
722 ;; "do": Pop state.
723 ;; boundary: Restore position if it's not at start, pop state. [*see below]
724 ;; other: See common state.
725 ;;
726 ;; State `catch':
727 ;; boundary: Goto state `catch-boundary'.
728 ;; other: Error, pop state, retry token.
729 ;;
730 ;; State `catch-boundary':
731 ;; "try": Pop state.
732 ;; "catch": Goto state `catch'.
733 ;; boundary: Error, pop state.
734 ;; other: See common state.
735 ;;
736 ;; [*] In the `while-boundary' state, we had pushed a 'while state, and were
737 ;; searching for a "do" which would have opened a do-while. If we didn't
738 ;; find it, we discard the analysis done since the "while", go back to this
739 ;; token in the buffer and restart the scanning there, this time WITHOUT
740 ;; pushing the 'while state onto the stack.
741 ;;
742 ;; In addition to the above there is some special handling of labels
743 ;; and macros.
747 macro-start
750 c-stmt-delim-chars-with-comma
751 c-stmt-delim-chars))
752 c-in-literal-cache c-maybe-labelp after-case:-pos saved
753 ;; Current position.
754 pos
755 ;; Position of last stmt boundary character (e.g. ;).
756 boundary-pos
757 ;; The position of the last sexp or bound that follows the
758 ;; first found colon, i.e. the start of the nonlabel part of
759 ;; the statement. It's `start' if a colon is found just after
760 ;; the start.
761 after-labels-pos
762 ;; Like `after-labels-pos', but the first such position inside
763 ;; a label, i.e. the start of the last label before the start
764 ;; of the nonlabel part of the statement.
765 last-label-pos
766 ;; The last position where a label is possible provided the
767 ;; statement started there. It's nil as long as no invalid
768 ;; label content has been found (according to
769 ;; `c-nonlabel-token-key'). It's `start' if no valid label
770 ;; content was found in the label. Note that we might still
771 ;; regard it a label if it starts with `c-label-kwds'.
772 label-good-pos
773 ;; Putative positions of the components of a bitfield declaration,
774 ;; e.g. "int foo : NUM_FOO_BITS ;"
775 bitfield-type-pos bitfield-id-pos bitfield-size-pos
776 ;; Symbol just scanned back over (e.g. 'while or 'boundary).
777 ;; See above.
778 sym
779 ;; Current state in the automaton. See above.
780 state
781 ;; Current saved positions. See above.
782 saved-pos
783 ;; Stack of conses (state . saved-pos).
784 stack
785 ;; Regexp which matches "for", "if", etc.
788 ;; Return value.
790 ;; Positions of the last three sexps or bounds we've stopped at.
791 tok ptok pptok)
801 ;; Try to skip back over unary operator characters, to register
802 ;; that we've moved.
806 ;; Protect post-++/-- operators just before a virtual semicolon.
810 ;; Skip back over any semicolon here. If it was a bare semicolon, we're
811 ;; done. Later on we ignore the boundaries for statements that don't
812 ;; contain any sexp. The only thing that is affected is that the error
813 ;; checking is a little less strict, and we really don't bother.
823 pos saved)
825 ;; Begin at start and not pos to detect macros if we stand
826 ;; directly after the #.
829 ;; Record this as the first token if not starting inside it.
833 ;; The following while loop goes back one sexp (balanced parens,
834 ;; etc. with contents, or symbol or suchlike) each iteration. This
835 ;; movement is accomplished with a call to c-backward-sexp approx 170
836 ;; lines below.
837 ;;
838 ;; The loop is exited only by throwing nil to the (catch 'loop ...):
839 ;; 1. On reaching the start of a macro;
840 ;; 2. On having passed a stmt boundary with the PDA stack empty;
841 ;; 3. On reaching the start of an Objective C method def;
842 ;; 4. From macro `c-bos-pop-state'; when the stack is empty;
843 ;; 5. From macro `c-bos-pop-state-and-retry' when the stack is empty.
847 ;; Are we in a macro, just after the opening #?
862 ;; Stop at the first token in the content of the macro.
866 ret 'macro
867 ignore-labels t))
870 ;; Do a round through the automaton if we've just passed a
871 ;; statement boundary or passed a "while"-like token.
879 ;; The PDA state handling.
880 ;;
881 ;; Refer to the description of the PDA in the opening
882 ;; comments. In the following OR form, the first leaf
883 ;; attempts to handles one of the specific actions detailed
884 ;; (e.g., finding token "if" whilst in state `else-boundary').
885 ;; We drop through to the second leaf (which handles common
886 ;; state) if no specific handler is found in the first cond.
887 ;; If a parsing error is detected (e.g. an "else" with no
888 ;; preceding "if"), we throw to the enclosing catch.
889 ;;
890 ;; Note that the (eq state 'else) means
891 ;; "we've just passed an else", NOT "we're looking for an
892 ;; else".
909 ;; Since this can cause backtracking we do a
910 ;; little more careful analysis to avoid it:
911 ;; If there's a label in front of the while
912 ;; it can't be part of a do-while.
941 ;; This is state common. We get here when the previous
942 ;; cond statement found no particular state handler.
944 ;; If we have a boundary at the start
945 ;; position we push a frame to go to the
946 ;; previous statement.
955 ;; Is this a real while, or a do-while?
956 ;; The next `when' triggers unless we are SURE that
957 ;; the `while' is not the tail end of a `do-while'.
960 ;; The following kludge is to prevent
961 ;; infinite recursion when called from
962 ;; c-awk-after-if-for-while-condition-p,
963 ;; or the like.
967 ;; Since this can cause backtracking we do a
968 ;; little more careful analysis to avoid it: If
969 ;; the while isn't followed by a (possibly
970 ;; virtual) semicolon it can't be a do-while.
979 ;; We're either past a statement boundary or at the
980 ;; start of a statement, so throw away any label data
981 ;; for the previous one.
983 last-label-pos nil
984 c-maybe-labelp nil))))
986 ;; Step to the previous sexp, but not if we crossed a
987 ;; boundary, since that doesn't consume an sexp.
991 ;; HERE IS THE SINGLE PLACE INSIDE THE PDA LOOP WHERE WE MOVE
992 ;; BACKWARDS THROUGH THE SOURCE.
996 ;; The end position of the area to search for statement
997 ;; barriers in this round.
1000 ;; Go back over exactly one logical sexp, taking proper
1001 ;; account of macros and escaped EOLs.
1005 ;; Give up if we hit an unbalanced block. Since the
1006 ;; stack won't be empty the code below will report a
1007 ;; suitable error.
1010 ;; Have we moved into a macro?
1013 ;; Have we crossed a statement boundary? If not,
1014 ;; keep going back until we find one or a "real" sexp.
1021 ;; Have we just gone back over an escaped NL? This
1022 ;; doesn't count as a sexp.
1025 ;; Have we crossed a statement boundary?
1028 ;; Are we at a macro beginning?
1030 c-opt-cpp-prefix
1036 ;; Just gone back over a brace block?
1042 ;; Just gone back over some paren block?
1046 before-sws-pos)))
1049 ;; Just gone back over an ordinary symbol of some sort?
1055 ptok tok
1056 tok boundary-pos
1058 ;; Like a C "continue". Analyze the next sexp.
1061 ;; ObjC method def?
1068 ;; Might we have a bitfield declaration, "<type> <id> : <size>"?
1071 ;; The : <size> and <id> fields?
1081 ;; The <type> field?
1089 bitfield-id-pos nil)))))
1091 ;; Handle labels.
1094 ;; `c-crosses-statement-barrier-p' has found a colon, so we
1095 ;; might be in a label now. Have we got a real label
1096 ;; (including a case label) or something like C++'s "public:"?
1097 ;; A case label might use an expression rather than a token.
1100 ;; Catch C++'s inheritance construct "class foo : bar".
1114 ; been found.
1116 ;; We're in a potential label and it's the first
1117 ;; time we've found something that isn't allowed in
1118 ;; one.
1121 ;; We've moved back by a sexp, so update the token positions.
1123 pptok ptok
1124 ptok tok
1130 ;; If the stack isn't empty there might be errors to report.
1139 ;; Need to investigate closer whether we've crossed
1140 ;; between a substatement and its containing statement.
1142 ptok
1143 pptok))
1150 after-labels-pos
1157 last-label-pos
1158 pos))
1160 ;; We're in a label. Maybe we should step to the statement
1161 ;; after it.
1166 ;; Might have jumped over several labels. Go to the last one.
1169 ;; Have we got "case <expression>:"?
1179 ;; Skip over the unary operators that can start the statement.
1182 ;; protect AWK post-inc/decrement operators, etc.
1187 ret)))
1190 "Return non-nil if there is a non-comment non-macro punctuation character
1191 between FROM and TO. FROM must not be in a string or comment. The returned
1192 value is the position of the first such character."
1203 "Return non-nil if buffer positions FROM to TO cross one or more
1204 statement or declaration boundaries. The returned value is actually
1205 the position of the earliest boundary char. FROM must not be within
1206 a string or comment.
1208 The variable `c-maybe-labelp' is set to the position of the first `:' that
1209 might start a label (i.e. not part of `::' and not preceded by `?'). If a
1210 single `?' is found, then `c-maybe-labelp' is cleared.
1212 For AWK, a statement which is terminated by an EOL (not a \; or a }) is
1214 the line. If this virtual semicolon is _at_ from, the function recognizes it.
1216 Note that this function might do hidden buffer changes. See the
1217 comment at the start of cc-engine.el for more info."
1219 ;; If the current language has CPP macros, insert # into skip-chars.
1222 c-opt-cpp-symbol ; usually "#"
1224 c-stmt-delim-chars))
1227 lit-range vsemi-pos)
1234 skip-chars
1238 ;; Virtual semicolon?
1248 ;; In a string/comment?
1255 ;; Ignore scope operators.
1259 ;; A question mark. Can't be a label, so stop
1260 ;; looking for more : and ?.
1263 ;; At a CPP construct?
1271 ;; In trailing space after an as yet undetected virtual semicolon?
1276 nil))))))
1279 "Return non-nil if the point is at the first token in a statement
1280 or somewhere in the syntactic whitespace before it.
1283 Any kind of declaration-like construct that occur outside function
1286 Note that this function might do hidden buffer changes. See the
1287 comment at the start of cc-engine.el for more info."
1291 c-maybe-labelp)
1302 "Return non-nil if the point is at the first token in an expression or
1303 statement, or somewhere in the syntactic whitespace before it.
1306 grammar sense: It's any sequence of expression tokens except commas,
1307 unless they are enclosed inside parentheses of some kind. Also, an
1308 expression never continues past an enclosing parenthesis, but it might
1309 contain parenthesis pairs of any sort except braces.
1311 Since expressions never cross statement boundaries, this function also
1312 recognizes statement beginnings, just like `c-at-statement-start-p'.
1314 Note that this function might do hidden buffer changes. See the
1315 comment at the start of cc-engine.el for more info."
1320 c-maybe-labelp)
1328 \f
1329 ;; A set of functions that covers various idiosyncrasies in
1330 ;; implementations of `forward-comment'.
1332 ;; Note: Some emacsen considers incorrectly that any line comment
1333 ;; ending with a backslash continues to the next line. I can't think
1334 ;; of any way to work around that in a reliable way without changing
1335 ;; the buffer, though. Suggestions welcome. ;) (No, temporarily
1336 ;; changing the syntax for backslash doesn't work since we must treat
1337 ;; escapes in string literals correctly.)
1340 "Move forward past whitespace and the closest following comment, if any.
1341 Return t if a comment was found, nil otherwise. In either case, the
1342 point is moved past the following whitespace. Line continuations,
1343 i.e. a backslashes followed by line breaks, are treated as whitespace.
1344 The line breaks that end line comments are considered to be the
1345 comment enders, so the point will be put on the beginning of the next
1346 line if it moved past a line comment.
1348 This function does not do any hidden buffer changes."
1356 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1357 ;; forwards at eob.
1358 nil
1360 ;; Emacs includes the ending newline in a b-style (c++)
1361 ;; comment, but XEmacs doesn't. We depend on the Emacs
1362 ;; behavior (which also is symmetric).
1366 t))))
1369 "Move forward past all following whitespace and comments.
1370 Line continuations, i.e. a backslashes followed by line breaks, are
1371 treated as whitespace.
1373 Note that this function might do hidden buffer changes. See the
1374 comment at the start of cc-engine.el for more info."
1377 ;; If forward-comment in at least XEmacs 21 is given a large
1378 ;; positive value, it'll loop all the way through if it hits
1379 ;; eob.
1381 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1382 ;; forwards at eob.
1387 t))))
1390 "Move backward past whitespace and the closest preceding comment, if any.
1391 Return t if a comment was found, nil otherwise. In either case, the
1392 point is moved past the preceding whitespace. Line continuations,
1393 i.e. a backslashes followed by line breaks, are treated as whitespace.
1394 The line breaks that end line comments are considered to be the
1395 comment enders, so the point cannot be at the end of the same line to
1396 move over a line comment.
1398 This function does not do any hidden buffer changes."
1401 ;; When we got newline terminated comments, forward-comment in all
1402 ;; supported emacsen so far will stop at eol of each line not
1403 ;; ending with a comment when moving backwards. This corrects for
1404 ;; that, and at the same time handles line continuations.
1412 ;; Some emacsen (e.g. Emacs 19.34) return t when moving
1413 ;; backwards at bob.
1414 nil
1416 ;; Leave point after the closest following newline if we've
1417 ;; backed up over any above, since forward-comment won't move
1418 ;; backward over a line comment if point is at the end of the
1419 ;; same line.
1424 ;; If forward-comment above succeeded and we're at eol
1425 ;; then the newline we moved over above didn't end a
1426 ;; line comment, so we give it another go.
1429 t))
1431 ;; Emacs <= 20 and XEmacs move back over the closer of a
1432 ;; block comment that lacks an opener.
1435 t)))))
1438 "Move backward past all preceding whitespace and comments.
1439 Line continuations, i.e. a backslashes followed by line breaks, are
1440 treated as whitespace. The line breaks that end line comments are
1441 considered to be the comment enders, so the point cannot be at the end
1442 of the same line to move over a line comment. Unlike
1443 c-backward-syntactic-ws, this function doesn't move back over
1444 preprocessor directives.
1446 Note that this function might do hidden buffer changes. See the
1447 comment at the start of cc-engine.el for more info."
1451 ;; `forward-comment' in some emacsen (e.g. XEmacs 21.4)
1452 ;; return t when moving backwards at bob.
1458 ;; Emacs <= 20 and XEmacs move back over the
1459 ;; closer of a block comment that lacks an opener.
1461 t)
1463 ;; XEmacs treats line continuations as whitespace but
1464 ;; only in the backward direction, which seems a bit
1465 ;; odd. Anyway, this is necessary for Emacs.
1470 t))))))
1472 \f
1473 ;; Tools for skipping over syntactic whitespace.
1475 ;; The following functions use text properties to cache searches over
1476 ;; large regions of syntactic whitespace. It works as follows:
1477 ;;
1478 ;; o If a syntactic whitespace region contains anything but simple
1479 ;; whitespace (i.e. space, tab and line breaks), the text property
1480 ;; `c-in-sws' is put over it. At places where we have stopped
1481 ;; within that region there's also a `c-is-sws' text property.
1482 ;; That since there typically are nested whitespace inside that
1483 ;; must be handled separately, e.g. whitespace inside a comment or
1484 ;; cpp directive. Thus, from one point with `c-is-sws' it's safe
1485 ;; to jump to another point with that property within the same
1486 ;; `c-in-sws' region. It can be likened to a ladder where
1487 ;; `c-in-sws' marks the bars and `c-is-sws' the rungs.
1488 ;;
1489 ;; o The `c-is-sws' property is put on the simple whitespace chars at
1490 ;; a "rung position" and also maybe on the first following char.
1491 ;; As many characters as can be conveniently found in this range
1492 ;; are marked, but no assumption can be made that the whole range
1493 ;; is marked (it could be clobbered by later changes, for
1494 ;; instance).
1495 ;;
1496 ;; Note that some part of the beginning of a sequence of simple
1497 ;; whitespace might be part of the end of a preceding line comment
1498 ;; or cpp directive and must not be considered part of the "rung".
1499 ;; Such whitespace is some amount of horizontal whitespace followed
1500 ;; by a newline. In the case of cpp directives it could also be
1501 ;; two newlines with horizontal whitespace between them.
1502 ;;
1503 ;; The reason to include the first following char is to cope with
1504 ;; "rung positions" that doesn't have any ordinary whitespace. If
1505 ;; `c-is-sws' is put on a token character it does not have
1506 ;; `c-in-sws' set simultaneously. That's the only case when that
1507 ;; can occur, and the reason for not extending the `c-in-sws'
1508 ;; region to cover it is that the `c-in-sws' region could then be
1509 ;; accidentally merged with a following one if the token is only
1510 ;; one character long.
1511 ;;
1512 ;; o On buffer changes the `c-in-sws' and `c-is-sws' properties are
1513 ;; removed in the changed region. If the change was inside
1514 ;; syntactic whitespace that means that the "ladder" is broken, but
1515 ;; a later call to `c-forward-sws' or `c-backward-sws' will use the
1516 ;; parts on either side and use an ordinary search only to "repair"
1517 ;; the gap.
1518 ;;
1519 ;; Special care needs to be taken if a region is removed: If there
1520 ;; are `c-in-sws' on both sides of it which do not connect inside
1521 ;; the region then they can't be joined. If e.g. a marked macro is
1522 ;; broken, syntactic whitespace inside the new text might be
1523 ;; marked. If those marks would become connected with the old
1524 ;; `c-in-sws' range around the macro then we could get a ladder
1525 ;; with one end outside the macro and the other at some whitespace
1526 ;; within it.
1527 ;;
1528 ;; The main motivation for this system is to increase the speed in
1529 ;; skipping over the large whitespace regions that can occur at the
1530 ;; top level in e.g. header files that contain a lot of comments and
1531 ;; cpp directives. For small comments inside code it's probably
1532 ;; slower than using `forward-comment' straightforwardly, but speed is
1533 ;; not a significant factor there anyway.
1535 ; (defface c-debug-is-sws-face
1536 ; '((t (:background "GreenYellow")))
1537 ; "Debug face to mark the `c-is-sws' property.")
1538 ; (defface c-debug-in-sws-face
1539 ; '((t (:underline t)))
1540 ; "Debug face to mark the `c-in-sws' property.")
1542 ; (defun c-debug-put-sws-faces ()
1543 ; ;; Put the sws debug faces on all the `c-is-sws' and `c-in-sws'
1544 ; ;; properties in the buffer.
1545 ; (interactive)
1546 ; (save-excursion
1547 ; (c-save-buffer-state (in-face)
1548 ; (goto-char (point-min))
1549 ; (setq in-face (if (get-text-property (point) 'c-is-sws)
1550 ; (point)))
1551 ; (while (progn
1552 ; (goto-char (next-single-property-change
1553 ; (point) 'c-is-sws nil (point-max)))
1554 ; (if in-face
1555 ; (progn
1556 ; (c-debug-add-face in-face (point) 'c-debug-is-sws-face)
1557 ; (setq in-face nil))
1558 ; (setq in-face (point)))
1559 ; (not (eobp))))
1560 ; (goto-char (point-min))
1561 ; (setq in-face (if (get-text-property (point) 'c-in-sws)
1562 ; (point)))
1563 ; (while (progn
1564 ; (goto-char (next-single-property-change
1565 ; (point) 'c-in-sws nil (point-max)))
1566 ; (if in-face
1567 ; (progn
1568 ; (c-debug-add-face in-face (point) 'c-debug-in-sws-face)
1569 ; (setq in-face nil))
1570 ; (setq in-face (point)))
1571 ; (not (eobp)))))))
1574 ;;`(message ,@args)
1575 )
1578 ;; This macro does a hidden buffer change.
1585 ;; This macro does a hidden buffer change.
1592 ;; This macro does a hidden buffer change.
1599 ;; This macro does a hidden buffer change.
1606 ;; This macro does a hidden buffer change.
1614 ;; Called from `after-change-functions'. Note that if
1615 ;; `c-forward-sws' or `c-backward-sws' are used outside
1616 ;; `c-save-buffer-state' or similar then this will remove the cache
1617 ;; properties right after they're added.
1618 ;;
1619 ;; This function does hidden buffer changes.
1622 ;; Adjust the end to remove the properties in any following simple
1623 ;; ws up to and including the next line break, if there is any
1624 ;; after the changed region. This is necessary e.g. when a rung
1625 ;; marked empty line is converted to a line comment by inserting
1626 ;; "//" before the line break. In that case the line break would
1627 ;; keep the rung mark which could make a later `c-backward-sws'
1628 ;; move into the line comment instead of over it.
1638 ;; Ensure that an `c-in-sws' range gets broken. Note that it isn't
1639 ;; safe to keep a range that was continuous before the change. E.g:
1640 ;;
1641 ;; #define foo
1642 ;; \
1643 ;; bar
1644 ;;
1645 ;; There can be a "ladder" between "#" and "b". Now, if the newline
1646 ;; after "foo" is removed then "bar" will become part of the cpp
1647 ;; directive instead of a syntactically relevant token. In that
1648 ;; case there's no longer syntactic ws from "#" to "b".
1655 ;; Used by `c-forward-syntactic-ws' to implement the unbounded search.
1656 ;;
1657 ;; This function might do hidden buffer changes.
1660 ;; unmarked part of the simple ws region.
1662 rung-is-marked next-rung-is-marked simple-ws-end
1663 ;; `safe-start' is set when it's safe to cache the start position.
1664 ;; It's not set if we've initially skipped over comments and line
1665 ;; continuations since we might have gone out through the end of a
1666 ;; macro then. This provision makes `c-forward-sws' not populate the
1667 ;; cache in the majority of cases, but otoh is `c-backward-sws' by far
1668 ;; more common.
1669 safe-start)
1671 ;; Skip simple ws and do a quick check on the following character to see
1672 ;; if it's anything that can't start syntactic ws, so we can bail out
1673 ;; early in the majority of cases when there just are a few ws chars.
1680 ;; Find the last rung position to avoid setting properties in all
1681 ;; the cases when the marked rung is complete.
1682 ;; (`next-single-property-change' is certain to move at least one
1683 ;; step forward.)
1686 ;; Got no marked rung here. Since the simple ws might have started
1687 ;; inside a line comment or cpp directive we must set `rung-pos' as
1688 ;; high as possible.
1697 ;; The following search is the main reason that `c-in-sws'
1698 ;; and `c-is-sws' aren't combined to one property.
1702 ;; If the `c-in-sws' region extended past the last
1703 ;; `c-is-sws' char we have to go back a bit.
1710 "c-forward-sws cached move %s -> %s (max %s)"
1717 ;; We'll loop here if there is simple ws after the last rung.
1718 ;; That means that there's been some change in it and it's
1719 ;; possible that we've stepped into another ladder, so extend
1720 ;; the previous one to join with it if there is one, and try to
1721 ;; use the cache again.
1723 "c-forward-sws extending rung with [%s..%s] (max %s)"
1726 ;; Remove any `c-in-sws' property from the last char of
1727 ;; the rung before we mark it with `c-is-sws', so that we
1728 ;; won't connect with the remains of a broken "ladder".
1734 last-put-in-sws-pos rung-pos)))
1741 ;; Skipped over comments. Don't cache at eob in case the buffer
1742 ;; is narrowed.
1753 ;; Skip a preprocessor directive.
1760 ;; Don't cache at eob in case the buffer is narrowed.
1763 ;; We've searched over a piece of non-white syntactic ws. See if this
1764 ;; can be cached.
1770 ;; Cache if we haven't skipped comments only, and if we started
1771 ;; either from a marked rung or from a completely uncached
1772 ;; position.
1777 ;; See if there's a marked rung in the encountered simple ws. If
1778 ;; so then we can cache, unless `safe-start' is nil. Even then
1779 ;; we need to do this to check if the cache can be used for the
1780 ;; next step.
