| blob | 9188ab2fbfd4eec13b61a61302a187c02cd5cf8f |
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.
1698 ;; The following search is the main reason that `c-in-sws'
1699 ;; and `c-is-sws' aren't combined to one property.
1703 ;; If the `c-in-sws' region extended past the last
1704 ;; `c-is-sws' char we have to go back a bit.
1711 "c-forward-sws cached move %s -> %s (max %s)"
1718 ;; We'll loop here if there is simple ws after the last rung.
1719 ;; That means that there's been some change in it and it's
1720 ;; possible that we've stepped into another ladder, so extend
1721 ;; the previous one to join with it if there is one, and try to
1722 ;; use the cache again.
1724 "c-forward-sws extending rung with [%s..%s] (max %s)"
1727 ;; Remove any `c-in-sws' property from the last char of
1728 ;; the rung before we mark it with `c-is-sws', so that we
1729 ;; won't connect with the remains of a broken "ladder".
1735 last-put-in-sws-pos rung-pos)))
1742 ;; Skipped over comments. Don't cache at eob in case the buffer
1743 ;; is narrowed.
1754 ;; Skip a preprocessor directive.
1761 ;; Don't cache at eob in case the buffer is narrowed.
1764 ;; We've searched over a piece of non-white syntactic ws. See if this
1765 ;; can be cached.
1771 ;; Cache if we haven't skipped comments only, and if we started
1772 ;; either from a marked rung or from a completely uncached
1773 ;; position.
1778 ;; See if there's a marked rung in the encountered simple ws. If
1779 ;; so then we can cache, unless `safe-start' is nil. Even then
1780 ;; we need to do this to check if the cache can be used for the
1781 ;; next step.
1785 safe-start))
1789 "c-forward-sws caching [%s..%s] - [%s..%s] (max %s)"
1793 ;; Remove the properties for any nested ws that might be cached.
1794 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
1795 ;; anyway.
1803 last-put-in-sws-pos rung-pos))
1805 ;; Remove any `c-in-sws' property from the last char of
1806 ;; the rung before we mark it with `c-is-sws', so that we
1807 ;; won't connect with the remains of a broken "ladder".
1810 rung-end-pos))
1813 "c-forward-sws not caching [%s..%s] - [%s..%s] (max %s)"
1817 ;; Set `rung-pos' for the next rung. It's the same thing here as
1818 ;; initially, except that the rung position is set as early as
1819 ;; possible since we can't be in the ending ws of a line comment or
1820 ;; cpp directive now.
1827 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
1828 ;; another one after the point (which might occur when editing inside a
1829 ;; comment or macro).
1833 "c-forward-sws clearing at %s for cache separation"
1834 last-put-in-sws-pos)
1838 ;; If at eob we have to clear the last character before the end
1839 ;; instead since the buffer might be narrowed and there might
1840 ;; be a `c-in-sws' after (point-max). In this case it's
1841 ;; necessary to clear both properties.
1843 "c-forward-sws clearing thoroughly at %s for cache separation"
1846 last-put-in-sws-pos))))
1847 ))))
1850 ;; Used by `c-backward-syntactic-ws' to implement the unbounded search.
1851 ;;
1852 ;; This function might do hidden buffer changes.
1855 ;; part of the simple ws region.
1857 rung-is-marked simple-ws-beg cmt-skip-pos)
1859 ;; Skip simple horizontal ws and do a quick check on the preceding
1860 ;; character to see if it's anything that can't end syntactic ws, so we can
1861 ;; bail out early in the majority of cases when there just are a few ws
1862 ;; chars. Newlines are complicated in the backward direction, so we can't
1863 ;; skip over them.
1870 ;; Try to find a rung position in the simple ws preceding point, so that
1871 ;; we can get a cache hit even if the last bit of the simple ws has
1872 ;; changed recently.
1878 ;; `rung-pos' will be the earliest marked position, which means that
1879 ;; there might be later unmarked parts in the simple ws region.
1880 ;; It's not worth the effort to fix that; the last part of the
1881 ;; simple ws is also typically edited often, so it could be wasted.
1893 ;; The following search is the main reason that `c-in-sws'
1894 ;; and `c-is-sws' aren't combined to one property.
1898 ;; If the `c-in-sws' region extended past the first
1899 ;; `c-is-sws' char we have to go forward a bit.
1904 "c-backward-sws cached move %s <- %s (min %s)"
1916 t
1918 nil))
1920 ;; We'll loop here if there is simple ws before the first rung.
1921 ;; That means that there's been some change in it and it's
1922 ;; possible that we've stepped into another ladder, so extend
1923 ;; the previous one to join with it if there is one, and try to
1924 ;; use the cache again.
1926 "c-backward-sws extending rung with [%s..%s] (min %s)"
1929 ;; Remove any `c-in-sws' property from the last char of
1930 ;; the rung before we mark it with `c-is-sws', so that we
1931 ;; won't connect with the remains of a broken "ladder".
1934 rung-pos)
1938 last-put-in-sws-pos rung-pos))
1947 ;; Inside a cpp directive. See if it should be skipped over.
1950 ;; Move back over all line continuations in the region skipped
1951 ;; over by `c-backward-comments'. If we go past it then we
1952 ;; started inside the cpp directive.
1961 ;; Don't move past the cpp directive if we began inside
1962 ;; it. Note that the position at the end of the last line
1963 ;; of the macro is also considered to be within it.
1965 nil)
1967 ;; It's worthwhile to spend a little bit of effort on finding
1968 ;; the end of the macro, to get a good `simple-ws-beg'
1969 ;; position for the cache. Note that `c-backward-comments'
1970 ;; could have stepped over some comments before going into
1971 ;; the macro, and then `simple-ws-beg' must be kept on the
1972 ;; same side of those comments.
1979 ;; Might happen if comments after the macro were skipped
1980 ;; over.
1984 t)))
1989 simple-ws-beg)
1990 ;; Skipped over comments. Must put point at the end of
1991 ;; the simple ws at point since we might be after a line
1992 ;; comment or cpp directive that's been partially