1784 safe-start))
1788 "c-forward-sws caching [%s..%s] - [%s..%s] (max %s)"
1792 ;; Remove the properties for any nested ws that might be cached.
1793 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
1794 ;; anyway.
1802 last-put-in-sws-pos rung-pos))
1804 ;; Remove any `c-in-sws' property from the last char of
1805 ;; the rung before we mark it with `c-is-sws', so that we
1806 ;; won't connect with the remains of a broken "ladder".
1809 rung-end-pos))
1812 "c-forward-sws not caching [%s..%s] - [%s..%s] (max %s)"
1816 ;; Set `rung-pos' for the next rung. It's the same thing here as
1817 ;; initially, except that the rung position is set as early as
1818 ;; possible since we can't be in the ending ws of a line comment or
1819 ;; cpp directive now.
1826 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
1827 ;; another one after the point (which might occur when editing inside a
1828 ;; comment or macro).
1832 "c-forward-sws clearing at %s for cache separation"
1833 last-put-in-sws-pos)
1837 ;; If at eob we have to clear the last character before the end
1838 ;; instead since the buffer might be narrowed and there might
1839 ;; be a `c-in-sws' after (point-max). In this case it's
1840 ;; necessary to clear both properties.
1842 "c-forward-sws clearing thoroughly at %s for cache separation"
1845 last-put-in-sws-pos))))
1846 )))
1849 ;; Used by `c-backward-syntactic-ws' to implement the unbounded search.
1850 ;;
1851 ;; This function might do hidden buffer changes.
1854 ;; part of the simple ws region.
1856 rung-is-marked simple-ws-beg cmt-skip-pos)
1858 ;; Skip simple horizontal ws and do a quick check on the preceding
1859 ;; character to see if it's anything that can't end syntactic ws, so we can
1860 ;; bail out early in the majority of cases when there just are a few ws
1861 ;; chars. Newlines are complicated in the backward direction, so we can't
1862 ;; skip over them.
1869 ;; Try to find a rung position in the simple ws preceding point, so that
1870 ;; we can get a cache hit even if the last bit of the simple ws has
1871 ;; changed recently.
1877 ;; `rung-pos' will be the earliest marked position, which means that
1878 ;; there might be later unmarked parts in the simple ws region.
1879 ;; It's not worth the effort to fix that; the last part of the
1880 ;; simple ws is also typically edited often, so it could be wasted.
1891 ;; The following search is the main reason that `c-in-sws'
1892 ;; and `c-is-sws' aren't combined to one property.
1896 ;; If the `c-in-sws' region extended past the first
1897 ;; `c-is-sws' char we have to go forward a bit.
1902 "c-backward-sws cached move %s <- %s (min %s)"
1914 t
1916 nil))
1918 ;; We'll loop here if there is simple ws before the first rung.
1919 ;; That means that there's been some change in it and it's
1920 ;; possible that we've stepped into another ladder, so extend
1921 ;; the previous one to join with it if there is one, and try to
1922 ;; use the cache again.
1924 "c-backward-sws extending rung with [%s..%s] (min %s)"
1927 ;; Remove any `c-in-sws' property from the last char of
1928 ;; the rung before we mark it with `c-is-sws', so that we
1929 ;; won't connect with the remains of a broken "ladder".
1932 rung-pos)
1936 last-put-in-sws-pos rung-pos))
1945 ;; Inside a cpp directive. See if it should be skipped over.
1948 ;; Move back over all line continuations in the region skipped
1949 ;; over by `c-backward-comments'. If we go past it then we
1950 ;; started inside the cpp directive.
1959 ;; Don't move past the cpp directive if we began inside
1960 ;; it. Note that the position at the end of the last line
1961 ;; of the macro is also considered to be within it.
1963 nil)
1965 ;; It's worthwhile to spend a little bit of effort on finding
1966 ;; the end of the macro, to get a good `simple-ws-beg'
1967 ;; position for the cache. Note that `c-backward-comments'
1968 ;; could have stepped over some comments before going into
1969 ;; the macro, and then `simple-ws-beg' must be kept on the
1970 ;; same side of those comments.
1977 ;; Might happen if comments after the macro were skipped
1978 ;; over.
1982 t)))
1987 simple-ws-beg)
1988 ;; Skipped over comments. Must put point at the end of
1989 ;; the simple ws at point since we might be after a line
1990 ;; comment or cpp directive that's been partially
1991 ;; narrowed out, and we can't risk marking the simple ws
1992 ;; at the end of it.
1994 t)))
1996 ;; We've searched over a piece of non-white syntactic ws. See if this
1997 ;; can be cached.
2002 ;; Cache if we started either from a marked rung or from a
2003 ;; completely uncached position.
2004 rung-is-marked
2007 ;; Cache if there's a marked rung in the encountered simple ws.
2015 "c-backward-sws caching [%s..%s] - [%s..%s] (min %s)"
2020 ;; Remove the properties for any nested ws that might be cached.
2021 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
2022 ;; anyway.
2027 ;; Remove any `c-in-sws' property from the last char of
2028 ;; the rung before we mark it with `c-is-sws', so that we
2029 ;; won't connect with the remains of a broken "ladder".
2032 rung-end-pos)
2035 last-put-in-sws-pos simple-ws-beg)
2036 rung-pos)
2041 "c-backward-sws not caching [%s..%s] - [%s..%s] (min %s)"
2047 ))
2049 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
2050 ;; another one before the point (which might occur when editing inside a
2051 ;; comment or macro).
2055 "c-backward-sws clearing at %s for cache separation"
2058 last-put-in-sws-pos))
2060 ;; If at bob and the buffer is narrowed, we have to clear the
2061 ;; character we're standing on instead since there might be a
2062 ;; `c-in-sws' before (point-min). In this case it's necessary
2063 ;; to clear both properties.
2065 "c-backward-sws clearing thoroughly at %s for cache separation"
2066 last-put-in-sws-pos)
2069 )))
2071 \f
2072 ;; Other whitespace tools
2074 ;; Is the region (beg end) WS, and is there WS (or BOB/EOB) next to the
2075 ;; region? This is a "heuristic" function. .....
2076 ;;
2077 ;; The motivation for the second bit is to check whether removing this
2078 ;; region would coalesce two symbols.
2079 ;;
2080 ;; FIXME!!! This function doesn't check virtual semicolons in any way. Be
2081 ;; careful about using this function for, e.g. AWK. (2007/3/7)
2090 \f
2091 ;; A system for finding noteworthy parens before the point.
2094 ;; A maximum comfortable scanning distance, e.g. between
2095 ;; `c-state-cache-good-pos' and "HERE" (where we call c-parse-state). When
2096 ;; this distance is exceeded, we take "emergency measures", e.g. by clearing
2097 ;; the cache and starting again from point-min or a beginning of defun. This
2098 ;; value can be tuned for efficiency or set to a lower value for testing.
2102 ;; The state cache used by `c-parse-state' to cut down the amount of
2103 ;; searching. It's the result from some earlier `c-parse-state' call. See
2104 ;; `c-parse-state''s doc string for details of its structure.
2105 ;;
2106 ;; The use of the cached info is more effective if the next
2107 ;; `c-parse-state' call is on a line close by the one the cached state
2108 ;; was made at; the cache can actually slow down a little if the
2109 ;; cached state was made very far back in the buffer. The cache is
2110 ;; most effective if `c-parse-state' is used on each line while moving
2111 ;; forward.
2115 ;; This is a position where `c-state-cache' is known to be correct, or
2116 ;; nil (see below). It's a position inside one of the recorded unclosed
2117 ;; parens or the top level, but not further nested inside any literal or
2118 ;; subparen that is closed before the last recorded position.
2119 ;;
2120 ;; The exact position is chosen to try to be close to yet earlier than
2121 ;; the position where `c-state-cache' will be called next. Right now
2122 ;; the heuristic is to set it to the position after the last found
2123 ;; closing paren (of any type) before the line on which
2124 ;; `c-parse-state' was called. That is chosen primarily to work well
2125 ;; with refontification of the current line.
2126 ;;
2127 ;; 2009-07-28: When `c-state-point-min' and the last position where
2128 ;; `c-parse-state' or for which `c-invalidate-state-cache' was called, are
2129 ;; both in the same literal, there is no such "good position", and
2130 ;; c-state-cache-good-pos is then nil. This is the ONLY circumstance in which
2131 ;; it can be nil. In this case, `c-state-point-min-literal' will be non-nil.
2132 ;;
2133 ;; 2009-06-12: In a brace desert, c-state-cache-good-pos may also be in
2134 ;; the middle of the desert, as long as it is not within a brace pair
2135 ;; recorded in `c-state-cache' or a paren/bracket pair.
2138 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2139 ;; We maintain a simple cache of positions which aren't in a literal, so as to
2140 ;; speed up testing for non-literality.
2142 ;; The approximate interval between entries in `c-state-nonlit-pos-cache'.
2146 ;; A list of buffer positions which are known not to be in a literal or a cpp
2147 ;; construct. This is ordered with higher positions at the front of the list.
2148 ;; Only those which are less than `c-state-nonlit-pos-cache-limit' are valid.
2152 ;; An upper limit on valid entries in `c-state-nonlit-pos-cache'. This is
2153 ;; reduced by buffer changes, and increased by invocations of
2154 ;; `c-state-literal-at'.
2158 ;; A list of buffer positions which are known not to be in a literal. This is
2159 ;; ordered with higher positions at the front of the list. Only those which
2160 ;; are less than `c-state-semi-nonlit-pos-cache-limit' are valid.
2164 ;; An upper limit on valid entries in `c-state-semi-nonlit-pos-cache'. This is
2165 ;; reduced by buffer changes, and increased by invocations of
2166 ;; `c-state-literal-at'. FIXME!!!
2169 ;; Do a parse-partial-sexp from FROM to TO, returning either
2170 ;; (STATE TYPE (BEG . END)) if TO is in a literal; or
2171 ;; (STATE) otherwise,
2172 ;; where STATE is the parsing state at TO, TYPE is the type of the literal
2173 ;; (one of 'c, 'c++, 'string) and (BEG . END) is the boundaries of the literal.
2174 ;;
2175 ;; Only elements 3 (in a string), 4 (in a comment), 5 (following a quote),
2176 ;; 7 (comment type) and 8 (start of comment/string) (and possibly 9) of
2177 ;; STATE are valid.
2180 ty)
2187 nil ; TARGETDEPTH
2188 nil ; STOPBEFORE
2189 s ; OLDSTATE
2196 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2197 ;; string, comment, or macro.
2198 ;;
2199 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2200 ;; MAY NOT contain any positions within macros, since macros are frequently
2201 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2202 ;; We cannot rely on this mechanism whilst determining a cache pos since
2203 ;; this function is also called from outwith `c-parse-state'.
2208 pos npos high-pos lit macro-beg macro-end)
2209 ;; Trim the cache to take account of buffer changes.
2221 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2225 ;; Test for being in a literal. If so, go to after it.
2232 ;; Test for being in a macro. If so, go to after it.
2247 ;; Add one extra element above HERE so as to to avoid the previous
2248 ;; expensive calculation when the next call is close to the current
2249 ;; one. This is especially useful when inside a large macro.
2254 pos))))
2257 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2258 ;; string or comment. It may be in a macro.
2263 pos npos high-pos lit macro-beg macro-end)
2264 ;; Trim the cache to take account of buffer changes.
2276 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2280 ;; Test for being in a literal. If so, go to after it.
2293 pos))))
2296 ;; If position HERE is inside a literal, return (START . END), the
2297 ;; boundaries of the literal (which may be outside the accessible bit of the
2298 ;; buffer). Otherwise, return nil.
2299 ;;
2300 ;; This function is almost the same as `c-literal-limits'. Previously, it
2301 ;; differed in that it was a lower level function, and that it rigorously
2302 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2303 ;; virtually identical to this function.
2311 ;; Return the start of the literal containing POS, or POS itself.
2313 pos))
2316 ;; Return a position outside of a string/comment/macro at or before POS.
2317 ;; STATE is the parse-partial-sexp state at POS.
2321 pos)))
2326 res))))
2328 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2329 ;; Stuff to do with point-min, and coping with any literal there.
2332 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2333 ;; narrowing is likely to affect the parens that are visible before the point.
2339 ;; These two variables define the literal, if any, containing point-min.
2340 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2341 ;; literal. If there's no literal there, they're both nil.
2345 ;; This is the earliest buffer-pos from which scanning can be done. It is
2346 ;; either the end of the literal containing point-min, or point-min itself.
2347 ;; It becomes nil if the buffer is changed earlier than this point.
2349 ;; Return the lowest valid scanning pos. This will be the end of the
2350 ;; literal enclosing point-min, or point-min itself.
2362 ;; Determine the properties of any literal containing POINT-MIN, setting the
2363 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2364 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2366 lit)
2381 c-state-point-min-lit-start nil
2382 c-state-min-scan-pos p-min)))))
2385 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2386 ;; A variable which signals a brace dessert - helpful for reducing the number
2387 ;; of fruitless backward scans.
2390 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2391 ;; that defun has searched backwards for a brace pair and not found one. Its
2392 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2393 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2394 ;; nil when at top level) and FROM is where the backward search started. It
2395 ;; is reset to nil in `c-invalidate-state-cache'.
2398 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2399 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2400 ;; list of like structure.
2402 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2403 ;; (default `c-state-cache') (or nil).
2410 ;; Return the address of the latest brace/bracket/paren (whether left or
2411 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2418 ;; Return the position just after the latest brace/bracket/paren (whether
2419 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2427 ;; From the state-cache, determine the buffer position from which we might
2428 ;; scan forward to HERE to update this cache. This position will be just
2429 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2430 ;; return the earliest position in the accessible region which isn't within
2431 ;; a literal. If the visible portion of the buffer is entirely within a
2432 ;; literal, return NIL.
2434 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2450 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2451 ;; Variables which keep track of preprocessor constructs.
2456 ;; These are the limits of the macro containing point at the previous call of
2457 ;; `c-parse-state', or nil.
2459 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2460 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2462 ;; Return a start position for building `c-state-cache' from
2463 ;; scratch. This will be at the top level, 2 defuns back.
2465 ;; Go back 2 bods, but ignore any bogus positions returned by
2466 ;; beginning-of-defun (i.e. open paren in column zero).
2476 ;; Return the position of the opening paren/brace/bracket before HERE- which
2477 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2478 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2479 ;;
2480 ;; ............................................
2481 ;; | |
2482 ;; ( [ ( .........#macro.. ) ( ) ] )
2483 ;; ^ ^ ^ ^
2484 ;; | | | |
2485 ;; return HERE- HERE+ TOP
2486 ;;
2487 ;; If there aren't enough opening paren/brace/brackets, return the position
2488 ;; of the outermost one found, or HERE- if there are none. If there are no
2489 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2490 ;; must not be inside literals. Only the accessible portion of the buffer
2491 ;; will be scanned.
2493 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2494 ;; `here'. Go round the next loop each time we pass over such a ")". These
2495 ;; probably match "("s before `here-'.
2507 ;; PART 2: Scan back before `here-' searching for the "("s
2508 ;; matching/mismatching the ")"s found above. We only need to direct the
2509 ;; caller to scan when we've encountered unmatched right parens.
2518 pos))
2521 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2522 ;; to minimize the amount of scanning. HERE is the pertinent position in
2523 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2524 ;; its head trimmed) is known to be good, or nil if there is no such
2525 ;; position.
2526 ;;
2527 ;; The return value is a list, one of the following:
2528 ;;
2529 ;; o - ('forward CACHE-POS START-POINT) - scan forward from START-POINT,
2530 ;; which is not less than CACHE-POS.
2531 ;; o - ('backward CACHE-POS nil) - scan backwards (from HERE).
2532 ;; o - ('BOD nil START-POINT) - scan forwards from START-POINT, which is at the
2533 ;; top level.
2534 ;; o - ('IN-LIT nil nil) - point is inside the literal containing point-min.
2535 ;; , where CACHE-POS is the highest position recorded in `c-state-cache' at
2536 ;; or below HERE.
2538 BOD-pos ; position of 2nd BOD before HERE.
2539 strategy ; 'forward, 'backward, 'BOD, or 'IN-LIT.
2540 start-point
2546 start-point nil
2547 cache-pos nil
2559 start-point cache-pos)))
2561 ;; Might we be better off starting from the top level, two defuns back,
2562 ;; instead?
2567 start-point BOD-pos)))
2570 strategy
2575 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2576 ;; Routines which change `c-state-cache' and associated values.
2578 ;; The region (more precisely, point-min) has changed since we
2579 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2581 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2582 ;; It would be possible to do a better job here and recalculate the top
2583 ;; only.
2587 c-state-cache-good-pos c-state-min-scan-pos
2588 c-state-brace-pair-desert nil))
2590 ;; point-min has MOVED FORWARD.
2592 ;; Is the new point-min inside a (different) literal?
2597 ;; Cut off a bit of the tail from `c-state-cache'.
2599 pa)
2607 c-state-cache-good-pos c-state-min-scan-pos)
2610 )))
2615 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2616 ;; of nesting (not necessarily immediately preceding), push a cons onto
2617 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2618 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2619 ;; UPPER-LIM.
2620 ;;
2621 ;; Return non-nil when this has been done.
2622 ;;
2623 ;; The situation it copes with is this transformation:
2624 ;;
2625 ;; OLD: { (.) {...........}
2626 ;; ^ ^
2627 ;; FROM HERE
2628 ;;
2629 ;; NEW: { {....} (.) {.........
2630 ;; ^ ^ ^
2631 ;; LOWER BRACE PAIR HERE or HERE
2632 ;;
2633 ;; This routine should be fast. Since it can get called a LOT, we maintain
2634 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2635 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2639 new-cons
2647 ;; If we're essentially repeating a fruitless search, just give up.
2651 ;; DESERT-LIM. Only search what we absolutely need to,
2656 ;; CACHE-LIM. This limit will be necessary when an opening
2657 ;; paren at `cache-pos' has just had its matching close paren
2658 ;; inserted. `cache-pos' continues to be a search bound, even
2659 ;; though the algorithm below would skip over the new paren
2660 ;; pair.
2671 ;; In the next pair of nested loops, the inner one moves back past a
2672 ;; pair of (mis-)matching parens or brackets; the outer one moves
2673 ;; back over a sequence of unmatched close brace/paren/bracket each
2674 ;; time round.
2688 ; brace, ignore it.
2691 ;; We've found the desired brace-pair.
2703 ;; We haven't found a brace pair. Record this in the cache.
2707 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2708 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2709 ;; "push" "a" brace pair onto `c-state-cache'.
2710 ;;
2711 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2712 ;; otherwise push it normally.
2713 ;;
2714 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2715 ;; latter is inside a macro, not being a macro containing
2716 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2717 ;; base pair. This latter case is assumed to be rare.
2718 ;;
2719 ;; Note: POINT is not preserved in this routine.
2729 c-state-cache)))
2730 ;; N.B. This defsubst codes one method for the simple, normal case,
2731 ;; and a more sophisticated, slower way for the general case. Don't
2732 ;; eliminate this defsubst - it's a speed optimization.
2736 ;; Scan the buffer from FROM to (point-max), adding elements into
2737 ;; `c-state-cache' for braces etc. Return a candidate for
2738 ;; `c-state-cache-good-pos'.
2739 ;;
2740 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2741 ;; any. Typically, it is immediately after it. It must not be inside a
2742 ;; literal.
2751 ; Is actually the pos. to scan for a (/{/[ from,
2752 ; which sometimes is after a silly )/}/].
2755 ; good-pos.
2761 ;; Each time round the following loop, we enter a successively deeper
2762 ;; level of brace/paren nesting. (Except sometimes we "continue at
2763 ;; the existing level".) `pa+1' is a pos inside an opening
2764 ;; brace/paren/bracket, usually just after it.
2767 ;; Each time round the next loop moves forward over an opening then
2768 ;; a closing brace/bracket/paren. This loop is white hot, so it
2769 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2770 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2771 ;; call of `scan-lists' signals an error, which happens when there
2772 ;; are no more b/b/p's to scan.
2783 ;; We've just entered a deeper nesting level.
2785 ;; Insert the brace pair (if present) and the single open
2786 ;; paren/brace/bracket into `c-state-cache' It cannot be
2787 ;; inside a macro, except one around point, because of what
2788 ;; `c-neutralize-syntax-in-CPP' has done.
2790 ;; Insert the opening brace/bracket/paren position.
2792 ;; Clear admin stuff for the next more nested part of the scan.
2796 ;; All open p/b/b's at this nesting level, if any, have probably
2797 ;; been closed by matching/mismatching ones. We're probably
2798 ;; finished - we just need to check for having found an
2799 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2800 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2803 ;; Record the final, innermost, brace-pair if there is one.
2806 ;; Determine a good pos
2814 here-bol))
2823 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2824 ;; not be in it when it is amended for position (point-max).
2825 ;; Additionally, the "outermost" open-brace entry before (point-max)
2826 ;; will be converted to a cons if the matching close-brace is scanned.
2827 ;;
2828 ;; GOOD-POS is a "maximal" "safe position" - there must be no open
2829 ;; parens/braces/brackets between GOOD-POS and (point-max).
2830 ;;
2831 ;; As a second thing, calculate the result of parse-partial-sexp at
2832 ;; PPS-POINT, w.r.t. GOOD-POS. The motivation here is that
2833 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2834 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2835 ;; needs to be FAST).
2836 ;;
2837 ;; Return a list (GOOD-POS SCAN-BACK-POS PPS-STATE), where
2838 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2839 ;; to be good (we aim for this to be as high as possible);
2840 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2841 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2842 ;; position to scan backwards from.
2843 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2853 ; or nil
2860 ; (point-max)), or nil.
2866 pps-state ; Will be 9 or 10 elements long.
2867 pos
2868 upper-lim ; ,beyond which `c-state-cache' entries are removed
2869 scan-back-pos
2870 pair-beg pps-point-state target-depth)
2872 ;; Remove entries beyond (point-max). Also remove any entries inside
2873 ;; a macro, unless (point-max) is in the same macro.
2882 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2883 ;; RBrace and indicate we'll need to search backwards for a previous
2884 ;; brace pair.
2891 ;; The next loop jumps forward out of a nested level of parens each
2892 ;; time round; the corresponding elements in `c-state-cache' are
2893 ;; removed. `pos' is just after the brace-pair or the open paren at
2894 ;; (car c-state-cache). There can be no open parens/braces/brackets
2895 ;; between `good-pos'/`good-pos-actual-macro-start' and (point-max),
2896 ;; due to the interface spec to this function.
2899 in-macro-start)))
2901 ; marked by a `category' text property.
2902 good-pos))
2911 ;; Do nothing when we've merely reached pps-point.
2912 )
2914 ;; Scan!
2918 target-depth
2919 nil pps-state))
2929 ;; We've just moved out of the paren pair containing the brace-pair
2930 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2931 ;; and is potentially where the open brace of a cons in
2932 ;; c-state-cache will be.
2938 ; containing Lparen
2942 ; together with enclosed brace pair.
2943 ;; (t nil) ; Ignore an unmated Rparen.
2944 )))
2948 nil nil ; TARGETDEPTH, STOPBEFORE
2949 pps-state)))
2951 ;; If the last paren pair we moved out of was actually a brace pair,
2952 ;; insert it into `c-state-cache'.
2957 c-state-cache)))
2962 ;; Strip stale elements of `c-state-cache' by moving backwards through the
2963 ;; buffer, and inform the caller of the scenario detected.
2964 ;;
2965 ;; HERE is the position we're setting `c-state-cache' for.
2966 ;; CACHE-POS is just after the latest recorded position in `c-state-cache'
2967 ;; before HERE, or a position at or near point-min which isn't in a
2968 ;; literal.
2969 ;;
2970 ;; This function must only be called only when (> `c-state-cache-good-pos'
2971 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
2972 ;; optimized to eliminate (or minimize) scanning between these two
2973 ;; positions.
2974 ;;
2975 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
2976 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
2977 ;; could become so after missing elements are inserted into
2978 ;; `c-state-cache'. This is JUST AFTER an opening or closing
2979 ;; brace/paren/bracket which is already in `c-state-cache' or just before
2980 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
2981 ;; before `here''s line, or the start of the literal containing it.
2982 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
2983 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
2984 ;; to scan backwards from.
2985 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
2986 ;; POS and HERE which aren't recorded in `c-state-cache'.
2987 ;;
2988 ;; The comments in this defun use "paren" to mean parenthesis or square
2989 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
2990 ;;
2991 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
2992 ;; | | | | | |
2993 ;; CP E here D C good
2995 pa ren ; positions of "(" and ")"
2996 dropped-cons ; whether the last element dropped from `c-state-cache'
2997 ; was a cons (representing a brace-pair)
2998 good-pos ; see above.
2999 lit ; (START . END) of a literal containing some point.
3000 here-lit-start here-lit-end ; bounds of literal containing `here'
3001 ; or `here' itself.
3006 ;; Remove completely irrelevant entries from `c-state-cache'.
3012 ;; At this stage, (> pos here);
3013 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3018 ;; CASE 1: The top of the cache is a brace pair which now encloses
3019 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3020 ;; knowledge of what's inside these braces, we have no alternative but
3021 ;; to direct the caller to scan the buffer from the opening brace.
3025 ; entry into a { entry, so the caller needs to
3026 ; search for a brace pair before the {.
3028 ;; `here' might be inside a literal. Check for this.
3033 ;; Has `here' just "newly entered" a macro?
3046 ;; `here' might be nested inside any depth of parens (or brackets but
3047 ;; not braces). Scan backwards to find the outermost such opening
3048 ;; paren, if there is one. This will be the scan position to return.
3055 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3056 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3057 ;; a brace pair preceding this, it will already be in `c-state-cache',
3058 ;; unless there was a brace pair after it, i.e. there'll only be one to
3059 ;; scan for if we've just deleted one.
3062 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3063 ;; Further forward scanning isn't needed, but we still need to find a
3064 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3074 ;; CASE 3: After a }/)/] before `here''s BOL.
3078 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3079 ;; literal containing it.
3084 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3085 ;; Externally visible routines.
3089 c-state-cache-good-pos 1
3090 c-state-nonlit-pos-cache nil
3091 c-state-nonlit-pos-cache-limit 1
3092 c-state-brace-pair-desert nil
3093 c-state-point-min 1
3094 c-state-point-min-lit-type nil
3095 c-state-point-min-lit-start nil
3096 c-state-min-scan-pos 1
3097 c-state-old-cpp-beg nil
3098 c-state-old-cpp-end nil)
3101 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3102 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3103 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3104 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3105 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3106 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3107 ;; (defun c-state-dump ()
3108 ;; ;; For debugging.
3109 ;; ;(message
3110 ;; (concat
3111 ;; (c-sc-qde c-state-cache)
3112 ;; (c-sc-de c-state-cache-good-pos)
3113 ;; (c-sc-qde c-state-nonlit-pos-cache)
3114 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3115 ;; (c-sc-qde c-state-brace-pair-desert)
3116 ;; (c-sc-de c-state-point-min)
3117 ;; (c-sc-de c-state-point-min-lit-type)
3118 ;; (c-sc-de c-state-point-min-lit-start)
3119 ;; (c-sc-de c-state-min-scan-pos)
3120 ;; (c-sc-de c-state-old-cpp-beg)
3121 ;; (c-sc-de c-state-old-cpp-end)))
3122 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3125 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3126 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3127 ;; left in a consistent state.
3128 ;;
3129 ;; This is much like `c-whack-state-after', but it never changes a paren
3130 ;; pair element into an open paren element. Doing that would mean that the
3131 ;; new open paren wouldn't have the required preceding paren pair element.
3132 ;;
3133 ;; This function is called from c-after-change.
3135 ;; The caches of non-literals:
3141 ;; `c-state-cache':
3142 ;; Case 1: if `here' is in a literal containing point-min, everything
3143 ;; becomes (or is already) nil.
3147 c-state-cache-good-pos nil
3148 c-state-min-scan-pos nil)
3150 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3151 ;; below `here'. To maintain its consistency, we may need to insert a new
3152 ;; brace pair.
3154 too-high-pa ; recorded {/(/[ next above here, or nil.
3155 dropped-cons ; was the last removed element a brace pair?
3156 pa)
3157 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3164 ;; Do we need to add in an earlier brace pair, having lopped one off?
3171 ;; The brace-pair desert marker:
3179 ;; Find and record all noteworthy parens between some good point earlier in
3180 ;; the file and point. That good point is at least the beginning of the
3181 ;; top-level construct we are in, or the beginning of the preceding
3182 ;; top-level construct if we aren't in one.
3183 ;;
3184 ;; The returned value is a list of the noteworthy parens with the last one
3185 ;; first. If an element in the list is an integer, it's the position of an
3186 ;; open paren (of any type) which has not been closed before the point. If
3187 ;; an element is a cons, it gives the position of a closed BRACE paren
3188 ;; pair[*]; the car is the start brace position and the cdr is the position
3189 ;; following the closing brace. Only the last closed brace paren pair
3190 ;; before each open paren and before the point is recorded, and thus the
3191 ;; state never contains two cons elements in succession. When a close brace
3192 ;; has no matching open brace (e.g., the matching brace is outside the
3193 ;; visible region), it is not represented in the returned value.
3194 ;;
3195 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3196 ;; This defun explicitly treats mismatching parens/braces/brackets as
3197 ;; matching. It is the open brace which makes it a "brace" pair.
3198 ;;
3199 ;; If POINT is within a macro, open parens and brace pairs within
3200 ;; THIS macro MIGHT be recorded. This depends on whether their
3201 ;; syntactic properties have been suppressed by
3202 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3203 ;;
3204 ;; Currently no characters which are given paren syntax with the
3205 ;; syntax-table property are recorded, i.e. angle bracket arglist
3206 ;; parens are never present here. Note that this might change.
3207 ;;
3208 ;; BUG: This function doesn't cope entirely well with unbalanced
3209 ;; parens in macros. (2008-12-11: this has probably been resolved
3210 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3211 ;; following case the brace before the macro isn't balanced with the
3212 ;; one after it:
3213 ;;
3214 ;; {
3215 ;; #define X {
3216 ;; }
3217 ;;
3218 ;; Note to maintainers: this function DOES get called with point
3219 ;; within comments and strings, so don't assume it doesn't!
3220 ;;
3221 ;; This function might do hidden buffer changes.
3224 strategy ; 'forward, 'backward etc..
3225 ;; Candidate positions to start scanning from:
3226 cache-pos ; highest position below HERE already existing in
3227 ; cache (or 1).
3228 good-pos
3229 start-point
3230 bopl-state
3231 res
3233 ;; If POINT-MIN has changed, adjust the cache
3237 ;; Strategy?
3245 c-state-cache-good-pos start-point))
3247 ;; SCAN!
3256 ; start-point)
3265 good-pos)))
3274 scan-backward-pos))
3279 good-pos)))
3283 c-state-cache-good-pos nil)))))
3285 c-state-cache)
3288 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3289 ;;
3290 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3291 ;; of all parens in preprocessor constructs, except for any such construct
3292 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3293 ;; worrying further about macros and template delimiters.
3298 c-state-old-cpp-beg
3305 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3306 ;; description of the functionality and return value.
3307 ;;
3308 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3309 ;; of all parens in preprocessor constructs, except for any such construct
3310 ;; containing point. We can then call `c-parse-state-1' without worrying
3311 ;; further about macros and template delimiters.
3318 here-cpp-beg)
3320 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3321 ;; subsystem.
3326 here-cpp-beg here-cpp-end
3332 )))
3334 ;; Debug tool to catch cache inconsistencies. This is called from
3335 ;; 000tests.el.
3348 c-state-cache-good-pos
3349 c-state-nonlit-pos-cache
3350 c-state-nonlit-pos-cache-limit
3351 c-state-brace-pair-desert
3352 c-state-point-min
3353 c-state-point-min-lit-type
3354 c-state-point-min-lit-start
3355 c-state-min-scan-pos
3356 c-state-old-cpp-beg
3357 c-state-old-cpp-end))))
3382 ;; The cache can actually go further back due to the ad-hoc way
3383 ;; the first paren is found, so try to whack off a bit of its
3384 ;; start before complaining.
3385 ;; (save-excursion
3386 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3387 ;; (c-beginning-of-defun-1)
3388 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3389 ;; (c-beginning-of-defun-1))
3390 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3391 ;; (message (concat "c-parse-state inconsistency at %s: "
3392 ;; "using cache: %s, from scratch: %s")
3393 ;; here res1 res2)))
3396 here res1 res2)
3400 res1))
3406 'c-debug-parse-state
3412 \f
3414 ;; Whack off any state information from PAREN-STATE which lies
3415 ;; before BUFPOS. Not destructive on PAREN-STATE.
3418 car)
3429 ;; Whack off any state information from PAREN-STATE which lies at or
3430 ;; after BUFPOS. Not destructive on PAREN-STATE.
3435 ;; just check the car, because in a balanced brace
3436 ;; expression, it must be impossible for the corresponding
3437 ;; close brace to be before point, but the open brace to
3438 ;; be after.
3440 nil ; whack it off
3442 ;; its possible that the open brace is before
3443 ;; bufpos, but the close brace is after. In that
3444 ;; case, convert this to a non-cons element. The
3445 ;; rest of the state is before bufpos, so we're
3446 ;; done.
3448 ;; we know that both the open and close braces are
3449 ;; before bufpos, so we also know that everything else
3450 ;; on state is before bufpos.
3453 nil ; whack it off
3454 ;; it's before bufpos, so everything else should too.
3457 nil)))
3460 ;; Return the bufpos of the innermost enclosing open paren before
3461 ;; bufpos, or nil if none was found.
3471 enclosingp))
3474 ;; Return the bufpos of the outermost enclosing open paren, or nil
3475 ;; if none was found.
3482 pos))
3485 ;; Return the closest "safe" position recorded on PAREN-STATE that
3486 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3487 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3488 ;; find the closest limit before a given limit that might be nil.
3489 ;;
3490 ;; A "safe" position is a position at or after a recorded open
3491 ;; paren, or after a recorded close paren. The returned position is
3492 ;; thus either the first position after a close brace, or the first
3493 ;; position after an enclosing paren, or at the enclosing paren in
3494 ;; case BUFPOS is immediately after it.
3504 ;; See below.
3507 ;; elem is the position at and not after the opening paren, so
3508 ;; we can go forward one more step unless it's equal to
3509 ;; bufpos. This is useful in some cases avoid an extra paren
3510 ;; level between the safe position and bufpos.
3515 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3516 ;; goes to the closest previous point that is known to be outside
3517 ;; any string literal or comment. `c-state-cache' is used if it has
3518 ;; a position in the vicinity.
3520 elem
3523 ;; Note: Similar code in `c-safe-position'. The
3524 ;; difference is that we accept a safe position at
3525 ;; the point and don't bother to go forward past open
3526 ;; parens.
3541 ;; The position is far back. Try `c-beginning-of-defun-1'
3542 ;; (although we can't be entirely sure it will go to a position
3543 ;; outside a comment or string in current emacsen). FIXME:
3544 ;; Consult `syntax-ppss' here.
3549 \f
3550 ;; Tools for scanning identifiers and other tokens.
3553 "Return non-nil if the point is on or directly after an identifier.
3554 Keywords are recognized and not considered identifiers. If an
3555 identifier is detected, the returned value is its starting position.
3556 If an identifier ends at the point and another begins at it \(can only
3557 happen in Pike) then the point for the preceding one is returned.
3559 Note that this function might do hidden buffer changes. See the
3560 comment at the start of cc-engine.el for more info."
3562 ;; FIXME: Shouldn't this function handle "operator" in C++?
3569 ;; Check for a normal (non-keyword) identifier.
3575 ;; Handle the `<operator> syntax in Pike.
3579 t
3586 ;; Handle the "operator +" syntax in C++.
3603 )))
3606 ;; If the point is at the end of a symbol then skip backward to the
3607 ;; beginning of it. Don't move otherwise. Return non-nil if point
3608 ;; moved.
3609 ;;
3610 ;; This function might do hidden buffer changes.
3613 ;; Handle the `<operator> syntax in Pike.
3619 t
3621 nil)))))
3624 ;; Move to the beginning of the current token. Do not move if not
3625 ;; in the middle of one. BACK-LIMIT may be used to bound the
3626 ;; backward search; if given it's assumed to be at the boundary
3627 ;; between two tokens. Return non-nil if the point is moved, nil
3628 ;; otherwise.
3629 ;;
3630 ;; This function might do hidden buffer changes.
3637 ;; `c-nonsymbol-token-regexp' should always match
3638 ;; since we've skipped backward over punctuator
3639 ;; or paren syntax, but consume one char in case
3640 ;; it doesn't so that we don't leave point before
3641 ;; some earlier incorrect token.
3648 ;; Move to the end of the current token. Do not move if not in the
3649 ;; middle of one. BACK-LIMIT may be used to bound the backward
3650 ;; search; if given it's assumed to be at the boundary between two
3651 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3652 ;;
3653 ;; This function might do hidden buffer changes.
3661 ;; `c-nonsymbol-token-regexp' should always match since
3662 ;; we've skipped backward over punctuator or paren
3663 ;; syntax, but move forward in case it doesn't so that
3664 ;; we don't leave point earlier than we started with.
3680 "Move forward by tokens.
3681 A token is defined as all symbols and identifiers which aren't
3683 treated properly). Point is always either left at the beginning of a
3684 token or not moved at all. COUNT specifies the number of tokens to
3685 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3686 moves to the next token beginning only if not already at one. If
3687 BALANCED is true, move over balanced parens, otherwise move into them.
3688 Also, if BALANCED is true, never move out of an enclosing paren.
3690 LIMIT sets the limit for the movement and defaults to the point limit.
3691 The case when LIMIT is set in the middle of a token, comment or macro
3692 is handled correctly, i.e. the point won't be left there.
3694 Return the number of tokens left to move \(positive or negative). If
3695 BALANCED is true, a move over a balanced paren counts as one. Note
3696 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3697 be returned. Thus, a return value of 0 guarantees that point is at
3698 the requested position and a return value less \(without signs) than
3699 COUNT guarantees that point is at the beginning of some token.
3701 Note that this function might do hidden buffer changes. See the
3702 comment at the start of cc-engine.el for more info."
3709 c-jump-syntax-balanced
3710 c-jump-syntax-unbalanced))
3715 ;; If count is zero we should jump if in the middle of a token.
3722 ;; Skip whitespace. Count this as a move if we did in
3723 ;; fact move.
3727 ;; Moved out of bounds. Make sure the returned count isn't zero.
3732 ;; Use `condition-case' to avoid having the limit tests
3733 ;; inside the loop.
3741 t)
3744 t)
3745 ;; `c-nonsymbol-token-regexp' above should always
3746 ;; match if there are correct tokens. Try to
3747 ;; widen to see if the limit was set in the
3748 ;; middle of one, else fall back to treating
3749 ;; the offending thing as a one character token.
3754 nil)
3757 t))))
3767 count)))
3770 "Move backward by tokens.
3771 See `c-forward-token-2' for details."
3779 c-jump-syntax-balanced
3780 c-jump-syntax-unbalanced))
3784 ;; The count is zero so try to skip to the beginning of the
3785 ;; current token.
3789 ;; The limit is inside the same token, so return 1.
3792 ;; We're not in the middle of a token. If there's
3793 ;; whitespace after the point then we must move backward,
3794 ;; so set count to 1 in that case.
3796 ;; If we're looking at a '#' that might start a cpp
3797 ;; directive then we have to do a more elaborate check.
3810 ;; Use `condition-case' to avoid having to check for buffer
3811 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3820 ;; This can be very inefficient if there's a long
3821 ;; sequence of operator tokens without any separation.
3822 ;; That doesn't happen in practice, anyway.
3832 count)))
3835 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3837 characters are jumped over character by character. This function is
3838 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3843 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3845 characters are jumped over character by character. This function is
3846 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3850 \f
3851 ;; Tools for doing searches restricted to syntactically relevant text.
3854 paren-level not-inside-token
3855 lookbehind-submatch)
3856 "Like `re-search-forward', but only report matches that are found
3857 in syntactically significant text. I.e. matches in comments, macros
3858 or string literals are ignored. The start point is assumed to be
3859 outside any comment, macro or string literal, or else the content of
3860 that region is taken as syntactically significant text.
3862 If PAREN-LEVEL is non-nil, an additional restriction is added to
3863 ignore matches in nested paren sexps. The search will also not go
3864 outside the current list sexp, which has the effect that if the point
3865 should be moved to BOUND when no match is found \(i.e. NOERROR is
3866 neither nil nor t), then it will be at the closing paren if the end of
3867 the current list sexp is encountered first.
3869 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3870 ignored. Things like multicharacter operators and special symbols
3872 constants.
3874 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3875 subexpression in REGEXP. The end of that submatch is used as the
3876 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
3877 isn't used or if that subexpression didn't match then the start
3879 subexpression is never tested before the starting position, so it
3883 subexpression can match the end of nonwhite syntactic whitespace,
3884 i.e. the end of comments or cpp directives. This since the function
3885 skips over such things before resuming the search. It's on the other
3887 matches syntactic whitespace.
3889 Bug: Unbalanced parens inside cpp directives are currently not handled
3890 correctly \(i.e. they don't get ignored as they should) when
3891 PAREN-LEVEL is set.
3893 Note that this function might do hidden buffer changes. See the
3894 comment at the start of cc-engine.el for more info."
3899 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
3902 tmp
3903 ;; Start position for the last search.
3904 search-pos
3905 ;; The `parse-partial-sexp' state between the start position
3906 ;; and the point.
3907 state
3908 ;; The current position after the last state update. The next
3909 ;; `parse-partial-sexp' continues from here.
3911 ;; The position at which to check the state and the state
3912 ;; there. This is separate from `state-pos' since we might
3913 ;; need to back up before doing the next search round.
3914 check-pos check-state
3915 ;; Last position known to end a token.
3917 ;; Set when a valid match is found.
3918 found)
3936 state-pos check-pos paren-level nil state))
3938 check-state state))
3940 ;; NOTE: If we got a look behind subexpression and get
3941 ;; an insignificant match in something that isn't
3942 ;; syntactic whitespace (i.e. strings or in nested
3943 ;; parentheses), then we can never skip more than a
3944 ;; single character from the match start position
3945 ;; (i.e. `state-pos' here) before continuing the
3946 ;; search. That since the look behind subexpression
3947 ;; might match the end of the insignificant region in
3948 ;; the next search.
3952 ;; Match inside a line comment. Skip to eol. Use
3953 ;; `re-search-forward' instead of `skip-chars-forward' to get
3954 ;; the right bound behavior.
3958 ;; Match inside a block comment. Skip to the '*/'.
3965 ;; Match in the middle of the opener of a block or line
3966 ;; comment.
3971 ;; The last `parse-partial-sexp' above might have
3972 ;; stopped short of the real check position if the end
3973 ;; of the current sexp was encountered in paren-level
3974 ;; mode. The checks above are always false in that
3975 ;; case, and since they can do better skipping in
3976 ;; lookbehind-submatch mode, we do them before
3977 ;; checking the paren level.
3981 ;; Check the paren level first since we're short of the
3982 ;; syntactic checking position if the end of the
3983 ;; current sexp was encountered by `parse-partial-sexp'.
3986 ;; Inside a nested paren sexp.
3988 ;; See the NOTE above.
3990 ;; Skip out of the paren quickly.
3994 ;; Have exited the current paren sexp.
3997 ;; The last `parse-partial-sexp' call above
3998 ;; has left us just after the closing paren
3999 ;; in this case, so we can modify the bound
4000 ;; to leave the point at the right position
4001 ;; upon return.
4003 nil)
4007 ;; Match inside a string.
4010 ;; See the NOTE above.
4012 ;; Skip to the end of the string before continuing.
4021 continue)))
4026 ;; Match inside a macro. Skip to the end of it.
4034 (< check-pos
4040 ;; Inside a token.
4042 ;; See the NOTE above.
4047 ;; A real match.
4049 nil)))
4051 ;; Should loop to search again, but take care to avoid
4052 ;; looping on the same spot.
4056 nil
4059 t))))
4065 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4072 ;; Search failed. Set point as appropriate.
4076 nil)))
4081 ;; Return the start of the literal point is in, or nil.
4082 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4083 ;; bound in the caller.
4085 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4086 ;; if it's outside comments and strings.
4089 ;; Pick a safe position as close to the point as possible.
4090 ;;
4091 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4092 ;; position.
4105 ;; Cache positions along the way to use if we have to back up more. We
4106 ;; cache every closing paren on the same level. If the paren cache is
4107 ;; relevant in this region then we're typically already on the same
4108 ;; level as the target position. Note that we might cache positions
4109 ;; after opening parens in case safe-pos is in a nested list. That's
4110 ;; both uncommon and harmless.
4118 ;; If the state contains the start of the containing sexp we cache that
4119 ;; position too, so that parse-partial-sexp in the next run has a bigger
4120 ;; chance of starting at the same level as the target position and thus
4121 ;; will get more good safe positions into the list.
4127 ;; Inside string or comment. Continue search at the
4128 ;; beginning of it.
4132 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4133 i.e. don't stop at positions inside syntactic whitespace or string
4134 literals. Preprocessor directives are also ignored, with the exception
4135 of the one that the point starts within, if any. If LIMIT is given,
4136 it's assumed to be at a syntactically relevant position.
4138 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4139 sexps, and the search will also not go outside the current paren sexp.
4140 However, if LIMIT or the buffer limit is reached inside a nested paren
4141 then the point will be left at the limit.
4143 Non-nil is returned if the point moved, nil otherwise.
4145 Note that this function might do hidden buffer changes. See the
4146 comment at the start of cc-engine.el for more info."
4149 state-2
4150 ;; A list of syntactically relevant positions in descending
4151 ;; order. It's used to avoid scanning repeatedly over
4152 ;; potentially large regions with `parse-partial-sexp' to verify
4153 ;; each position. Used in `c-ssb-lit-begin'
4154 safe-pos-list
4155 ;; The result from `c-beginning-of-macro' at the start position or the
4156 ;; start position itself if it isn't within a macro. Evaluated on
4157 ;; demand.
4158 start-macro-beg
4159 ;; The earliest position after the current one with the same paren
4160 ;; level. Used only when `paren-level' is set.
4161 lit-beg
4166 ;; The next loop "tries" to find the end point each time round,
4167 ;; loops when it hasn't succeeded.
4175 ;; Don't stop inside a literal
4178 t)
4186 ;; Not at the right level.
4189 ;; We stop above if we go out of a paren.
4190 ;; Now check whether it precedes or is
4191 ;; nested in the starting sexp.
4195 pps-end-pos paren-level-pos
4196 nil nil state-2))
4199 ;; We've stopped short of the starting position
4200 ;; so the hit was inside a nested list. Go up
4201 ;; until we are at the right level.
4210 nil)
4211 t))
4214 nil))
4216 ;; The hit was outside the list at the start
4217 ;; position. Go to the start of the list and exit.
4219 nil))
4222 ;; Inside a macro.
4230 t
4232 ;; It's inside the same macro we started in so it's
4233 ;; a relevant match.
4235 nil))))))
4239 ;; Skip syntactic ws afterwards so that we don't stop at the
4240 ;; end of a comment if `skip-chars' is something like "^/".
4244 ;; We might want to extend this with more useful return values in
4245 ;; the future.
4248 ;; The following is an alternative implementation of
4249 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4250 ;; track of the syntactic context. It turned out to be generally
4251 ;; slower than the one above which uses forward checks from earlier
4252 ;; safe positions.
4253 ;;
4254 ;;(defconst c-ssb-stop-re
4255 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4256 ;; ;; stop at to avoid going into comments and literals.
4257 ;; (concat
4258 ;; ;; Match comment end syntax and string literal syntax. Also match
4259 ;; ;; '/' for block comment endings (not covered by comment end
4260 ;; ;; syntax).
4261 ;; "\\s>\\|/\\|\\s\""
4262 ;; (if (memq 'gen-string-delim c-emacs-features)
4263 ;; "\\|\\s|"
4264 ;; "")
4265 ;; (if (memq 'gen-comment-delim c-emacs-features)
4266 ;; "\\|\\s!"
4267 ;; "")))
4268 ;;
4269 ;;(defconst c-ssb-stop-paren-re
4270 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4271 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4272 ;;
4273 ;;(defconst c-ssb-sexp-end-re
4274 ;; ;; Regexp matching the ending syntax of a complex sexp.
4275 ;; (concat c-string-limit-regexp "\\|\\s)"))
4276 ;;
4277 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4278 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4279 ;;i.e. don't stop at positions inside syntactic whitespace or string
4280 ;;literals. Preprocessor directives are also ignored. However, if the
4281 ;;point is within a comment, string literal or preprocessor directory to
4282 ;;begin with, its contents is treated as syntactically relevant chars.
4283 ;;If LIMIT is given, it limits the backward search and the point will be
4284 ;;left there if no earlier position is found.
4285 ;;
4286 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4287 ;;sexps, and the search will also not go outside the current paren sexp.
4288 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4289 ;;then the point will be left at the limit.
4290 ;;
4291 ;;Non-nil is returned if the point moved, nil otherwise.
4292 ;;
4293 ;;Note that this function might do hidden buffer changes. See the
4294 ;;comment at the start of cc-engine.el for more info."
4295 ;;
4296 ;; (save-restriction
4297 ;; (when limit
4298 ;; (narrow-to-region limit (point-max)))
4299 ;;
4300 ;; (let ((start (point)))
4301 ;; (catch 'done
4302 ;; (while (let ((last-pos (point))
4303 ;; (stop-pos (progn
4304 ;; (skip-chars-backward skip-chars)
4305 ;; (point))))
4306 ;;
4307 ;; ;; Skip back over the same region as
4308 ;; ;; `skip-chars-backward' above, but keep to
4309 ;; ;; syntactically relevant positions.
4310 ;; (goto-char last-pos)
4311 ;; (while (and
4312 ;; ;; `re-search-backward' with a single char regexp
4313 ;; ;; should be fast.
4314 ;; (re-search-backward
4315 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4316 ;; stop-pos 'move)
4317 ;;
4318 ;; (progn
4319 ;; (cond
4320 ;; ((looking-at "\\s(")
4321 ;; ;; `paren-level' is set and we've found the
4322 ;; ;; start of the containing paren.
4323 ;; (forward-char)
4324 ;; (throw 'done t))
4325 ;;
4326 ;; ((looking-at c-ssb-sexp-end-re)
4327 ;; ;; We're at the end of a string literal or paren
4328 ;; ;; sexp (if `paren-level' is set).
4329 ;; (forward-char)
4330 ;; (condition-case nil
4331 ;; (c-backward-sexp)
4332 ;; (error
4333 ;; (goto-char limit)
4334 ;; (throw 'done t))))
4335 ;;
4336 ;; (t
4337 ;; (forward-char)
4338 ;; ;; At the end of some syntactic ws or possibly
4339 ;; ;; after a plain '/' operator.
4340 ;; (let ((pos (point)))
4341 ;; (c-backward-syntactic-ws)
4342 ;; (if (= pos (point))
4343 ;; ;; Was a plain '/' operator. Go past it.
4344 ;; (backward-char)))))
4345 ;;
4346 ;; (> (point) stop-pos))))
4347 ;;
4348 ;; ;; Now the point is either at `stop-pos' or at some
4349 ;; ;; position further back if `stop-pos' was at a
4350 ;; ;; syntactically irrelevant place.
4351 ;;
4352 ;; ;; Skip additional syntactic ws so that we don't stop
4353 ;; ;; at the end of a comment if `skip-chars' is
4354 ;; ;; something like "^/".
4355 ;; (c-backward-syntactic-ws)
4356 ;;
4357 ;; (< (point) stop-pos))))
4358 ;;
4359 ;; ;; We might want to extend this with more useful return values
4360 ;; ;; in the future.
4361 ;; (/= (point) start))))
4363 \f
4364 ;; Tools for handling comments and string literals.
4367 "Return the type of literal point is in, if any.
4368 The return value is `c' if in a C-style comment, `c++' if in a C++
4369 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4370 is non-nil and in a preprocessor line, or nil if somewhere else.
4371 Optional LIM is used as the backward limit of the search. If omitted,
4372 or nil, `c-beginning-of-defun' is used.
4374 The last point calculated is cached if the cache is enabled, i.e. if
4375 `c-in-literal-cache' is bound to a two element vector.
4377 Note that this function might do hidden buffer changes. See the
4378 comment at the start of cc-engine.el for more info."
4389 "Return a cons of the beginning and end positions of the comment or
4390 string surrounding point (including both delimiters), or nil if point
4392 to start parsing from. If NEAR is non-nil, then the limits of any
4394 spaces and tabs between point and the literal. The search for such a
4395 literal is done first in forward direction. If NOT-IN-DELIMITER is
4396 non-nil, the case when point is inside a starting delimiter won't be
4397 recognized. This only has effect for comments which have starting
4398 delimiters with more than one character.
4400 Note that this function might do hidden buffer changes. See the
4401 comment at the start of cc-engine.el for more info."
4418 ;; We're standing in a comment starter.
4424 ;; Search forward for a literal.
4435 ;; Search backward.
4446 ;; Block comment. Due to the nature of line
4447 ;; comments, they will always be covered by the
4448 ;; normal case above.
4451 ;; If LIM is bogus, beg will be bogus.
4455 ))))
4457 ;; In case external callers use this; it did have a docstring.
4461 "If the argument is a cons of two buffer positions (such as returned by
4462 `c-literal-limits'), and that range contains a C++ style line comment,
4463 then an extended range is returned that contains all adjacent line
4464 comments (i.e. all comments that starts in the same column with no
4465 empty lines or non-whitespace characters between them). Otherwise the
4466 argument is returned.
4468 Note that this function might do hidden buffer changes. See the
4469 comment at the start of cc-engine.el for more info."
4480 ;; Got to take care in the backward direction to handle
4481 ;; comments which are preceded by code.
4495 range)
4499 "Convenience function that given the result of `c-literal-limits',
4500 returns nil or the type of literal that the range surrounds, one
4501 of the symbols 'c, 'c++ or 'string. It's much faster than using
4502 `c-in-literal' and is intended to be used when you need both the
4503 type of a literal and its limits.
4505 Note that this function might do hidden buffer changes. See the
4506 comment at the start of cc-engine.el for more info."
4517 range))
4520 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4521 ;; This doesn't preserve point.
4530 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4531 ;; (default point). This is done by going back further in the buffer then
4532 ;; searching forward for literals. The position found won't be in a
4533 ;; literal. We start searching for the sought position TRY-SIZE (default
4534 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4535 ;; :-)
4543 stack elt size
4546 ;; Move forward one literal each time round this loop.
4547 ;; Move forward to the start of a comment or string.
4549 pos
4550 start
4551 nil ; target-depth
4552 nil ; stop-before
4553 s ; state
4556 ;; Gather details of the non-literal-bit - starting pos and size.
4560 pos))
4564 ;; Move forward to the end of the comment/string.
4568 start
4569 nil ; target-depth
4570 nil ; stop-before
4571 s ; state
4575 ;; Now try and find enough non-literal characters recorded on the stack.
4576 ;; Go back one recorded literal each time round this loop.
4578 stack)
4583 ;; Have we found enough yet?
4593 ;; Return a buffer position about HOW-FAR non-literal characters forward
4594 ;; from START-POS (default point), which must not be inside a literal.
4601 ;; Scan over counted characters.
4603 pos
4605 nil ; target-depth
4606 nil ; stop-before
4607 s ; state
4611 ;; Scan over literal characters.
4614 pos
4616 nil ; target-depth
4617 nil ; stop-before
4618 s ; state
4623 \f
4624 ;; `c-find-decl-spots' and accompanying stuff.
4626 ;; Variables used in `c-find-decl-spots' to cache the search done for
4627 ;; the first declaration in the last call. When that function starts,
4628 ;; it needs to back up over syntactic whitespace to look at the last
4629 ;; token before the region being searched. That can sometimes cause
4630 ;; moves back and forth over a quite large region of comments and
4631 ;; macros, which would be repeated for each changed character when
4632 ;; we're called during fontification, since font-lock refontifies the
4633 ;; current line for each change. Thus it's worthwhile to cache the
4634 ;; first match.
4635 ;;
4636 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4637 ;; the syntactic whitespace less or equal to some start position.
4638 ;; There's no cached value if it's nil.
4639 ;;
4640 ;; `c-find-decl-match-pos' is the match position if
4641 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4642 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4653 ; (defface c-debug-decl-spot-face
4654 ; '((t (:background "Turquoise")))
4655 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4656 ; (defface c-debug-decl-sws-face
4657 ; '((t (:background "Khaki")))
4658 ; "Debug face to mark the syntactic whitespace between the declaration
4659 ; spots and the preceding token end.")
4675 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4676 ;; but it contains lots of free variables that refer to things
4677 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4678 ;; if there is a match, otherwise at `cfd-limit'.
4679 ;;
4680 ;; This macro might do hidden buffer changes.
4683 ;; Find the next property match position if we haven't got one already.
4694 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4695 ;; got one already.
4705 ;; Match. Check if it's inside a comment or string literal.
4708 ;; Matched the end of a token preceding a decl spot.
4712 ;; Matched a token that start a decl spot.
4715 c-literal-faces)
4717 ;; No match. Finish up and exit the loop.
4719 nil)
4721 ;; Skip out of comments and string literals.
4728 ;; If we matched at the decl start, we have to back up over the
4729 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4730 ;; any decl spots in the syntactic ws.
4735 ;; Choose whichever match is closer to the start.
4738 cfd-re-match nil)
4740 cfd-prop-match nil))
4745 ;; Skip forward past comments only so we don't skip macros.
4747 ;; Set the position to continue at. We can avoid going over
4748 ;; the comments skipped above a second time, but it's possible
4749 ;; that the comment skipping has taken us past `cfd-prop-match'
4750 ;; since the property might be used inside comments.
4756 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4757 ;; label from the point to CFD-LIMIT.
4758 ;;
4759 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4760 ;; arguments: The first is the end position of the token preceding the spot,
4761 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4762 ;; the match is inside a macro. Point should be moved forward by at least
4763 ;; one token.
4764 ;;
4765 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4766 ;; it should return non-nil to ensure that the next search will find them.
4767 ;;
4768 ;; Such a spot is:
4769 ;; o The first token after bob.
4770 ;; o The first token after the end of submatch 1 in
4771 ;; `c-decl-prefix-or-start-re' when that submatch matches.
4772 ;; o The start of each `c-decl-prefix-or-start-re' match when
4773 ;; submatch 1 doesn't match.
4774 ;; o The first token after the end of each occurrence of the
4775 ;; `c-type' text property with the value `c-decl-end', provided
4776 ;; `c-type-decl-end-used' is set.
4777 ;;
4778 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4779 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4780 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4781 ;;
4782 ;; If the match is inside a macro then the buffer is narrowed to the
4783 ;; end of it, so that CFD-FUN can investigate the following tokens
4784 ;; without matching something that begins inside a macro and ends
4785 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4786 ;; CFD-FACE-CHECKLIST checks exist.
4787 ;;
4788 ;; The spots are visited approximately in order from top to bottom.
4789 ;; It's however the positions where `c-decl-prefix-or-start-re'
4790 ;; matches and where `c-decl-end' properties are found that are in
4791 ;; order. Since the spots often are at the following token, they
4792 ;; might be visited out of order insofar as more spots are reported
4793 ;; later on within the syntactic whitespace between the match
4794 ;; positions and their spots.
4795 ;;
4796 ;; It's assumed that comments and strings are fontified in the
4797 ;; searched range.
4798 ;;
4799 ;; This is mainly used in fontification, and so has an elaborate
4800 ;; cache to handle repeated calls from the same start position; see
4801 ;; the variables above.
4802 ;;
4803 ;; All variables in this function begin with `cfd-' to avoid name
4804 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4805 ;;
4806 ;; This function might do hidden buffer changes.
4810 ;; The end of the token preceding the decl spot last found
4811 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4812 ;; no match.
4813 cfd-re-match
4814 ;; The end position of the last `c-decl-prefix-or-start-re'
4815 ;; match. If this is greater than `cfd-continue-pos', the
4816 ;; next regexp search is started here instead.
4818 ;; The end of the last `c-decl-end' found by
4819 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4820 ;; match. If searching for the property isn't needed then we
4821 ;; disable it by setting it to `cfd-limit' directly.
4823 ;; The end of the token preceding the decl spot last found by
4824 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4825 ;; bob. `cfd-limit' if there's no match. In other words,
4826 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4828 ;; The position to continue searching at.
4829 cfd-continue-pos
4830 ;; The position of the last "real" token we've stopped at.
4831 ;; This can be greater than `cfd-continue-pos' when we get
4832 ;; hits inside macros or at `c-decl-end' positions inside
4833 ;; comments.
4835 ;; The end position of the last entered macro.
4838 ;; Initialize by finding a syntactically relevant start position
4839 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4840 ;; search unless we're at bob.
4843 ;; Must back up a bit since we look for the end of the previous
4844 ;; statement or declaration, which is earlier than the first
4845 ;; returned match.
4848 ;; First we need to move to a syntactically relevant position.
4849 ;; Begin by backing out of comment or string literals.
4852 ;; Try to use the faces to back up to the start of the
4853 ;; literal. FIXME: What if the point is on a declaration
4854 ;; inside a comment?
4860 ;; XEmacs doesn't fontify the quotes surrounding string
4861 ;; literals.
4870 ;; Don't trust the literal to contain only literal faces
4871 ;; (the font lock package might not have fontified the
4872 ;; start of it at all, for instance) so check that we have
4873 ;; arrived at something that looks like a start or else
4874 ;; resort to `c-literal-limits'.
4881 ;; The start is in a literal. If the limit is in the same
4882 ;; one we don't have to find a syntactic position etc. We
4883 ;; only check that if the limit is at or before bonl to save
4884 ;; time; it covers the by far most common case when font-lock
4885 ;; refontifies the current line only.
4896 ;; Completely inside a literal. Set up variables to trig the
4897 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
4898 ;; find a suitable start position.
4901 ;; Check if the region might be completely inside a macro, to
4902 ;; optimize that like the completely-inside-literal above.
4909 ;; (Maybe) completely inside a macro. Only need to trig the
4910 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
4911 ;; set things up.
4913 start-in-macro t))
4916 ;; Back out of any macro so we don't miss any declaration
4917 ;; that could follow after it.
4921 ;; Now we're at a proper syntactically relevant position so we
4922 ;; can use the cache. But first clear it if it applied
4923 ;; further down.
4928 ;; Don't have to do this if the cache is relevant here,
4929 ;; typically if the same line is refontified again. If
4930 ;; we're just some syntactic whitespace further down we can
4931 ;; still use the cache to limit the skipping.
4934 ;; If we hit `c-find-decl-syntactic-pos' and
4935 ;; `c-find-decl-match-pos' is set then we install the cached
4936 ;; values. If we hit `c-find-decl-syntactic-pos' and
4937 ;; `c-find-decl-match-pos' is nil then we know there's no decl
4938 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
4939 ;; and so we can continue the search from this point. If we
4940 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
4941 ;; the right spot to begin searching anyway.
4943 c-find-decl-match-pos)
4945 cfd-continue-pos syntactic-pos)
4950 ;; Always consider bob a match to get the first
4951 ;; declaration in the file. Do this separately instead of
4952 ;; letting `c-decl-prefix-or-start-re' match bob, so that
4953 ;; regexp always can consume at least one character to
4954 ;; ensure that we won't get stuck in an infinite loop.
4959 ;; Do an initial search now. In the bob case above it's
4960 ;; only done to search for a `c-decl-end' spot.
4964 cfd-match-pos)))))
4966 ;; Advance `cfd-continue-pos' if it's before the start position.
4967 ;; The closest continue position that might have effect at or
4968 ;; after the start depends on what we started in. This also
4969 ;; finds a suitable start position in the special cases when the
4970 ;; region is completely within a literal or macro.
4975 ;; If we're in a macro then it's the closest preceding token
4976 ;; in the macro. Check this before `start-in-literal',
4977 ;; since if we're inside a literal in a macro, the preceding
4978 ;; token is earlier than any `c-decl-end' spot inside the
4979 ;; literal (comment).
4981 ;; The only syntactic ws in macros are comments.
4987 ;; If we're in a comment it can only be the closest
4988 ;; preceding `c-decl-end' position within that comment, if
4989 ;; any. Go back to the beginning of such a property so that
4990 ;; `c-find-decl-prefix-search' will find the end of it.
4991 ;; (Can't stop at the end and install it directly on
4992 ;; `cfd-prop-match' since that variable might be cleared
4993 ;; after `cfd-fun' below.)
4994 ;;
4995 ;; Note that if the literal is a string then the property
4996 ;; search will simply skip to the beginning of it right
4997 ;; away.
5009 ;; Didn't find any property inside the comment, so we can
5010 ;; skip it entirely. (This won't skip past a string, but
5011 ;; that'll be handled quickly by the next
5012 ;; `c-find-decl-prefix-search' anyway.)
5018 ;; If we started in normal code, the only match that might
5019 ;; apply before the start is what we already got in
5020 ;; `cfd-match-pos' so we can continue at the start position.
5021 ;; (Note that we don't get here if the first match is below
5022 ;; it.)
5025 ;; Delete found matches if they are before our new continue
5026 ;; position, so that `c-find-decl-prefix-search' won't back up
5027 ;; to them later on.
5035 ;; This is the normal case and we got a proper syntactic
5036 ;; position. If there's a match then it's always outside
5037 ;; macros and comments, so advance to the next token and set
5038 ;; `cfd-token-pos'. The loop below will later go back using
5039 ;; `cfd-continue-pos' to fix declarations inside the
5040 ;; syntactic ws.
5048 ;; Have one of the special cases when the region is completely
5049 ;; within a literal or macro. `cfd-continue-pos' is set to a
5050 ;; good start position for the search, so do it.
5053 ;; Now loop. Round what? (ACM, 2006/7/5). We already got the first match.
5060 ;; Kludge to filter out matches on the "<" that
5061 ;; aren't open parens, for the sake of languages
5062 ;; that got `c-recognize-<>-arglists' set.
5067 ;; If `cfd-continue-pos' is less or equal to
5068 ;; `cfd-token-pos', we've got a hit inside a macro
5069 ;; that's in the syntactic whitespace before the last
5070 ;; "real" declaration we've checked. If they're equal
5071 ;; we've arrived at the declaration a second time, so
5072 ;; there's nothing to do.
5076 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5077 ;; we're still searching for declarations embedded in
5078 ;; the syntactic whitespace. In that case we need
5079 ;; only to skip comments and not macros, since they
5080 ;; can't be nested, and that's already been done in
5081 ;; `c-find-decl-prefix-search'.
5086 ;; Continue if the following token fails the
5087 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5094 t)))
5105 ;; Narrow to the end of the macro if we got a hit inside
5106 ;; one, to avoid recognizing things that start inside the
5107 ;; macro and end outside it.
5109 ;; Not in the same macro as in the previous round.
5120 t
5123 t)
5124 ;; The matched token was the last thing in the macro,
5125 ;; so the whole match is bogus.
5127 nil))))
5134 ;; Restore limits if we did macro narrowing above.
5141 ; cfd-match-pos, etc.
5143 \f
5144 ;; A cache for found types.
5146 ;; Buffer local variable that contains an obarray with the types we've
5147 ;; found. If a declaration is recognized somewhere we record the
5148 ;; fully qualified identifier in it to recognize it as a type
5149 ;; elsewhere in the file too. This is not accurate since we do not
5150 ;; bother with the scoping rules of the languages, but in practice the
5151 ;; same name is seldom used as both a type and something else in a
5152 ;; file, and we only use this as a last resort in ambiguous cases (see
5153 ;; `c-forward-decl-or-cast-1').
5154 ;;
5155 ;; Not every type need be in this cache. However, things which have
5156 ;; ceased to be types must be removed from it.
5157 ;;
5158 ;; Template types in C++ are added here too but with the template
5159 ;; arglist replaced with "<>" in references or "<" for the one in the
5160 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5161 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5162 ;; template specs can be fairly sized programs in themselves) and
5163 ;; improves the hit ratio (it's a type regardless of the template
5164 ;; args; it's just not the same type, but we're only interested in
5165 ;; recognizing types, not telling distinct types apart). Note that
5166 ;; template types in references are added here too; from the example
5167 ;; above there will also be an entry "Foo<".
5172 ;; Clears `c-found-types'.
5176 ;; Add the given region as a type in `c-found-types'. If the region
5177 ;; doesn't match an existing type but there is a type which is equal
5178 ;; to the given one except that the last character is missing, then
5179 ;; the shorter type is removed. That's done to avoid adding all
5180 ;; prefixes of a type as it's being entered and font locked. This
5181 ;; doesn't cover cases like when characters are removed from a type
5182 ;; or added in the middle. We'd need the position of point when the
5183 ;; font locking is invoked to solve this well.
5184 ;;
5185 ;; This function might do hidden buffer changes.
5192 ;; Remove the "NAME" from c-found-types, if present.
5196 ;; Return non-nil if the given region contains a type in
5197 ;; `c-found-types'.
5198 ;;
5199 ;; This function might do hidden buffer changes.
5201 c-found-types))
5204 ;; Return all the types in `c-found-types' as a sorted list of
5205 ;; strings.
5209 type-list)))
5210 c-found-types)
5213 ;; Shut up the byte compiler.
5217 ;; An after change function which, in conjunction with the info in
5218 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5219 ;; from `c-found-types', should this type have become stale. For
5220 ;; example, this happens to "foo" when "foo \n bar();" becomes
5221 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5222 ;; the fontification.
5223 ;;
5224 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5225 ;; type?
5245 ;; Changing the amount of (already existing) whitespace - don't do anything.
5250 ;; The syntactic relationship which defined a "found type" has been
5251 ;; destroyed.
5254 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5255 )))
5257 \f
5258 ;; Setting and removing syntax properties on < and > in languages (C++
5259 ;; and Java) where they can be template/generic delimiters as well as
5260 ;; their normal meaning of "less/greater than".
5262 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5263 ;; be delimiters, they are marked as such with the category properties
5264 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5266 ;; STRATEGY:
5267 ;;
5268 ;; It is impossible to determine with certainty whether a <..> pair in
5269 ;; C++ is two comparison operators or is template delimiters, unless
5270 ;; one duplicates a lot of a C++ compiler. For example, the following
5271 ;; code fragment:
5272 ;;
5273 ;; foo (a < b, c > d) ;
5274 ;;
5275 ;; could be a function call with two integer parameters (each a
5276 ;; relational expression), or it could be a constructor for class foo
5277 ;; taking one parameter d of templated type "a < b, c >". They are
5278 ;; somewhat easier to distinguish in Java.
5279 ;;
5280 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5281 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5282 ;; individually when their context so indicated. This gave rise to
5283 ;; intractable problems when one of a matching pair was deleted, or
5284 ;; pulled into a literal.]
5285 ;;
5286 ;; At each buffer change, the syntax-table properties are removed in a
5287 ;; before-change function and reapplied, when needed, in an
5288 ;; after-change function. It is far more important that the
5289 ;; properties get removed when they they are spurious than that they
5290 ;; be present when wanted.
5291 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5293 ;; POS (default point) is at a < character. If it is marked with
5294 ;; open paren syntax-table text property, remove the property,
5295 ;; together with the close paren property on the matching > (if
5296 ;; any).
5311 ;; POS (default point) is at a > character. If it is marked with
5312 ;; close paren syntax-table property, remove the property, together
5313 ;; with the open paren property on the matching < (if any).
5328 ;; POS (default point) is at a < or > character. If it has an
5329 ;; open/close paren syntax-table property, remove this property both
5330 ;; from the current character and its partner (which will also be
5331 ;; thusly marked).
5341 ;; POS (default point) is at a < character. If it is both marked
5342 ;; with open/close paren syntax-table property, and has a matching >
5343 ;; (also marked) which is after LIM, remove the property both from
5344 ;; the current > and its partner. Return t when this happens, nil
5345 ;; when it doesn't.
5359 t)))
5362 ;; POS (default point) is at a > character. If it is both marked
5363 ;; with open/close paren syntax-table property, and has a matching <
5364 ;; (also marked) which is before LIM, remove the property both from
5365 ;; the current < and its partner. Return t when this happens, nil
5366 ;; when it doesn't.
5380 t)))
5382 ;; Set by c-common-init in cc-mode.el.
5387 ;; Unmark certain pairs of "< .... >" which are currently marked as
5388 ;; template/generic delimiters. (This marking is via syntax-table
5389 ;; text properties).
5390 ;;
5391 ;; These pairs are those which are in the current "statement" (i.e.,
5392 ;; the region between the {, }, or ; before BEG and the one after
5393 ;; END), and which enclose any part of the interval (BEG END).
5394 ;;
5395 ;; Note that in C++ (?and Java), template/generic parens cannot
5396 ;; enclose a brace or semicolon, so we use these as bounds on the
5397 ;; region we must work on.
5398 ;;
5399 ;; This function is called from before-change-functions (via
5400 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5401 ;; and point is undefined, both at entry and exit.
5402 ;;
5403 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5404 ;; 2010-01-29.
5408 new-beg new-end need-new-beg need-new-end)
5409 ;; Locate the barrier before the changed region
5414 ;; Remove the syntax-table properties from each pertinent <...> pair.
5415 ;; Firsly, the ones with the < before beg and > after beg.
5420 ;; Locate the barrier after END.
5425 ;; Remove syntax-table properties from the remaining pertinent <...>
5426 ;; pairs, those with a > after end and < before end.
5431 ;; Extend the fontification region, if needed.
5443 ;; This is called from `after-change-functions' when
5444 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5445 ;; chars with paren syntax become part of another operator like "<<"
5446 ;; or ">=".
5447 ;;
5448 ;; This function might do hidden buffer changes.
5481 \f
5482 ;; Handling of small scale constructs like types and names.
5484 ;; Dynamically bound variable that instructs `c-forward-type' to also
5485 ;; treat possible types (i.e. those that it normally returns 'maybe or
5486 ;; 'found for) as actual types (and always return 'found for them).
5487 ;; This means that it records them in `c-record-type-identifiers' if
5488 ;; that is set, and that it adds them to `c-found-types'.
5491 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5492 ;; mark up successfully parsed arglists with paren syntax properties on
5493 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5494 ;; `c-type' property of each argument separating comma.
5495 ;;
5496 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5497 ;; all arglists for side effects (i.e. recording types), otherwise it
5498 ;; exploits any existing paren syntax properties to quickly jump to the
5499 ;; end of already parsed arglists.
5500 ;;
5501 ;; Marking up the arglists is not the default since doing that correctly
5502 ;; depends on a proper value for `c-restricted-<>-arglists'.
5505 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5506 ;; not accept arglists that contain binary operators.
5507 ;;
5508 ;; This is primarily used to handle C++ template arglists. C++
5509 ;; disambiguates them by checking whether the preceding name is a
5510 ;; template or not. We can't do that, so we assume it is a template
5511 ;; if it can be parsed as one. That usually works well since
5512 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5513 ;; in almost all cases would be pointless.
5514 ;;
5515 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5516 ;; should let the comma separate the function arguments instead. And
5517 ;; in a context where the value of the expression is taken, e.g. in
5518 ;; "if (a < b || c > d)", it's probably not a template.
5521 ;; Dynamically bound variables that instructs
5522 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5523 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5524 ;; `c-forward-label' to record the ranges of all the type and
5525 ;; reference identifiers they encounter. They will build lists on
5526 ;; these variables where each element is a cons of the buffer
5527 ;; positions surrounding each identifier. This recording is only
5528 ;; activated when `c-record-type-identifiers' is non-nil.
5529 ;;
5530 ;; All known types that can't be identifiers are recorded, and also
5531 ;; other possible types if `c-promote-possible-types' is set.
5532 ;; Recording is however disabled inside angle bracket arglists that
5533 ;; are encountered inside names and other angle bracket arglists.
5534 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5535 ;; instead.
5536 ;;
5537 ;; Only the names in C++ template style references (e.g. "tmpl" in
5538 ;; "tmpl<a,b>::foo") are recorded as references, other references
5539 ;; aren't handled here.
5540 ;;
5541 ;; `c-forward-label' records the label identifier(s) on
5542 ;; `c-record-ref-identifiers'.
5546 ;; This variable will receive a cons cell of the range of the last
5547 ;; single identifier symbol stepped over by `c-forward-name' if it's
5548 ;; successful. This is the range that should be put on one of the
5549 ;; record lists above by the caller. It's assigned nil if there's no
5550 ;; such symbol in the name.
5555 ;; Always true.
5565 ;; Always true.
5573 ;; Dynamically bound variable that instructs `c-forward-type' to
5574 ;; record the ranges of types that only are found. Behaves otherwise
5575 ;; like `c-record-type-identifiers'.
5579 ;; Used internally in `c-forward-keyword-clause' to move forward
5580 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5581 ;; possibly is prefixed by keywords and their associated clauses.
5582 ;; Try with a type/name first to not trip up on those that begin
5583 ;; with a keyword. Return t if a known or found type is moved
5584 ;; over. The point is clobbered if nil is returned. If range
5585 ;; recording is enabled, the identifier is recorded on as a type
5586 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5587 ;;
5588 ;; This macro might do hidden buffer changes.
5593 nil
5600 t)))
5603 ;; Used internally in `c-forward-keyword-clause' to move forward
5604 ;; over a comma separated list of types or names using
5605 ;; `c-forward-keyword-prefixed-id'.
5606 ;;
5607 ;; This macro might do hidden buffer changes.
5618 ;; Submatch MATCH in the current match data is assumed to surround a
5619 ;; token. If it's a keyword, move over it and any immediately
5620 ;; following clauses associated with it, stopping at the start of
5621 ;; the next token. t is returned in that case, otherwise the point
5622 ;; stays and nil is returned. The kind of clauses that are
5623 ;; recognized are those specified by `c-type-list-kwds',
5624 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5625 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5626 ;; and `c-<>-arglist-kwds'.
5627 ;;
5628 ;; This function records identifier ranges on
5629 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5630 ;; `c-record-type-identifiers' is non-nil.
5631 ;;
5632 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5633 ;; apply directly after the keyword, the type list is moved over
5634 ;; only when there is no unaccounted token before it (i.e. a token
5635 ;; that isn't moved over due to some other keyword list). The
5636 ;; identifier ranges in the list are still recorded if that should
5637 ;; be done, though.
5638 ;;
5639 ;; This function might do hidden buffer changes.
5642 ;; The call to `c-forward-<>-arglist' below is made after
5643 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5644 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5645 ;; should therefore be nil.
5657 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5662 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5667 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5674 ;; Use `c-forward-type' on every identifier we can find
5675 ;; inside the paren, to record the types.
5700 ;; If we are at the colon already, we move over the type
5701 ;; list after it.
5707 ;; Not at the colon, so stop here. But the identifier
5708 ;; ranges in the type list later on should still be
5709 ;; recorded.
5712 ;; If a keyword matched both one of the types above and
5713 ;; this one, we match `c-colon-type-list-re' after the
5714 ;; clause matched above.
5721 ;; There's a type after the `c-colon-type-list-re' match
5722 ;; after a keyword in `c-colon-type-list-kwds'.
5726 t)))
5728 ;; cc-mode requires cc-fonts.
5732 ;; The point is assumed to be at a "<". Try to treat it as the open
5733 ;; paren of an angle bracket arglist and move forward to the
5734 ;; corresponding ">". If successful, the point is left after the
5735 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5736 ;; returned. If ALL-TYPES is t then all encountered arguments in
5737 ;; the arglist that might be types are treated as found types.
5738 ;;
5739 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5740 ;; function handles text properties on the angle brackets and argument
5741 ;; separating commas.
5742 ;;
5743 ;; `c-restricted-<>-arglists' controls how lenient the template
5744 ;; arglist recognition should be.
5745 ;;
5746 ;; This function records identifier ranges on
5747 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5748 ;; `c-record-type-identifiers' is non-nil.
5749 ;;
5750 ;; This function might do hidden buffer changes.
5753 ;; If `c-record-type-identifiers' is set then activate
5754 ;; recording of any found types that constitute an argument in
5755 ;; the arglist.
5763 ;; `nconc' doesn't mind that the tail of
5764 ;; `c-record-found-types' is t.
5767 t)
5770 nil)))
5773 ;; Recursive part of `c-forward-<>-arglist'.
5774 ;;
5775 ;; This function might do hidden buffer changes.
5778 ;; Cover this so that any recorded found type ranges are
5779 ;; automatically lost if it turns out to not be an angle
5780 ;; bracket arglist. It's propagated through the return value
5781 ;; on successful completion.
5783 ;; List that collects the positions after the argument
5784 ;; separating ',' in the arglist.
5785 arg-start-pos)
5786 ;; If the '<' has paren open syntax then we've marked it as an angle
5787 ;; bracket arglist before, so skip to the end.
5795 t
5796 ;; Got unmatched paren angle brackets. We don't clear the paren
5797 ;; syntax properties and retry, on the basis that it's very
5798 ;; unlikely that paren angle brackets become operators by code
5799 ;; manipulation. It's far more likely that it doesn't match due
5800 ;; to narrowing or some temporary change.
5802 nil))
5807 ;; go forward one non-alphanumeric character (group) per iteration of
5808 ;; this loop.
5815 ;; All encountered identifiers are types, so set the
5816 ;; promote flag and parse the type.
5839 ;; Note: These regexps exploit the match order in \| so
5840 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5842 ;; Stop on ',', '|', '&', '+' and '-' to catch
5843 ;; common binary operators that could be between
5844 ;; two comparison expressions "a<b" and "c>d".
5846 nil t t))
5850 ;; Either an operator starting with '>' or the end of
5851 ;; the angle bracket arglist.
5858 ;; The angle bracket arglist is finished.
5870 ;; Either an operator starting with '<' or a nested arglist.
5874 ;; The '<' begins a multi-char operator.
5878 ;; We're at a nested <.....>
5884 ;; There's always an identifier before an angle
5885 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
5886 ;; or `c-<>-arglist-kwds'.
5903 ;; It was an angle bracket arglist.
5906 ;; Record the identifier before the template as a type
5907 ;; or reference depending on whether the arglist is last
5908 ;; in a qualified identifier.
5918 ;; At a "less than" operator.
5921 )))
5928 ;; Just another argument. Record the position. The
5929 ;; type check stuff that made us stop at it is at
5930 ;; the top of the loop.
5934 ;; Got a character that can't be in an angle bracket
5935 ;; arglist argument. Abort using `throw', since
5936 ;; it's useless to try to find a surrounding arglist
5937 ;; if we're nested.
5943 ;; The point is assumed to be directly after a ">". Try to treat it
5944 ;; as the close paren of an angle bracket arglist and move back to
5945 ;; the corresponding "<". If successful, the point is left at
5946 ;; the "<" and t is returned, otherwise the point isn't moved and
5947 ;; nil is returned. ALL-TYPES is passed on to
5948 ;; `c-forward-<>-arglist'.
5949 ;;
5950 ;; If the optional LIMIT is given, it bounds the backward search.
5951 ;; It's then assumed to be at a syntactically relevant position.
5952 ;;
5953 ;; This is a wrapper around `c-forward-<>-arglist'. See that
5954 ;; function for more details.
5963 t
5964 ;; See corresponding note in `c-forward-<>-arglist'.
5966 nil)
5973 t
5974 ;; Stopped at bob or a char that isn't allowed in an
5975 ;; arglist, so we've failed.
5977 nil)
5982 ;; If we moved then the "<" was part of some
5983 ;; multicharacter token.
5984 t
5990 ;; Matched the arglist. Break the while.
5992 nil)
5994 ;; We started from a non-paren ">" inside an
5995 ;; arglist.
5997 nil)
5999 ;; Matched a shorter arglist. Can be a nested
6000 ;; one so continue looking.
6002 t))
6003 t))))))
6008 ;; Move forward over a complete name if at the beginning of one,
6009 ;; stopping at the next following token. A keyword, as such,
6010 ;; doesn't count as a name. If the point is not at something that
6011 ;; is recognized as a name then it stays put.
6012 ;;
6013 ;; A name could be something as simple as "foo" in C or something as
6014 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6015 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6016 ;; int>::*volatile const" in C++ (this function is actually little
6017 ;; more than a `looking-at' call in all modes except those that,
6018 ;; like C++, have `c-recognize-<>-arglists' set).
6019 ;;
6020 ;; Return
6021 ;; o - nil if no name is found;
6022 ;; o - 'template if it's an identifier ending with an angle bracket
6023 ;; arglist;
6024 ;; o - 'operator of it's an operator identifier;
6025 ;; o - t if it's some other kind of name.
6026 ;;
6027 ;; This function records identifier ranges on
6028 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6029 ;; `c-record-type-identifiers' is non-nil.
6030 ;;
6031 ;; This function might do hidden buffer changes.
6034 ;; Turn off `c-promote-possible-types' here since we might
6035 ;; call `c-forward-<>-arglist' and we don't want it to promote
6036 ;; every suspect thing in the arglist to a type. We're
6037 ;; typically called from `c-forward-type' in this case, and
6038 ;; the caller only wants the top level type that it finds to
6039 ;; be promoted.
6040 c-promote-possible-types)
6046 ;; Check for keyword. We go to the last symbol in
6047 ;; `c-identifier-key' first.
6060 ;; "template" is only valid inside an
6061 ;; identifier if preceded by "::".
6066 t))
6068 ;; Handle a C++ operator or template identifier.
6072 ;; Got "... ::template".
6076 res subres))))
6079 ;; Got a cast operator.
6083 ;; Now we should match a sequence of either
6084 ;; '*', '&' or a name followed by ":: *",
6085 ;; where each can be followed by a sequence
6086 ;; of `c-opt-type-modifier-key'.
6089 t)
6099 t))))
6107 ;; Got some other operator.
6115 nil)
6117 ;; `id-start' is equal to `id-end' if we've jumped over
6118 ;; an identifier that doesn't end with a symbol token.
6119 ;; That can occur e.g. for Java import directives on the
6120 ;; form "foo.bar.*".
6127 res t)))
6137 ;; Got a concatenated identifier. This handles the
6138 ;; cases with tricky syntactic whitespace that aren't
6139 ;; covered in `c-identifier-key'.
6142 t)
6146 ;; Maybe an angle bracket arglist.
6154 c-last-identifier-range nil)
6159 ;; Continue if there's an identifier concatenation
6160 ;; operator after the template argument.
6166 t)
6171 nil)))
6172 )))))
6175 res))
6178 ;; Move forward over a type spec if at the beginning of one,
6179 ;; stopping at the next following token. The keyword "typedef"
6180 ;; isn't part of a type spec here.
6181 ;;
6182 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6183 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6184 ;; The current (2009-03-10) intention is to convert all uses of
6185 ;; `c-forward-type' to call with this parameter set, then to
6186 ;; eliminate it.
6187 ;;
6188 ;; Return
6189 ;; o - t if it's a known type that can't be a name or other
6190 ;; expression;
6191 ;; o - 'known if it's an otherwise known type (according to
6192 ;; `*-font-lock-extra-types');
6193 ;; o - 'prefix if it's a known prefix of a type;
6194 ;; o - 'found if it's a type that matches one in `c-found-types';
6195 ;; o - 'maybe if it's an identifier that might be a type; or
6196 ;; o - nil if it can't be a type (the point isn't moved then).
6197 ;;
6198 ;; The point is assumed to be at the beginning of a token.
6199 ;;
6200 ;; Note that this function doesn't skip past the brace definition
6201 ;; that might be considered part of the type, e.g.
6202 ;; "enum {a, b, c} foo".
6203 ;;
6204 ;; This function records identifier ranges on
6205 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6206 ;; `c-record-type-identifiers' is non-nil.
6207 ;;
6208 ;; This function might do hidden buffer changes.
6216 ;; Skip leading type modifiers. If any are found we know it's a
6217 ;; prefix of a type.
6226 ; "typedef".
6234 ;; In many languages the name can be used without the
6235 ;; prefix, so we add it to `c-found-types'.
6238 c-last-identifier-range)
6260 id-range c-last-identifier-range))))
6272 id-end)
6274 ;; Looking at a primitive or known type identifier. We've
6275 ;; checked for a name first so that we don't go here if the
6276 ;; known type match only is a prefix of another name.
6287 ;; There might be more keywords for the type.
6315 ;; A normal identifier.
6325 ;; It's an identifier that has been used as
6326 ;; a type somewhere else.
6327 'found
6328 ;; It's an identifier that might be a type.
6331 ;; A template is a type.
6335 ;; Otherwise it's an operator identifier, which is not a type.
6340 ;; Skip trailing type modifiers. If any are found we know it's
6341 ;; a type.
6347 ;; Step over any type suffix operator. Do not let the existence
6348 ;; of these alter the classification of the found type, since
6349 ;; these operators typically are allowed in normal expressions
6350 ;; too.
6357 ;; Look for a trailing operator that concatenates the type
6358 ;; with a following one, and if so step past that one through
6359 ;; a recursive call. Note that we don't record concatenated
6360 ;; types in `c-found-types' - it's the component types that
6361 ;; are recorded when appropriate.
6364 c-promote-possible-types))
6365 ;; If we can't promote then set `c-record-found-types' so that
6366 ;; we can merge in the types from the second part afterwards if
6367 ;; it turns out to be a known type there.
6370 subres)
6379 ;; If either operand certainly is a type then both are, but we
6380 ;; don't let the existence of the operator itself promote two
6381 ;; uncertain types to a certain one.
6400 ;; Merge in the ranges of any types found by the second
6401 ;; `c-forward-type'.
6403 ;; `nconc' doesn't mind that the tail of
6404 ;; `c-record-found-types' is t.
6406 c-record-type-identifiers))))
6414 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6416 res))
6419 ;; Used for Java code only at the moment. Assumes point is on the
6420 ;; @, moves forward an annotation. returns nil if there is no
6421 ;; annotation at point.
6428 t)))
6430 \f
6431 ;; Handling of large scale constructs like statements and declarations.
6433 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6434 ;; defsubst or perhaps even a defun, but it contains lots of free
6435 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6437 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6438 ;; of types when parsing a declaration, which means that it
6439 ;; sometimes consumes the identifier in the declaration as a type.
6440 ;; This is used to "backtrack" and make the last type be treated as
6441 ;; an identifier instead.
6444 ;; These identifiers are bound only in the inner let.
6446 identifier-start type-start
6447 got-parens nil
6448 got-identifier t
6449 got-suffix t
6450 got-suffix-after-parens id-start
6454 t
6455 backup-at-type))
6457 id-start backup-id-start)
6459 id-start start-pos))
6461 ;; When these flags already are set we've found specifiers that
6462 ;; unconditionally signal these attributes - backtracking doesn't
6463 ;; change that. So keep them set in that case.
6470 ;; This identifier is bound only in the inner let.
6474 ;; Move forward over a declaration or a cast if at the start of one.
6475 ;; The point is assumed to be at the start of some token. Nil is
6476 ;; returned if no declaration or cast is recognized, and the point
6477 ;; is clobbered in that case.
6478 ;;
6479 ;; If a declaration is parsed:
6480 ;;
6481 ;; The point is left at the first token after the first complete
6482 ;; declarator, if there is one. The return value is a cons where
6483 ;; the car is the position of the first token in the declarator. (See
6484 ;; below for the cdr.)
6485 ;; Some examples:
6486 ;;
6487 ;; void foo (int a, char *b) stuff ...
6488 ;; car ^ ^ point
6489 ;; float (*a)[], b;
6490 ;; car ^ ^ point
6491 ;; unsigned int a = c_style_initializer, b;
6492 ;; car ^ ^ point
6493 ;; unsigned int a (cplusplus_style_initializer), b;
6494 ;; car ^ ^ point (might change)
6495 ;; class Foo : public Bar {}
6496 ;; car ^ ^ point
6497 ;; class PikeClass (int a, string b) stuff ...
6498 ;; car ^ ^ point
6499 ;; enum bool;
6500 ;; car ^ ^ point
6501 ;; enum bool flag;
6502 ;; car ^ ^ point
6503 ;; void cplusplus_function (int x) throw (Bad);
6504 ;; car ^ ^ point
6505 ;; Foo::Foo (int b) : Base (b) {}
6506 ;; car ^ ^ point
6507 ;;
6508 ;; The cdr of the return value is non-nil when a
6509 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6510 ;; Specifically it is a dotted pair (A . B) where B is t when a
6511 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6512 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6513 ;; specifier is present. I.e., (some of) the declared
6514 ;; identifier(s) are types.
6515 ;;
6516 ;; If a cast is parsed:
6517 ;;
6518 ;; The point is left at the first token after the closing paren of
6519 ;; the cast. The return value is `cast'. Note that the start
6520 ;; position must be at the first token inside the cast parenthesis
6521 ;; to recognize it.
6522 ;;
6523 ;; PRECEDING-TOKEN-END is the first position after the preceding
6524 ;; token, i.e. on the other side of the syntactic ws from the point.
6525 ;; Use a value less than or equal to (point-min) if the point is at
6526 ;; the first token in (the visible part of) the buffer.
6527 ;;
6528 ;; CONTEXT is a symbol that describes the context at the point:
6529 ;; 'decl In a comma-separated declaration context (typically
6530 ;; inside a function declaration arglist).
6531 ;; '<> In an angle bracket arglist.
6532 ;; 'arglist Some other type of arglist.
6533 ;; nil Some other context or unknown context. Includes
6534 ;; within the parens of an if, for, ... construct.
6535 ;;
6536 ;; LAST-CAST-END is the first token after the closing paren of a
6537 ;; preceding cast, or nil if none is known. If
6538 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6539 ;; the position after the closest preceding call where a cast was
6540 ;; matched. In that case it's used to discover chains of casts like
6541 ;; "(a) (b) c".
6542 ;;
6543 ;; This function records identifier ranges on
6544 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6545 ;; `c-record-type-identifiers' is non-nil.
6546 ;;
6547 ;; This function might do hidden buffer changes.
6550 ;; first type, i.e. after any leading specifiers. It might
6551 ;; also point at the beginning of the preceding syntactic
6552 ;; whitespace.
6554 ;; Set to the result of `c-forward-type'.
6555 at-type
6556 ;; The position of the first token in what we currently
6557 ;; believe is the type in the declaration or cast, after any
6558 ;; specifiers and their associated clauses.
6559 type-start
6560 ;; The position of the first token in what we currently
6561 ;; believe is the declarator for the first identifier. Set
6562 ;; when the type is found, and moved forward over any
6563 ;; `c-decl-hangon-kwds' and their associated clauses that
6564 ;; occurs after the type.
6565 id-start
6566 ;; These store `at-type', `type-start' and `id-start' of the
6567 ;; identifier before the one in those variables. The previous
6568 ;; identifier might turn out to be the real type in a
6569 ;; declaration if the last one has to be the declarator in it.
6570 ;; If `backup-at-type' is nil then the other variables have
6571 ;; undefined values.
6572 backup-at-type backup-type-start backup-id-start
6573 ;; Set if we've found a specifier (apart from "typedef") that makes
6574 ;; the defined identifier(s) types.
6575 at-type-decl
6576 ;; Set if we've a "typedef" keyword.
6577 at-typedef
6578 ;; Set if we've found a specifier that can start a declaration
6579 ;; where there's no type.
6580 maybe-typeless
6581 ;; If a specifier is found that also can be a type prefix,
6582 ;; these flags are set instead of those above. If we need to
6583 ;; back up an identifier, they are copied to the real flag
6584 ;; variables. Thus they only take effect if we fail to
6585 ;; interpret it as a type.
6586 backup-at-type-decl backup-maybe-typeless
6587 ;; Whether we've found a declaration or a cast. We might know
6588 ;; this before we've found the type in it. It's 'ids if we've
6589 ;; found two consecutive identifiers (usually a sure sign, but
6590 ;; we should allow that in labels too), and t if we've found a
6591 ;; specifier keyword (a 100% sure sign).
6592 at-decl-or-cast
6593 ;; Set when we need to back up to parse this as a declaration
6594 ;; but not as a cast.
6595 backup-if-not-cast
6596 ;; For casts, the return position.
6597 cast-end
6598 ;; Save `c-record-type-identifiers' and
6599 ;; `c-record-ref-identifiers' since ranges are recorded
6600 ;; speculatively and should be thrown away if it turns out
6601 ;; that it isn't a declaration or cast.
6608 ;; Check for a type. Unknown symbols are treated as possible
6609 ;; types, but they could also be specifiers disguised through
6610 ;; macros like __INLINE__, so we recognize both types and known
6611 ;; specifiers after them too.
6615 ;; Look for a specifier keyword clause.
6627 ;; Found a known or possible type or a prefix of a known type.
6630 ;; Got two identifiers with nothing but whitespace
6631 ;; between them. That can only happen in declarations.
6635 ;; If the previous identifier is a found type we
6636 ;; record it as a real one; it might be some sort of
6637 ;; alias for a prefix like "unsigned".
6644 backup-type-start type-start
6645 backup-id-start id-start
6646 at-type found-type
6647 type-start start
6649 ;; The previous ambiguous specifier/type turned out
6650 ;; to be a type since we've parsed another one after
6651 ;; it, so clear these backup flags.
6652 backup-at-type-decl nil
6653 backup-maybe-typeless nil))
6657 ;; Handle known specifier keywords and
6658 ;; `c-decl-hangon-kwds' which can occur after known
6659 ;; types.
6662 ;; It's a hang-on keyword that can occur anywhere.
6666 ;; Move the identifier start position if
6667 ;; we've passed a type.
6669 ;; Otherwise treat this as a specifier and
6670 ;; move the fallback position.
6674 ;; It's an ordinary specifier so we know that
6675 ;; anything before this can't be the type.
6677 start-pos kwd-clause-end)
6680 ;; It's ambiguous whether this keyword is a
6681 ;; specifier or a type prefix, so set the backup
6682 ;; flags. (It's assumed that `c-forward-type'
6683 ;; moved further than `c-forward-keyword-clause'.)
6691 ;; This test only happens after we've scanned a type.
6692 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6697 ;; Haven't matched a type so it's an unambiguous
6698 ;; specifier keyword and we know we're in a
6699 ;; declaration.
6704 ;; If the type isn't known we continue so that we'll jump
6705 ;; over all specifiers and type identifiers. The reason
6706 ;; to do this for a known type prefix is to make things
6707 ;; like "unsigned INT16" work.
6712 ;; If a known type was found, we still need to skip over any
6713 ;; hangon keyword clauses after it. Otherwise it has already
6714 ;; been done in the loop above.
6720 ;; A prefix type is itself a primitive type when it's not
6721 ;; followed by another type.
6725 ;; Got no type but set things up to continue anyway to handle
6726 ;; the various cases when a declaration doesn't start with a
6727 ;; type.
6732 ;; If it's C++ then check if the last "type" ends on the form
6733 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
6734 ;; (con|de)structor.
6745 ;; Cheating in the handling of syntactic ws below.
6752 name))
6753 ;; It is a (con|de)structor name. In that case the
6754 ;; declaration is typeless so zap out any preceding
6755 ;; identifier(s) that we might have taken as types.
6758 backup-at-type nil
6759 id-start type-start))))))
6761 ;; Check for and step over a type decl expression after the thing
6762 ;; that is or might be a type. This can't be skipped since we
6763 ;; need the correct end position of the declarator for
6764 ;; `max-type-decl-end-*'.
6766 ;; True if there's a non-open-paren match of
6767 ;; `c-type-decl-prefix-key'.
6768 got-prefix
6769 ;; True if the declarator is surrounded by a parenthesis pair.
6770 got-parens
6771 ;; True if there is an identifier in the declarator.
6772 got-identifier
6773 ;; True if there's a non-close-paren match of
6774 ;; `c-type-decl-suffix-key'.
6775 got-suffix
6776 ;; True if there's a prefix match outside the outermost
6777 ;; paren pair that surrounds the declarator.
6778 got-prefix-before-parens
6779 ;; True if there's a suffix match outside the outermost
6780 ;; paren pair that surrounds the declarator. The value is
6781 ;; the position of the first suffix match.
6782 got-suffix-after-parens
6783 ;; True if we've parsed the type decl to a token that is
6784 ;; known to end declarations in this context.
6785 at-decl-end
6786 ;; The earlier values of `at-type' and `type-start' if we've
6787 ;; shifted the type backwards.
6788 identifier-type identifier-start
6789 ;; If `c-parse-and-markup-<>-arglists' is set we need to
6790 ;; turn it off during the name skipping below to avoid
6791 ;; getting `c-type' properties that might be bogus. That
6792 ;; can happen since we don't know if
6793 ;; `c-restricted-<>-arglists' will be correct inside the
6794 ;; arglist paren that gets entered.
6799 ;; Skip over type decl prefix operators. (Note similar code in
6800 ;; `c-font-lock-declarators'.)
6804 ;; If the second submatch matches in C++ then
6805 ;; we're looking at an identifier that's a
6806 ;; prefix only if it specifies a member pointer.
6809 ;; We only check for a trailing "::" and
6810 ;; let the "*" that should follow be
6811 ;; matched in the next round.
6813 ;; It turned out to be the real identifier,
6814 ;; so stop.
6815 nil))
6816 t))
6830 ;; Skip over an identifier.
6835 ;; Skip over type decl suffix operators.
6842 t)
6846 t)
6852 ;; No suffix matched. We might have matched the
6853 ;; identifier as a type and the open paren of a
6854 ;; function arglist as a type decl prefix. In that
6855 ;; case we should "backtrack": Reinterpret the last
6856 ;; type as the identifier, move out of the arglist and
6857 ;; continue searching for suffix operators.
6858 ;;
6859 ;; Do this even if there's no preceding type, to cope
6860 ;; with old style function declarations in K&R C,
6861 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
6862 ;; style declarations. That isn't applicable in an
6863 ;; arglist context, though.
6868 maybe-typeless
6869 backup-maybe-typeless
6876 t))
6883 at-type)
6884 ;; Have found no identifier but `c-typeless-decl-kwds' has
6885 ;; matched so we know we're inside a declaration. The
6886 ;; preceding type must be the identifier instead.
6890 at-decl-or-cast
6893 ;; CASE 1
6895 ;; Encountered something inside parens that isn't matched by
6896 ;; the `c-type-decl-*' regexps, so it's not a type decl
6897 ;; expression. Try to skip out to the same paren depth to
6898 ;; not confuse the cast check below.
6900 ;; If we've found a specifier keyword then it's a
6901 ;; declaration regardless.
6909 ;; Now we've collected info about various characteristics of
6910 ;; the construct we're looking at. Below follows a decision
6911 ;; tree based on that. It's ordered to check more certain
6912 ;; signs before less certain ones.
6917 ;; CASE 2
6920 ;; Got another identifier directly after the type, so it's a
6921 ;; declaration.
6928 maybe-typeless
6929 backup-maybe-typeless))
6930 ;; Got a declaration of the form "foo bar (gnu);" where we've
6931 ;; recognized "bar" as the type and "gnu" as the declarator.
6932 ;; In this case it's however more likely that "bar" is the
6933 ;; declarator and "gnu" a function argument or initializer (if
6934 ;; `c-recognize-paren-inits' is set), since the parens around
6935 ;; "gnu" would be superfluous if it's a declarator. Shift the
6936 ;; type one step backward.
6939 ;; Found no identifier.
6945 ;; CASE 3
6947 ;; Got a plain list of identifiers. If a colon follows it's
6948 ;; a valid label, or maybe a bitfield. Otherwise the last
6949 ;; one probably is the declared identifier and we should
6950 ;; back up to the previous type, providing it isn't a cast.
6954 ;; If we've found a specifier keyword then it's a
6955 ;; declaration regardless.
6966 ;; CASE 4
6970 ;; Got a plain list of identifiers followed by some suffix.
6971 ;; If this isn't a cast then the last identifier probably is
6972 ;; the declared one and we should back up to the previous
6973 ;; type.
6977 ;; CASE 5
6979 ;; If the type is known we know that there can't be any
6980 ;; identifier somewhere else, and it's only in declarations in
6981 ;; e.g. function prototypes and in casts that the identifier may
6982 ;; be left out.
6986 ;; Only got a single identifier (parsed as a type so far).
6987 ;; CASE 6
6989 ;; Check that the identifier isn't at the start of an
6990 ;; expression.
6991 at-decl-end
6994 ;; Inside an arglist that contains declarations. If K&R
6995 ;; style declarations and parenthesis style initializers
6996 ;; aren't allowed then the single identifier must be a
6997 ;; type, else we require that it's known or found
6998 ;; (primitive types are handled above).
7003 ;; Inside a template arglist. Accept known and found
7004 ;; types; other identifiers could just as well be
7005 ;; constants in C++.
7008 ;; CASE 7
7009 ;; Can't be a valid declaration or cast, but if we've found a
7010 ;; specifier it can't be anything else either, so treat it as
7011 ;; an invalid/unfinished declaration or cast.
7019 maybe-typeless
7020 backup-maybe-typeless
7024 c-after-suffixed-type-maybe-decl-key))))))
7025 ;; Got an empty paren pair and a preceding type that probably
7026 ;; really is the identifier. Shift the type backwards to make
7027 ;; the last one the identifier. This is analogous to the
7028 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7029 ;; above.
7030 ;;
7031 ;; Exception: In addition to the conditions in that
7032 ;; "backtracking" code, do not shift backward if we're not
7033 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7034 ;; Since there's no preceding type, the shift would mean that
7035 ;; the declaration is typeless. But if the regexp doesn't match
7036 ;; then we will simply fall through in the tests below and not
7037 ;; recognize it at all, so it's better to try it as an abstract
7038 ;; declarator instead.
7041 ;; Still no identifier.
7042 ;; CASE 8
7044 ;; Require `got-prefix' together with either `got-parens' or
7045 ;; `got-suffix' to recognize it as an abstract declarator:
7046 ;; `got-parens' only is probably an empty function call.
7047 ;; `got-suffix' only can build an ordinary expression together
7048 ;; with the preceding identifier which we've taken as a type.
7049 ;; We could actually accept on `got-prefix' only, but that can
7050 ;; easily occur temporarily while writing an expression so we
7051 ;; avoid that case anyway. We could do a better job if we knew
7052 ;; the point when the fontification was invoked.
7055 ;; CASE 9
7059 got-suffix-after-parens
7061 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7062 ;; normal function call after all (or perhaps a C++ style object
7063 ;; instantiation expression).
7066 ;; CASE 10
7068 ;; By now we've located the type in the declaration that we know
7069 ;; we're in.
7072 ;; CASE 11
7080 ;; Shift the type backward in the case that there's a
7081 ;; single identifier inside parens. That can only
7082 ;; occur in K&R style function declarations so it's
7083 ;; more likely that it really is a function call.
7084 ;; Therefore we only do this after
7085 ;; `c-after-suffixed-type-decl-key' has matched.
7087 got-suffix-after-parens))
7088 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7091 ;; CASE 12
7094 ;; It's a declaration if a known type precedes it and it can't be a
7095 ;; function call.
7098 ;; If we get here we can't tell if this is a type decl or a normal
7099 ;; expression by looking at it alone. (That's under the assumption
7100 ;; that normal expressions always can look like type decl expressions,
7101 ;; which isn't really true but the cases where it doesn't hold are so
7102 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7103 ;; the effort to look for them.)
7106 ;; If this is a declaration it should end here or its initializer(*)
7107 ;; should start here, so check for allowed separation tokens. Note
7108 ;; that this rule doesn't work e.g. with a K&R arglist after a
7109 ;; function header.
7110 ;;
7111 ;; *) Don't check for C++ style initializers using parens
7112 ;; since those already have been matched as suffixes.
7113 ;;
7114 ;; If `at-decl-or-cast' is then we've found some other sign that
7115 ;; it's a declaration or cast, so then it's probably an
7116 ;; invalid/unfinished one.
7119 ;; Below are tests that only should be applied when we're certain to
7120 ;; not have parsed halfway through an expression.
7122 ;; CASE 14
7124 ;; The expression starts with a known type so treat it as a
7125 ;; declaration.
7128 ;; CASE 15
7130 ;; In C++ we check if the identifier is a known type, since
7131 ;; (con|de)structors use the class name as identifier.
7132 ;; We've always shifted over the identifier as a type and
7133 ;; then backed up again in this case.
7134 identifier-type
7137 ;; `at-type' probably won't be 'found for
7138 ;; destructors since the "~" is then part of the
7139 ;; type name being checked against the list of
7140 ;; known types, so do a check without that
7141 ;; operator.
7146 c-identifier-syntax-table
7150 ;; We have already parsed the type earlier,
7151 ;; so it'd be possible to cache the end
7152 ;; position instead of redoing it here, but
7153 ;; then we'd need to keep track of another
7154 ;; position everywhere.
7162 ;; CASE 16
7164 at-type
7168 ;; Got something like "foo * bar;". Since we're not inside an
7169 ;; arglist it would be a meaningless expression because the
7170 ;; result isn't used. We therefore choose to recognize it as
7171 ;; a declaration. Do not allow a suffix since it could then
7172 ;; be a function call.
7175 ;; CASE 17
7180 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7181 ;; be an odd expression or it could be a declaration. Treat
7182 ;; it as a declaration if "a" has been used as a type
7183 ;; somewhere else (if it's a known type we won't get here).
7186 ;; CASE 18
7192 ;; Got a type followed by an abstract declarator. If `got-prefix'
7193 ;; is set it's something like "a *" without anything after it. If
7194 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7195 ;; or similar, which we accept only if the context rules out
7196 ;; expressions.
7199 ;; If we had a complete symbol table here (which rules out
7200 ;; `c-found-types') we should return t due to the disambiguation rule
7201 ;; (in at least C++) that anything that can be parsed as a declaration
7202 ;; is a declaration. Now we're being more defensive and prefer to
7203 ;; highlight things like "foo (bar);" as a declaration only if we're
7204 ;; inside an arglist that contains declarations.
7207 ;; The point is now after the type decl expression.
7210 ;; Check for a cast.
7213 c-cast-parens
7215 ;; Should be the first type/identifier in a cast paren.
7219 ;; The closing paren should follow.
7224 ;; There should be a primary expression after it.
7233 ;; Check if the expression begins with a prefix keyword.
7236 ;; Expression begins with an ambiguous operator. Treat
7237 ;; it as a cast if it's a type decl or if we've
7238 ;; recognized the type somewhere else.
7241 ;; Unless it's a keyword, it's the beginning of a primary
7242 ;; expression.
7244 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7245 ;; that it matched a whole one so that we don't e.g. confuse
7246 ;; the operator '-' with '->'. It's ok if it matches further,
7247 ;; though, since it e.g. can match the float '.5' while the
7248 ;; operator regexp only matches '.'.
7252 ;; There should either be a cast before it or something that isn't an
7253 ;; identifier or close paren.
7261 ;; It's a symbol. Accept it only if it's one of the
7262 ;; keywords that can precede an expression (without
7263 ;; surrounding parens).
7266 ;; Check that it isn't a close paren (block close is ok,
7267 ;; though).
7269 ;; Check that it isn't a nonsymbol identifier.
7272 ;; Handle the cast.
7282 ;; We're at a declaration. Highlight the type and the following
7283 ;; declarators.
7289 ;; Make sure to propagate the `c-decl-arg-start' property to
7290 ;; the next argument if it's set in this one, to cope with
7291 ;; interactive refontification.
7305 ;; False alarm. Restore the recorded ranges.
7307 c-record-ref-identifiers save-rec-ref-ids)
7308 nil))))
7311 ;; Assuming that point is at the beginning of a token, check if it starts a
7312 ;; label and if so move over it and return non-nil (t in default situations,
7313 ;; specific symbols (see below) for interesting situations), otherwise don't
7314 ;; move and return nil. "Label" here means "most things with a colon".
7315 ;;
7316 ;; More precisely, a "label" is regarded as one of:
7317 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7318 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7319 ;; bare "case", should the colon be missing. We return t;
7320 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7321 ;; return t;
7322 ;; (iv) One of QT's "extended" C++ variants of
7323 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7324 ;; Returns the symbol `qt-2kwds-colon'.
7325 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7326 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7327 ;; colon). Currently (2006-03), this applies only to Objective C's
7328 ;; keywords "@private", "@protected", and "@public". Returns t.
7329 ;;
7330 ;; One of the things which will NOT be recognized as a label is a bit-field
7331 ;; element of a struct, something like "int foo:5".
7332 ;;
7333 ;; The end of the label is taken to be just after the colon, or the end of
7334 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7335 ;; after the end on return. The terminating char gets marked with
7336 ;; `c-decl-end' to improve recognition of the following declaration or
7337 ;; statement.
7338 ;;
7339 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7340 ;; label, if any, has already been marked up like that.
7341 ;;
7342 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7343 ;; after the preceding token, i.e. on the other side of the
7344 ;; syntactic ws from the point. Use a value less than or equal to
7345 ;; (point-min) if the point is at the first token in (the visible
7346 ;; part of) the buffer.
7347 ;;
7348 ;; The optional LIMIT limits the forward scan for the colon.
7349 ;;
7350 ;; This function records the ranges of the label symbols on
7351 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7352 ;; non-nil.
7353 ;;
7354 ;; This function might do hidden buffer changes.
7357 label-end
7358 qt-symbol-idx
7359 macro-start ; if we're in one.
7360 label-type
7361 kwd)
7363 ;; "case" or "default" (Doesn't apply to AWK).
7366 ;; Record only the keyword itself for fontification, since in
7367 ;; case labels the following is a constant expression and not
7368 ;; a label.
7372 ;; Find the label end.
7376 ;; Stop on chars that aren't allowed in expressions,
7377 ;; and on operator chars that would be meaningless
7378 ;; there. FIXME: This doesn't cope with ?: operators.
7386 t)
7388 ;; It's an unfinished label. We consider the keyword enough
7389 ;; to recognize it as a label, so that it gets fontified.
7390 ;; Leave the point at the end of it, but don't put any
7391 ;; `c-decl-end' marker.
7393 t))))
7395 ;; @private, @protected, @public, in Objective C, or similar.
7398 ;; For a `c-opt-extra-label-key' match, we record the whole
7399 ;; thing for fontification. That's to get the leading '@' in
7400 ;; Objective-C protection labels fontified.
7407 ;; All other cases of labels.
7410 ;; A colon label must have something before the colon.
7413 ;; Check that we're not after a token that can't precede a label.
7415 ;; Trivially succeeds when there's no preceding token.
7416 ;; Succeeds when we're at a virtual semicolon.
7425 ;; Check if we're after a label, if we're after a closing
7426 ;; paren that belong to statement, and with
7427 ;; `c-label-prefix-re'. It's done in different order
7428 ;; depending on `assume-markup' since the checks have
7429 ;; different expensiveness.
7456 ;; Might be after another label, so check it recursively.
7460 ;; Essentially the same as the
7461 ;; `c-syntactic-re-search-forward' regexp below.
7467 ;; Note: the following should work instead of the
7468 ;; narrow-to-region above. Investigate why not,
7469 ;; sometime. ACM, 2006-03-31.
7470 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7471 ;; macro-start t)
7473 ;; If the caller turned on recording for us,
7474 ;; it shouldn't apply when we check the
7475 ;; preceding label.
7476 c-record-type-identifiers)
7477 ;; A label can't start at a cpp directive. Check for
7478 ;; this, since c-forward-syntactic-ws would foul up on it.
7483 ;; Point is still at the beginning of the possible label construct.
7484 ;;
7485 ;; Check that the next nonsymbol token is ":", or that we're in one
7486 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7487 ;; arguments. FIXME: Should build this regexp from the language
7488 ;; constants.
7490 ;; public: protected: private:
7500 ;; QT double keyword like "protected slots:" or goto target.
7519 'qt-1kwd-colon
7531 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7541 t)
7544 ;; Record the identifiers in the label for fontification, unless
7545 ;; it begins with `c-label-kwds' in which case the following
7546 ;; identifiers are part of a (constant) expression that
7547 ;; shouldn't be fontified.
7559 ;; Not a label.
7561 label-type))
7564 ;; Assuming the point is at the beginning of a token, try to move
7565 ;; forward to the end of the Objective-C directive that starts
7566 ;; there. Return t if a directive was fully recognized, otherwise
7567 ;; the point is moved as far as one could be successfully parsed and
7568 ;; nil is returned.
7569 ;;
7570 ;; This function records identifier ranges on
7571 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7572 ;; `c-record-type-identifiers' is non-nil.
7573 ;;
7574 ;; This function might do hidden buffer changes.
7577 start-char
7579 lim
7580 ;; Turn off recognition of angle bracket arglists while parsing
7581 ;; types here since the protocol reference list might then be
7582 ;; considered part of the preceding name or superclass-name.
7593 t)
7602 ;; Handle the name of the class itself.
7604 ; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7605 ; at EOB.
7612 ;; Look for ": superclass-name" or "( category-name )".
7623 ;; Look for a protocol reference list.
7629 t))))
7635 t)
7638 nil)))
7641 ;; Go to the first non-whitespace after the colon that starts a
7642 ;; multiple inheritance introduction. Optional LIM is the farthest
7643 ;; back we should search.
7644 ;;
7645 ;; This function might do hidden buffer changes.
7653 ;; Return nil if we aren't in a method definition, otherwise the
7654 ;; position of the initial [+-].
7655 ;;
7656 ;; This function might do hidden buffer changes.
7662 ))
7664 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7666 ;; Return non-nil if point is within a gcc \"asm\" block.
7667 ;;
7668 ;; This should be called with point inside an argument list.
7669 ;;
7670 ;; Only one level of enclosing parentheses is considered, so for
7671 ;; instance `nil' is returned when in a function call within an asm
7672 ;; operand.
7673 ;;
7674 ;; This function might do hidden buffer changes.
7686 a function.
7688 More precisely, being at the top-level means that point is either
7689 outside any enclosing block (such as a function definition), or
7690 directly inside a class, namespace or other block that contains
7691 another declaration level.
7693 If point is not at the top-level (e.g. it is inside a method
7694 definition), then nil is returned. Otherwise, if point is at a
7695 top-level not enclosed within a class definition, t is returned.
7696 Otherwise, a 2-vector is returned where the zeroth element is the
7697 buffer position of the start of the class declaration, and the first
7698 element is the buffer position of the enclosing class's opening
7699 brace.
7701 Note that this function might do hidden buffer changes. See the
7702 comment at the start of cc-engine.el for more info."
7708 ;; Return non-nil if the point is in the region after the argument
7709 ;; list of a function and its opening brace (or semicolon in case it
7710 ;; got no body). If there are K&R style argument declarations in
7711 ;; that region, the point has to be inside the first one for this
7712 ;; function to recognize it.
7713 ;;
7714 ;; If successful, the point is moved to the first token after the
7715 ;; function header (see `c-forward-decl-or-cast-1' for details) and
7716 ;; the position of the opening paren of the function arglist is
7717 ;; returned.
7718 ;;
7719 ;; The point is clobbered if not successful.
7720 ;;
7721 ;; LIM is used as bound for backward buffer searches.
7722 ;;
7723 ;; This function might do hidden buffer changes.
7736 ;; There should not be a '=' or ',' between beg and the
7737 ;; start of the declaration since that means we were in the
7738 ;; "expression part" of the declaration.
7743 ;; Check that there's an arglist paren in the
7744 ;; declaration.
7747 ;; The declarator is a paren expression, so skip past it
7748 ;; so that we don't get stuck on that instead of the
7749 ;; function arglist.
7753 ;; Don't trip up on "operator ()".
7760 ;; Return the position of the first argument declaration if point is
7761 ;; inside a K&R style argument declaration list, nil otherwise.
7762 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
7763 ;; position that bounds the backward search for the argument list.
7764 ;;
7765 ;; Point must be within a possible K&R region, e.g. just before a top-level
7766 ;; "{". It must be outside of parens and brackets. The test can return
7767 ;; false positives otherwise.
7768 ;;
7769 ;; This function might do hidden buffer changes.
7773 ;; If we're in a macro, our search range is restricted to it. Narrow to
7774 ;; the searchable range.
7778 before-lparen after-rparen
7780 ; we give up
7783 ;; Search backwards for the defun's argument list. We give up if we
7784 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
7785 ;; a knr region) or BOB.
7786 ;;
7787 ;; The criterion for a paren structure being the arg list is:
7788 ;; o - there is non-WS stuff after it but before any "{"; AND
7789 ;; o - the token after it isn't a ";" AND
7790 ;; o - it is preceded by either an identifier (the function name) or
7791 ;; a macro expansion like "DEFUN (...)"; AND
7792 ;; o - its content is a non-empty comma-separated list of identifiers
7793 ;; (an empty arg list won't have a knr region).
7794 ;;
7795 ;; The following snippet illustrates these rules:
7796 ;; int foo (bar, baz, yuk)
7797 ;; int bar [] ;
7798 ;; int (*baz) (my_type) ;
7799 ;; int (*) (void) (*yuk) (void) ;
7800 ;; {
7811 ; parens/brackets.
7815 ;; We're inside a paren. Could it be our argument list....?
7821 ;; FIXME!!! What about macros between the parens? 2007/01/20
7824 ;; It can't be the arg list if next token is ; or {
7829 ;; Is the thing preceding the list an identifier (the
7830 ;; function name), or a macro expansion?
7840 ;; Have we got a non-empty list of comma-separated
7841 ;; identifiers?
7855 ;; ...Yes. We've identified the function's argument list.
7861 ;; ...No. The current parens aren't the function's arg list.
7868 ;; skip forward over conditional at point, including any predicate
7869 ;; statements in parentheses. No error checking is performed.
7870 ;;
7871 ;; This function might do hidden buffer changes.
7873 ;; else if()
7878 ;; do, else, try, finally
7883 ;; for, if, while, switch, catch, synchronized, foreach
7887 ;; If looking at the token after a conditional then return the
7888 ;; position of its start, otherwise return nil.
7889 ;;
7890 ;; This function might do hidden buffer changes.
7900 ;; If the point is after an operator identifier that isn't handled
7901 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
7902 ;; position of the start of that identifier is returned. nil is
7903 ;; returned otherwise. The point may be anywhere in the syntactic
7904 ;; whitespace after the last token of the operator identifier.
7905 ;;
7906 ;; This function might do hidden buffer changes.
7918 ;; Assuming point is at a brace that opens a statement block of some
7919 ;; kind, move to the proper anchor point for that block. It might
7920 ;; need to be adjusted further by c-add-stmt-syntax, but the
7921 ;; position at return is suitable as start position for that
7922 ;; function.
7923 ;;
7924 ;; This function might do hidden buffer changes.
7931 ;; Assuming point is at a brace that opens the block of a top level
7932 ;; declaration of some kind, move to the proper anchor point for
7933 ;; that block.
7934 ;;
7935 ;; This function might do hidden buffer changes.
7940 ;; Search forward for the end of the "header" of the current
7941 ;; declaration. That's the position where the definition body
7942 ;; starts, or the first variable initializer, or the ending
7943 ;; semicolon. I.e. search forward for the closest following
7944 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
7945 ;; _after_ the first found token, or at point-max if none is found.
7946 ;;
7947 ;; This function might do hidden buffer changes.
7952 ;; In C++ we need to take special care to handle operator
7953 ;; tokens and those pesky template brackets.
7958 ;; Handle operator identifiers, i.e. ignore any
7959 ;; operator token preceded by "operator".
7966 t
7968 nil)))))
7977 ;; Go to the beginning of the current declaration, or the beginning
7978 ;; of the previous one if already at the start of it. Point won't
7979 ;; be moved out of any surrounding paren. Return a cons cell of the
7980 ;; form (MOVE . KNR-POS). MOVE is like the return value from
7981 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
7982 ;; style argument declarations (and they are to be recognized) then
7983 ;; KNR-POS is set to the start of the first such argument
7984 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
7985 ;; position that bounds the backward search.
7986 ;;
7987 ;; NB: Cases where the declaration continues after the block, as in
7988 ;; "struct foo { ... } bar;", are currently recognized as two
7989 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
7990 ;;
7991 ;; This function might do hidden buffer changes.
7997 ;; `c-beginning-of-statement-1' stops at a block start, but we
7998 ;; want to continue if the block doesn't begin a top level
7999 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8000 ;; or an open paren.
8002 ;; Go back one "statement" each time round the loop until we're just
8003 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8004 ;; an ObjC method. This will move over a multiple declaration whose
8005 ;; components are comma separated.
8007 ;; Must check with c-opt-method-key in ObjC mode.
8017 ;; Check that we don't move from the first thing in a
8018 ;; macro to its header.
8024 move tentative-move))
8029 ;; Handle K&R argdecls. Back up after the "statement" jumped
8030 ;; over by `c-beginning-of-statement-1', unless it was the
8031 ;; function body, in which case we're sitting on the opening
8032 ;; brace now. Then test if we're in a K&R argdecl region and
8033 ;; that we started at the other side of the first argdecl in
8034 ;; it.
8044 'previous
8046 knr-argdecl-start))
8049 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8050 ;; statement, so the result will be 'previous if we've moved over any.
8051 ;; So change our result back to 'same if necessary.
8052 ;;
8053 ;; If they were brace list initializers we might not have moved over a
8054 ;; declaration boundary though, so change it to 'same if we've moved
8055 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8056 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8057 ;; potentially can search over a large amount of text.). Take special
8058 ;; pains not to get mislead by C++'s "operator=", and the like.
8061 c++-template-syntax-table
8067 ; no more, or get to one which isn't an "operator ="
8070 c-overloadable-operators-regexp
8071 c-opt-op-identifier-prefix
8086 ;; Assuming point is at the start of a declaration (as detected by
8087 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8088 ;; `c-beginning-of-decl-1', this function handles the case when a
8089 ;; block is followed by identifiers in e.g. struct declarations in C
8090 ;; or C++. If a proper end was found then t is returned, otherwise
8091 ;; point is moved as far as possible within the current sexp and nil
8092 ;; is returned. This function doesn't handle macros; use
8093 ;; `c-end-of-macro' instead in those cases.
8094 ;;
8095 ;; This function might do hidden buffer changes.
8098 c++-template-syntax-table
8106 ;; Stopped at the ';' in a K&R argdecl section which is
8107 ;; detected using the same criteria as in
8108 ;; `c-beginning-of-decl-1'. Move to the following block
8109 ;; start.
8113 ;; Encountered a block in the declaration. Jump over it.
8124 ;; Check for `c-opt-block-decls-with-vars-key'
8125 ;; before the first paren.
8128 c-opt-block-decls-with-vars-key
8130 lim t t t)
8133 ;; Check that the first following paren is
8134 ;; the block.
8136 lim t t t)
8138 ;; The declaration doesn't have any of the
8139 ;; `c-opt-block-decls-with-vars' keywords in the
8140 ;; beginning, so it ends here at the end of the block.
8148 nil)))
8151 ;; Assuming the point is at an open brace, check if it starts a
8152 ;; block that contains another declaration level, i.e. that isn't a
8153 ;; statement block or a brace list, and if so return non-nil.
8154 ;;
8155 ;; If the check is successful, the return value is the start of the
8156 ;; keyword that tells what kind of construct it is, i.e. typically
8157 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8158 ;; the point will be at the start of the construct, before any
8159 ;; leading specifiers, otherwise it's at the returned position.
8160 ;;
8161 ;; The point is clobbered if the check is unsuccessful.
8162 ;;
8163 ;; CONTAINING-SEXP is the position of the open of the surrounding
8164 ;; paren, or nil if none.
8165 ;;
8166 ;; The optional LIMIT limits the backward search for the start of
8167 ;; the construct. It's assumed to be at a syntactically relevant
8168 ;; position.
8169 ;;
8170 ;; If any template arglists are found in the searched region before
8171 ;; the open brace, they get marked with paren syntax.
8172 ;;
8173 ;; This function might do hidden buffer changes.
8180 ;; Could be at the end of a template arglist.
8191 ;; Note: Can't get bogus hits inside template arglists below since they
8192 ;; have gotten paren syntax above.
8194 ;; If `goto-start' is set we begin by searching for the
8195 ;; first possible position of a leading specifier list.
8196 ;; The `c-decl-block-key' search continues from there since
8197 ;; we know it can't match earlier.
8200 open-brace t t)
8202 t)
8203 t)
8210 ;; Found a keyword that can't be a type?
8213 ;; Can be a type too, in which case it's the return type of a
8214 ;; function (under the assumption that no declaration level
8215 ;; block construct starts with a type).
8218 ;; Jumped over a type, but it could be a declaration keyword
8219 ;; followed by the declared identifier that we've jumped over
8220 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8221 ;; then we should be at the declarator now, so check for a
8222 ;; valid declarator start.
8223 ;;
8224 ;; Note: This doesn't cope with the case when a declared
8225 ;; identifier is followed by e.g. '(' in a language where '('
8226 ;; also might be part of a declarator expression. Currently
8227 ;; there's no such language.
8231 ;; In Pike a list of modifiers may be followed by a brace
8232 ;; to make them apply to many identifiers. Note that the
8233 ;; match data will be empty on return in this case.
8239 ;; Use this variant to avoid yet another special regexp.
8243 t)))
8245 ;; Got a match.
8248 ;; Back up over any preceding specifiers and their clauses
8249 ;; by going forward from `first-specifier-pos', which is the
8250 ;; earliest possible position where the specifier list can
8251 ;; start.
8257 ;; Accept any plain symbol token on the ground that
8258 ;; it's a specifier masked through a macro (just
8259 ;; like `c-forward-decl-or-cast-1' skip forward over
8260 ;; such tokens).
8261 ;;
8262 ;; Could be more restrictive wrt invalid keywords,
8263 ;; but that'd only occur in invalid code so there's
8264 ;; no use spending effort on it.
8270 ;; Can't parse a declaration preamble and is still
8271 ;; before `kwd-start'. That means `first-specifier-pos'
8272 ;; was in some earlier construct. Search again.
8276 ;; Got no preamble before the block declaration keyword.
8282 kwd-start)))
8285 ;; Check if the closest containing paren sexp is a declaration
8286 ;; block, returning a 2 element vector in that case. Aref 0
8287 ;; contains the bufpos at boi of the class key line, and aref 1
8288 ;; contains the bufpos of the open brace. This function is an
8289 ;; obsolete wrapper for `c-looking-at-decl-block'.
8290 ;;
8291 ;; This function might do hidden buffer changes.
8299 nil))
8304 ;; Pull the next open brace from PS (which has the form of paren-state),
8305 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
8313 ;; Return the buffer position of the most enclosing decl-block brace (in the
8314 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8315 ;; none was found.
8324 open-brace))
8327 ;; Return the position of the L-brace if point is inside a brace list
8328 ;; initialization of an array, etc. This is an approximate function,
8329 ;; designed for speed over accuracy. It will not find every bracelist, but
8330 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8331 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8332 ;; is everywhere else.
8341 b-pos)))
8344 ;; return the buffer position of the beginning of the brace list
8345 ;; statement if we're inside a brace list, otherwise return nil.
8346 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8347 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8348 ;; braces
8349 ;;
8350 ;; N.B.: This algorithm can potentially get confused by cpp macros
8351 ;; placed in inconvenient locations. It's a trade-off we make for
8352 ;; speed.
8353 ;;
8354 ;; This function might do hidden buffer changes.
8356 ;; This will pick up brace list declarations.
8369 ;; this will pick up array/aggregate init lists, even if they are nested.
8372 ;; Pike can have class definitions anywhere, so we must
8373 ;; check for the class key here.
8375 c-decl-block-key))
8376 bufpos braceassignp lim next-containing)
8378 containing-sexp)
8389 ;; We're in an in-expression block of some kind. Do not
8390 ;; check nesting. We deliberately set the limit to the
8391 ;; containing sexp, so that c-looking-at-inexpr-block
8392 ;; doesn't check for an identifier before it.
8394 ;; see if the open brace is preceded by = or [...] in
8395 ;; this statement, but watch out for operator=
8398 ;; Checks to do only on the first sexp before the brace.
8401 ;; In Java, an initialization brace list may follow
8402 ;; directly after "new Foo[]", so check for a "new"
8403 ;; earlier.
8409 ;; Carry on looking if this is an
8410 ;; identifier (may contain "." in Java)
8411 ;; or another "[]" sexp.
8414 ;; Checks to do on all sexps before the brace, up to the
8415 ;; beginning of the statement.
8423 ;; We've seen a =, but must check earlier tokens so
8424 ;; that it isn't something that should be ignored.
8430 ;; Check for operator =
8433 nil)
8434 ;; Check for `<opchar>= in Pike.
8437 ;; Special case for Pikes
8438 ;; `[]=, since '[' is not in
8439 ;; the punctuation class.
8442 nil)
8444 ;; make sure we're not in a C++ template
8445 ;; argument assignment
8457 ))))
8458 nil)
8465 ;; Brace lists can't contain a semicolon, so we're done.
8467 ;; Go up one level.
8469 lim nil
8470 next-containing nil))
8471 ;; we've hit the beginning of the aggregate list
8475 )
8476 bufpos))
8477 ))
8480 ;; If we're looking at the start of a pike-style list, ie `({Â })',
8481 ;; `([Â ])', `(<Â >)' etc, a cons of a cons of its starting and ending
8482 ;; positions and its entry in c-special-brace-lists is returned, nil
8483 ;; otherwise. The ending position is nil if the list is still open.
8484 ;; LIM is the limit for forward search. The point may either be at
8485 ;; the `(' or at the following paren character. Tries to check the
8486 ;; matching closer, but assumes it's correct if no balanced paren is
8487 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8488 ;; a special brace list).
8489 ;;
8490 ;; This function might do hidden buffer changes.
8495 inner-beg end type)
8510 nil)))))
8520 ;; Check balancing of the inner paren
8521 ;; below.
8524 t)
8525 ;; If the inner char isn't a paren then
8526 ;; we can't check balancing, so just
8527 ;; check the char before the outer
8528 ;; closing paren.
8543 ;; Return non-nil if between two statements or declarations, assuming
8544 ;; point is not inside a literal or comment.
8545 ;;
8546 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8547 ;; are recommended instead.
8548 ;;
8549 ;; This function might do hidden buffer changes.
8554 ;; Return non-nil if we're looking at the beginning of a block
8555 ;; inside an expression. The value returned is actually a cons of
8556 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8557 ;; position of the beginning of the construct.
8558 ;;
8559 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8560 ;; position of the closest containing list. If it's nil, the
8561 ;; containing paren isn't used to decide whether we're inside an
8562 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8563 ;; needs to be farther back.
8564 ;;
8565 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8566 ;; brace block might be done. It should only be used when the
8567 ;; construct can be assumed to be complete, i.e. when the original
8568 ;; starting position was further down than that.
8569 ;;
8570 ;; This function might do hidden buffer changes.
8575 ;; Look at the character after point only as a last resort
8576 ;; when we can't disambiguate.
8596 ;; If the class definition is at the start of
8597 ;; a statement, we don't consider it an
8598 ;; in-expression class.
8606 ;; Also, in Pike we treat it as an
8607 ;; in-expression class if it's used in an
8608 ;; object clone expression.
8624 nil)
8632 ;; Accept several square bracket sexps for
8633 ;; Java array initializations.
8641 block-follows
8642 containing-sexp
8651 nil
8654 res))))
8657 ;; Returns non-nil if we're looking at the end of an in-expression
8658 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
8659 ;; PAREN-STATE is the paren state relevant at the current position.
8660 ;;
8661 ;; This function might do hidden buffer changes.
8663 ;; We currently only recognize a block.
8666 containing-sexp)
8675 paren-state)
8676 containing-sexp)))))
8679 ;; Is there a "virtual semicolon" at POS or point?
8680 ;; (See cc-defs.el for full details of "virtual semicolons".)
8681 ;;
8682 ;; This is true when point is at the last non syntactic WS position on the
8683 ;; line, there is a macro call last on the line, and this particular macro's
8684 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
8685 ;; semicolon.
8693 c-macro-with-semi-re
8696 ;; Check we've got nothing after this except comments and empty lines
8697 ;; joined by escaped EOLs.
8710 nil)
8718 ;; Check for one of the listed macros being before point.
8722 t))
8730 \f
8731 ;; `c-guess-basic-syntax' and the functions that precedes it below
8732 ;; implements the main decision tree for determining the syntactic
8733 ;; analysis of the current line of code.
8735 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
8736 ;; auto newline analysis.
8740 ;; BRACEPOS is the position of a brace in a construct like "namespace
8741 ;; Bar {". Return the anchor point in this construct; this is the
8742 ;; earliest symbol on the brace's line which isn't earlier than
8743 ;; "namespace".
8744 ;;
8745 ;; Currently (2007-08-17), "like namespace" means "matches
8746 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
8747 ;; or anything like that.
8757 ;; A simple function to prepend a new syntax element to
8758 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
8759 ;; should always be dynamically bound but since we read it first
8760 ;; we'll fail properly anyway if this function is misused.
8762 c-syntactic-context)))
8765 ;; Like `c-add-syntax' but appends to the end of the syntax list.
8766 ;; (Normally not necessary.)
8771 syntax-extra-args
8772 stop-at-boi-only
8773 containing-sexp
8774 paren-state)
8775 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
8776 ;; needed with further syntax elements of the types `substatement',
8777 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
8778 ;; `defun-block-intro'.
8779 ;;
8780 ;; Do the generic processing to anchor the given syntax symbol on
8781 ;; the preceding statement: Skip over any labels and containing
8782 ;; statements on the same line, and then search backward until we
8783 ;; find a statement or block start that begins at boi without a
8784 ;; label or comment.
8785 ;;
8786 ;; Point is assumed to be at the prospective anchor point for the
8787 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
8788 ;; skip past open parens and containing statements. Most of the added
8789 ;; syntax elements will get the same anchor point - the exception is
8790 ;; for an anchor in a construct like "namespace"[*] - this is as early
8791 ;; as possible in the construct but on the same line as the {.
8792 ;;
8793 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
8794 ;;
8795 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
8796 ;; syntax symbol. They are appended after the anchor point.
8797 ;;
8798 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
8799 ;; if the current statement starts there.
8800 ;;
8801 ;; Note: It's not a problem if PAREN-STATE "overshoots"
8802 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
8803 ;;
8804 ;; This function might do hidden buffer changes.
8807 ;; This is by far the most common case, so let's give it special
8808 ;; treatment.
8813 ;; Set when we're on a label, so that we don't stop there.
8814 ;; FIXME: To be complete we should check if we're on a label
8815 ;; now at the start.
8816 on-label)
8818 ;; Use point as the anchor point for "namespace", "extern", etc.
8822 syntax-extra-args)
8824 ;; Loop while we have to back out of containing blocks.
8829 ;; Loop while we have to back up statements.
8831 on-label
8834 ;; Skip past any comments that stands between the
8835 ;; statement start and boi.
8843 ;; Skip to the beginning of this statement or backward
8844 ;; another one.
8852 ;; If we didn't move we're at the start of a block and
8853 ;; have to continue outside it.
8862 ;; Special case to avoid deeper and deeper indentation
8863 ;; of "else if" clauses.
8864 )
8869 ;; If stop-at-boi-only is nil, we shouldn't back up
8870 ;; over previous or containing statements to try to
8871 ;; reach boi, so go back to the last position and
8872 ;; exit.
8879 ;; If we've moved into another statement then we
8880 ;; should no longer try to stop in the middle of a
8881 ;; line.
8884 ;; Record this as a substatement if we skipped up one
8885 ;; level.
8888 )))
8890 containing-sexp)
8892 ;; Now we have to go out of this block.
8895 ;; Don't stop in the middle of a special brace list opener
8896 ;; like "({".
8908 ;; Analyze the construct in front of the block we've stepped out
8909 ;; from and add the right syntactic element for it.
8912 step-type)
8915 ;; Stepped out of a parenthesis block, so we're in an
8916 ;; expression now.
8928 ;; Stepped out of an in-expression statement. This
8929 ;; syntactic element won't get an anchor pos.
8932 ;; A parenthesis normally belongs to an arglist.
8940 on-label nil))
8942 ;; Stepped out of a brace block.
8954 containing-sexp)))
8956 'defun-block-intro
8958 nil))
8962 c-other-decl-block-key-in-symbols-alist))
8969 ;; Always done if the open brace was at boi. The
8970 ;; c-beginning-of-statement-1 call above is necessary
8971 ;; anyway, to decide the type of block-intro to add.
8974 ))
8976 ;; Fill in the current point as the anchor for all the symbols
8977 ;; added above.
8986 )))
8989 containing-decl-open
8990 containing-decl-start
8991 containing-decl-kwd
8992 paren-state)
8993 ;; The inclass and class-close syntactic symbols are added in
8994 ;; several places and some work is needed to fix everything.
8995 ;; Therefore it's collected here.
8996 ;;
8997 ;; This function might do hidden buffer changes.
9002 containing-decl-open)
9004 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9005 ;; here, but we have to do like this for compatibility.
9015 char-after-ip
9016 beg-of-same-or-containing-stmt
9017 containing-sexp
9018 paren-state)
9019 ;; This function contains the decision tree reached through both
9020 ;; cases 18 and 10. It's a continued statement or top level
9021 ;; construct of some kind.
9022 ;;
9023 ;; This function might do hidden buffer changes.
9030 ;; (CASE A removed.)
9031 ;; CASE B: open braces for class or brace-lists
9038 ;; CASE B.1: class-open
9045 ;; CASE B.2: brace-list-open
9050 indent-point t t t)))
9051 ;; The most semantically accurate symbol here is
9052 ;; brace-list-open, but we normally report it simply as a
9053 ;; statement-cont. The reason is that one normally adjusts
9054 ;; brace-list-open for brace lists as top-level constructs,
9055 ;; and brace lists inside statements is a completely different
9056 ;; context. C.f. case 5A.3.
9059 ;; Turn off the dwim above when we're
9060 ;; analyzing the nature of the brace
9061 ;; for the auto newline feature.
9062 'brace-list-open
9064 nil nil
9065 containing-sexp paren-state))
9067 ;; CASE B.3: The body of a function declared inside a normal
9068 ;; block. Can occur e.g. in Pike and when using gcc
9069 ;; extensions, but watch out for macros followed by blocks.
9070 ;; C.f. cases E, 16F and 17G.
9076 ;; Turn off recognition of constructs that lacks a
9077 ;; type in this case, since that's more likely to be
9078 ;; a macro followed by a block.
9081 containing-sexp paren-state))
9083 ;; CASE B.4: Continued statement with block open. The most
9084 ;; accurate analysis is perhaps `statement-cont' together with
9085 ;; `block-open' but we play DWIM and use `substatement-open'
9086 ;; instead. The rationale is that this typically is a macro
9087 ;; followed by a block which makes it very similar to a
9088 ;; statement with a substatement block.
9091 containing-sexp paren-state))
9092 ))
9094 ;; CASE C: iostream insertion or extraction operator
9098 ;; If there is no preceding streamop in the statement
9099 ;; then indent this line as a normal statement-cont.
9103 t))))
9105 ;; CASE E: In the "K&R region" of a function declared inside a
9106 ;; normal block. C.f. case B.3.
9108 ;; Check that the next token is a '{'. This works as
9109 ;; long as no language that allows nested function
9110 ;; definitions allows stuff like member init lists, K&R
9111 ;; declarations or throws clauses there.
9112 ;;
9113 ;; Note that we do a forward search for something ahead
9114 ;; of the indentation line here. That's not good since
9115 ;; the user might not have typed it yet. Unfortunately
9116 ;; it's exceedingly tricky to recognize a function
9117 ;; prototype in a code block without resorting to this.
9125 ;; Turn off recognition of constructs that lacks a
9126 ;; type in this case, since that's more likely to be
9127 ;; a macro followed by a block.
9130 containing-sexp paren-state))
9132 ;;CASE F: continued statement and the only preceding items are
9133 ;;annotations.
9141 t)
9148 ;; CASE G: a template list continuation?
9149 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9167 ;; In a nested template arglist.
9173 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9174 ;; template aware.
9177 ;; CASE D: continued statement.
9181 containing-sexp paren-state))
9182 )))
9184 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9185 ;; 2005/11/29).
9186 ;;;###autoload
9188 "Return the syntactic context of the current line."
9194 ;; A whole ugly bunch of various temporary variables. Have
9195 ;; to declare them here since it's not possible to declare
9196 ;; a variable with only the scope of a cond test and the
9197 ;; following result clauses, and most of this function is a
9198 ;; single gigantic cond. :P
9199 literal char-before-ip before-ws-ip char-after-ip macro-start
9200 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9201 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9202 containing-<
9203 ;; The following record some positions for the containing
9204 ;; declaration block if we're directly within one:
9205 ;; `containing-decl-open' is the position of the open
9206 ;; brace. `containing-decl-start' is the start of the
9207 ;; declaration. `containing-decl-kwd' is the keyword
9208 ;; symbol of the keyword that tells what kind of block it
9209 ;; is.
9210 containing-decl-open
9211 containing-decl-start
9212 containing-decl-kwd
9213 ;; The open paren of the closest surrounding sexp or nil if
9214 ;; there is none.
9215 containing-sexp
9216 ;; The position after the closest preceding brace sexp
9217 ;; (nested sexps are ignored), or the position after
9218 ;; `containing-sexp' if there is none, or (point-min) if
9219 ;; `containing-sexp' is nil.
9220 lim
9221 ;; The paren state outside `containing-sexp', or at
9222 ;; `indent-point' if `containing-sexp' is nil.
9224 ;; There's always at most one syntactic element which got
9225 ;; an anchor pos. It's stored in syntactic-relpos.
9226 syntactic-relpos
9229 ;; Check if we're directly inside an enclosing declaration
9230 ;; level block.
9239 containing-sexp)
9240 t)))
9243 containing-sexp nil)
9248 ;; Init some position variables.
9257 ;; Ignore balanced paren. The next entry
9258 ;; can't be another one.
9261 ;; If there is no surrounding open paren then
9262 ;; put the last balanced pair back on paren-state.
9264 containing-sexp nil)))
9272 ;; If we're in a parenthesis list then ',' delimits the
9273 ;; "statements" rather than being an operator (with the
9274 ;; exception of the "for" clause). This difference is
9275 ;; typically only noticeable when statements are used in macro
9276 ;; arglists.
9280 ;; cache char before and after indent point, and move point to
9281 ;; the most likely position to perform the majority of tests
9290 ;; are we in a literal?
9293 ;; now figure out syntactic qualities of the current line
9296 ;; CASE 1: in a string.
9300 ;; CASE 2: in a C or C++ style comment.
9302 ;; This is a kludge for XEmacs where we use
9303 ;; `buffer-syntactic-context', which doesn't correctly
9304 ;; recognize "\*/" to end a block comment.
9305 ;; `parse-partial-sexp' which is used by
9306 ;; `c-literal-limits' will however do that in most
9307 ;; versions, which results in that we get nil from
9308 ;; `c-literal-limits' even when `c-in-literal' claims
9309 ;; we're inside a comment.
9313 ;; CASE 3: in a cpp preprocessor macro continuation.
9323 ;; If at the beginning of the body of a #define
9324 ;; directive then analyze as cpp-define-intro
9325 ;; only. Go on with the syntactic analysis
9326 ;; otherwise. in-macro-expr is set if we're in a
9327 ;; cpp expression, i.e. before the #define body
9328 ;; or anywhere in a non-#define directive.
9335 nil)))))
9339 ;; CASE 11: an else clause?
9343 containing-sexp paren-state))
9345 ;; CASE 12: while closure of a do/while construct?
9353 containing-sexp paren-state))
9355 ;; CASE 13: A catch or finally clause? This case is simpler
9356 ;; than if-else and do-while, because a block is required
9357 ;; after every try, catch and finally.
9365 t)
9369 t)
9372 t))
9377 containing-sexp paren-state))
9379 ;; CASE 18: A substatement we can recognize by keyword.
9389 ;; Ought to cache the result from the
9390 ;; c-beginning-of-statement-1 calls here.
9400 ;; Step up to the containing statement if we
9401 ;; stayed in the same one.
9409 ;; There was no containing statement after all.
9411 placeholder))
9413 ;; Require a parenthesis after these keywords.
9414 ;; Necessary to catch e.g. synchronized in Java,
9415 ;; which can be used both as statement and
9416 ;; modifier.
9422 ;; CASE 18A: Simple substatement.
9428 containing-sexp paren-state))
9434 containing-sexp paren-state))
9437 containing-sexp paren-state))))
9439 ;; CASE 18B: Some other substatement. This is shared
9440 ;; with case 10.
9442 char-after-ip
9443 placeholder
9444 lim
9445 paren-state)))
9447 ;; CASE 14: A case or default label
9453 containing-sexp))
9456 ;; Got a bogus label at the top level. In lack of better
9457 ;; alternatives, anchor it on (point-min).
9460 ;; CASE 15: any other label
9467 containing-decl-open
9468 containing-decl-start
9469 containing-decl-kwd
9470 paren-state))
9471 ;; Append access-label with the same anchor point as
9472 ;; inclass gets.
9478 containing-sexp))
9483 ;; If the surrounding statement is a switch then
9484 ;; let's analyze all labels as switch labels, so
9485 ;; that they get lined up consistently.
9486 'case-label
9492 ;; A label on the top level. Treat it as a class
9493 ;; context. (point-min) is the closest we get to the
9494 ;; class open brace.
9497 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9498 ;; 17E.
9501 containing-sexp
9502 ;; Have to turn on the heuristics after
9503 ;; the point even though it doesn't work
9504 ;; very well. C.f. test case class-16.pike.
9505 t))
9510 ;; It's a statement block or an anonymous class.
9512 ;; It's a Pike lambda. Check whether we are between the
9513 ;; lambda keyword and the argument list or at the defun
9514 ;; opener.
9516 'inline-open
9522 paren-state)
9526 ;; CASE 5: Line is inside a declaration level block or at top level.
9530 ;; CASE 5A: we are looking at a defun, brace list, class,
9531 ;; or inline-inclass method opening brace
9538 ;; CASE 5A.1: Non-class declaration block open.
9553 ;; Special case for extern-lang-open.
9554 'extern-lang-open
9556 nil t containing-sexp paren-state))
9558 ;; CASE 5A.2: we are looking at a class opening brace
9567 ;; CASE 5A.3: brace list open
9592 ))))
9596 ;; We're in Java and have found that the open brace
9597 ;; belongs to a "new Foo[]" initialization list,
9598 ;; which means the brace list is part of an
9599 ;; expression and not a top level definition. We
9600 ;; therefore treat it as any topmost continuation
9601 ;; even though the semantically correct symbol still
9602 ;; is brace-list-open, on the same grounds as in
9603 ;; case B.2.
9609 ;; CASE 5A.4: inline defun open
9615 containing-decl-open
9616 containing-decl-start
9617 containing-decl-kwd
9618 paren-state))
9620 ;; CASE 5A.5: ordinary defun open
9628 ;; Bogus to use bol here, but it's the legacy. (Resolved,
9629 ;; 2007-11-09)
9630 ))))
9632 ;; CASE 5B: After a function header but before the body (or
9633 ;; the ending semicolon if there's no body).
9640 ;; CASE 5B.1: Member init list.
9645 ;; There is no preceding member init clause.
9646 ;; Indent relative to the beginning of indentation
9647 ;; for the topmost-intro line that contains the
9648 ;; prototype's open paren.
9651 ;; Indent relative to the first member init clause.
9656 ;; CASE 5B.2: K&R arg decl intro
9663 containing-decl-open
9664 containing-decl-start
9665 containing-decl-kwd
9666 paren-state)))
9668 ;; CASE 5B.4: Nether region after a C++ or Java func
9669 ;; decl, which could include a `throws' declaration.
9673 )))
9675 ;; CASE 5C: inheritance line. could be first inheritance
9676 ;; line, or continuation of a multiple inheritance
9684 ;; watch out for scope operator
9689 )))
9698 ;; for Java
9703 cont done)
9708 done t))
9712 )
9714 injava-inher)
9717 ))
9720 ;; CASE 5C.1: non-hanging colon on an inher intro
9724 ;; don't add inclass symbol since relative point already
9725 ;; contains any class offset
9726 )
9728 ;; CASE 5C.2: hanging colon on an inher intro
9734 containing-decl-open
9735 containing-decl-start
9736 containing-decl-kwd
9737 paren-state)))
9739 ;; CASE 5C.3: in a Java implements/extends
9753 )))
9755 ;; CASE 5C.4: a continued inheritance line
9759 ;; don't add inclass symbol since relative point already
9760 ;; contains any class offset
9761 )))
9763 ;; CASE 5D: this could be a top-level initialization, a
9764 ;; member init list continuation, or a template argument
9765 ;; list continuation.
9767 ;; Note: We use the fact that lim is always after any
9768 ;; preceding brace sexp.
9778 t))
9783 t)
9789 nil)
9790 t))
9792 ;; NB: No c-after-special-operator-id stuff in this
9793 ;; clause - we assume only C++ needs it.
9798 ;; CASE 5D.3: perhaps a template list continuation?
9813 ;; In a nested template arglist.
9819 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9820 ;; template aware.
9823 ;; CASE 5D.4: perhaps a multiple inheritance line?
9835 t)
9840 ;; CASE 5D.5: Continuation of the "expression part" of a
9841 ;; top level construct. Or, perhaps, an unrecognized construct.
9853 ;; A preceding comma at the top level means that a
9854 ;; new variable declaration starts here. Use
9855 ;; topmost-intro-cont for it, for consistency with
9856 ;; the first variable declaration. C.f. case 5N.
9859 nil nil containing-sexp paren-state))
9860 ))
9862 ;; CASE 5F: Close of a non-class declaration level block.
9866 ;; This is inconsistent: Should use `containing-decl-open'
9867 ;; here if it's at boi, like in case 5J.
9871 ;; Special case for compatibility with the
9872 ;; extern-lang syntactic symbols.
9873 'extern-lang-close
9876 nil t
9878 paren-state))
9880 ;; CASE 5G: we are looking at the brace which closes the
9881 ;; enclosing nested class decl
9886 containing-decl-open
9887 containing-decl-start
9888 containing-decl-kwd
9889 paren-state))
9891 ;; CASE 5H: we could be looking at subsequent knr-argdecls
9898 ;; Do an extra check to avoid tripping up on
9899 ;; statements that occur in invalid contexts
9900 ;; (e.g. in macro bodies where we don't really
9901 ;; know the context of what we're looking at).
9908 ;; CASE 5I: ObjC method definition.
9913 ;; Handle the case when it's the first (non-comment)
9914 ;; thing in the buffer. Can't look for a 'same return
9915 ;; value from cbos1 since ObjC directives currently
9916 ;; aren't recognized fully, so that we get 'same
9917 ;; instead of 'previous if it moved over a preceding
9918 ;; directive.
9922 ;; CASE 5P: AWK pattern or function or continuation
9923 ;; thereof.
9929 'topmost-intro-cont
9931 nil nil
9932 containing-sexp paren-state))
9934 ;; CASE 5N: At a variable declaration that follows a class
9935 ;; definition or some other block declaration that doesn't
9936 ;; end at the closing '}'. C.f. case 5D.5.
9945 ;; Speed up the backward search a bit.
9950 ;; The '}' is unbalanced.
9951 nil
9956 containing-sexp paren-state))
9958 ;; NOTE: The point is at the end of the previous token here.
9960 ;; CASE 5J: we are at the topmost level, make
9961 ;; sure we skip back past any access specifiers
9963 ;; A macro continuation line is never at top level.
9985 nil)))))
9986 ;; For historic reasons we anchor at bol of the last
9987 ;; line of the previous declaration. That's clearly
9988 ;; highly bogus and useless, and it makes our lives hard
9989 ;; to remain compatible. :P
10000 ;; Special case for compatibility with the
10001 ;; extern-lang syntactic symbols.
10002 'inextern-lang
10005 nil t
10007 paren-state))
10009 containing-decl-open
10010 containing-decl-start
10011 containing-decl-kwd
10012 paren-state)))
10014 macro-start
10019 ;; CASE 5K: we are at an ObjC method definition
10020 ;; continuation line.
10029 ;; CASE 5L: we are at the first argument of a template
10030 ;; arglist that begins on the previous line.
10038 ;; CASE 5Q: we are at a statement within a macro.
10043 ;;CASE 5N: We are at a topmost continuation line and the only
10044 ;;preceding items are annotations.
10051 t)
10057 ;; CASE 5M: we are at a topmost continuation line
10068 ))
10071 ;; (CASE 6 has been removed.)
10073 ;; CASE 7: line is an expression, not a statement. Most
10074 ;; likely we are either in a function prototype or a function
10075 ;; call argument list
10083 ;; CASE 7A: we are looking at the arglist closing paren.
10084 ;; C.f. case 7F.
10096 paren-state))
10098 ;; CASE 7B: Looking at the opening brace of an
10099 ;; in-expression block or brace list. C.f. cases 4, 16A
10100 ;; and 17E.
10104 paren-state))
10108 placeholder
10111 paren-state)
10112 containing-sexp)))
10113 ;; placeholder is nil if it's a block directly in
10114 ;; a function arglist. That makes us skip out of
10115 ;; this case.
10116 )))
10121 paren-state)
10125 ;; CASE 7C: we are looking at the first argument in an empty
10126 ;; argument list. Use arglist-close if we're actually
10127 ;; looking at a close paren or bracket.
10139 paren-state))
10141 ;; CASE 7D: we are inside a conditional test clause. treat
10142 ;; these things as statements
10152 ))
10154 ;; CASE 7E: maybe a continued ObjC method call. This is the
10155 ;; case when we are inside a [] bracketed exp, and what
10156 ;; precede the opening bracket is not an identifier.
10164 )))
10166 ;; CASE 7F: we are looking at an arglist continuation line,
10167 ;; but the preceding argument is on the same line as the
10168 ;; opening paren. This case includes multi-line
10169 ;; mathematical paren groupings, but we could be on a
10170 ;; for-list continuation line. C.f. case 7A.
10187 paren-state))
10189 ;; CASE 7G: we are looking at just a normal arglist
10190 ;; continuation line
10193 ))
10195 ;; CASE 8: func-local multi-inheritance line
10205 ;; CASE 8A: non-hanging colon on an inher intro
10210 ;; CASE 8B: hanging colon on an inher intro
10214 ;; CASE 8C: a continued inheritance line
10218 )))
10220 ;; CASE 9: we are inside a brace-list
10230 ;; CASE 9A: In the middle of a special brace list opener.
10250 ;; CASE 9B: brace-list-close brace
10252 ;; Check special brace list closer.
10259 ;; We were between the special close char and the `)'.
10262 ;; We were before the special close char.
10266 ;; Normal brace list check.
10277 ;; Prepare for the rest of the cases below by going to the
10278 ;; token following the opening brace
10291 ;; CASE 9C: we're looking at the first line in a brace-list
10302 ;; CASE 9D: this is just a later brace-list-entry or
10303 ;; brace-entry-open
10312 ))
10313 ))))
10315 ;; CASE 10: A continued statement or top level construct.
10325 ;; This is shared with case 18.
10327 char-after-ip
10328 placeholder
10329 containing-sexp
10330 paren-state))
10332 ;; CASE 16: block close brace, possibly closing the defun or
10333 ;; the class
10335 ;; From here on we have the next containing sexp in lim.
10340 ;; CASE 16E: Closing a statement block? This catches
10341 ;; cases where it's preceded by a statement keyword,
10342 ;; which works even when used in an "invalid" context,
10343 ;; e.g. a macro argument.
10348 ;; CASE 16A: closing a lambda defun or an in-expression
10349 ;; block? C.f. cases 4, 7B and 17E.
10352 nil))
10354 'inline-close
10364 paren-state)
10368 ;; CASE 16B: does this close an inline or a function in
10369 ;; a non-class declaration level block?
10376 nil))
10386 ;; CASE 16F: Can be a defun-close of a function declared
10387 ;; in a statement block, e.g. in Pike or when using gcc
10388 ;; extensions, but watch out for macros followed by
10389 ;; blocks. Let it through to be handled below.
10390 ;; C.f. cases B.3 and 17G.
10396 ;; Turn off recognition of constructs that
10397 ;; lacks a type in this case, since that's more
10398 ;; likely to be a macro followed by a block.
10405 ;; CASE 16C: If there is an enclosing brace then this is
10406 ;; a block close since defun closes inside declaration
10407 ;; level blocks have been handled above.
10409 ;; If the block is preceded by a case/switch label on
10410 ;; the same line, we anchor at the first preceding label
10411 ;; at boi. The default handling in c-add-stmt-syntax
10412 ;; really fixes it better, but we do like this to keep
10413 ;; the indentation compatible with version 5.28 and
10414 ;; earlier. C.f. case 17H.
10421 ;; c-backward-to-block-anchor not necessary here; those
10422 ;; situations are handled in case 16E above.
10425 ;; CASE 16D: Only top level defun close left.
10431 paren-state))
10432 ))
10434 ;; CASE 19: line is an expression, not a statement, and is directly
10435 ;; contained by a template delimiter. Most likely, we are in a
10436 ;; template arglist within a statement. This case is based on CASE
10437 ;; 7. At some point in the future, we may wish to create more
10438 ;; syntactic symbols such as `template-intro',
10439 ;; `template-cont-nonempty', etc., and distinguish between them as we
10440 ;; do for `arglist-intro' etc. (2009-12-07).
10446 ; '<') before indent-point.
10454 paren-state))
10456 ;; CASE 17: Statement or defun catchall.
10459 ;; Back up statements until we find one that starts at boi.
10462 containing-sexp)))
10466 nil)
10471 ;; CASE 17B: continued statement
10475 containing-sexp paren-state))
10477 ;; CASE 17A: After a case/default label?
10485 'statement-case-open
10487 nil t containing-sexp paren-state))
10489 ;; CASE 17D: any old statement
10496 containing-sexp paren-state)
10500 ;; CASE 17I: Inside a substatement block.
10502 ;; The following tests are all based on containing-sexp.
10504 ;; From here on we have the next containing sexp in lim.
10509 lim paren-state)
10513 ;; CASE 17E: first statement in an in-expression block.
10514 ;; C.f. cases 4, 7B and 16A.
10517 nil))
10519 'defun-block-intro
10528 paren-state)
10534 ;; CASE 17F: first statement in an inline, or first
10535 ;; statement in a top-level defun. we can tell this is it
10536 ;; if there are no enclosing braces that haven't been
10537 ;; narrowed out by a class (i.e. don't use bod here).
10540 paren-state)))
10546 nil))))
10551 ;; CASE 17G: First statement in a function declared inside
10552 ;; a normal block. This can occur in Pike and with
10553 ;; e.g. the gcc extensions, but watch out for macros
10554 ;; followed by blocks. C.f. cases B.3 and 16F.
10560 ;; Turn off recognition of constructs that lacks
10561 ;; a type in this case, since that's more likely
10562 ;; to be a macro followed by a block.
10568 lim paren-state))
10570 ;; CASE 17H: First statement in a block.
10572 ;; If the block is preceded by a case/switch label on the
10573 ;; same line, we anchor at the first preceding label at
10574 ;; boi. The default handling in c-add-stmt-syntax is
10575 ;; really fixes it better, but we do like this to keep the
10576 ;; indentation compatible with version 5.28 and earlier.
10577 ;; C.f. case 16C.
10584 ;; c-backward-to-block-anchor not necessary here; those
10585 ;; situations are handled in case 17I above.
10587 lim paren-state))
10590 ))
10591 )
10593 ;; now we need to look at any modifiers
10597 ;; are we looking at a comment only line?
10602 ;; we might want to give additional offset to friends (in C++).
10607 ;; Set syntactic-relpos.
10613 nil)
10614 t))
10617 ;; Start of or a continuation of a preprocessor directive?
10632 ;; If inside a cpp expression, i.e. anywhere in a
10633 ;; cpp directive except a #define body, we only let
10634 ;; through the syntactic analysis that is internal
10635 ;; in the expression. That means the arglist
10636 ;; elements, if they are anchored inside the cpp
10637 ;; expression.
10646 ;; Inside a #define body and the syntactic analysis is
10647 ;; anchored on the start of the #define. In this case
10648 ;; we add cpp-define-intro to get the extra
10649 ;; indentation of the #define body.
10652 ;; return the syntax
10653 c-syntactic-context)))
10655 \f
10656 ;; Indentation calculation.
10659 ;; offset can be a number, a function, a variable, a list, or one of
10660 ;; the symbols + or -
10661 ;;
10662 ;; This function might do hidden buffer changes.
10681 langelem symbol))
10687 offset symbol)
10688 nil)
10702 Error evaluating offset %S for %s: \
10703 Cannot combine absolute offset %S with relative %S in `%s' method"
10709 Error evaluating offset %S for %s: \
10710 Cannot combine relative offset %S with absolute %S in `%s' method"
10715 res))
10733 Error evaluating offset %S for %s: \
10734 Cannot combine absolute offsets %S and %S in `add' method"
10738 res))
10747 res))))
10754 nil))))
10758 res
10760 offset symbol res)
10761 nil)))
10764 ;; Get offset from LANGELEM which is a list beginning with the
10765 ;; syntactic symbol and followed by any analysis data it provides.
10766 ;; That data may be zero or more elements, but if at least one is
10767 ;; given then the first is the anchor position (or nil). The symbol
10768 ;; is matched against `c-offsets-alist' and the offset calculated
10769 ;; from that is returned.
10770 ;;
10771 ;; This function might do hidden buffer changes.
10781 offset
10785 ))
10788 ;; This is a compatibility wrapper for `c-calc-offset' in case
10789 ;; someone is calling it directly. It takes an old style syntactic
10790 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
10791 ;; new list form.
10792 ;;
10793 ;; This function might do hidden buffer changes.
10800 ;; Calculate the syntactic indentation from a syntactic description
10801 ;; as returned by `c-guess-syntax'.
10802 ;;
10803 ;; Note that topmost-intro always has an anchor position at bol, for
10804 ;; historical reasons. It's often used together with other symbols
10805 ;; that has more sane positions. Since we always use the first
10806 ;; found anchor position, we rely on that these other symbols always
10807 ;; precede topmost-intro in the LANGELEMS list.
10808 ;;
10809 ;; This function might do hidden buffer changes.
10817 ;; Got an absolute column that overrides any indentation
10818 ;; we've collected so far, but not the relative
10819 ;; indentation we might get for the nested structures
10820 ;; further down the langelems list.
10824 ;; Got a relative change of the current calculated
10825 ;; indentation.
10828 ;; Use the anchor position from the first syntactic
10829 ;; element with one.
10839 indent)))
10841 \f
10844 ;;; cc-engine.el ends here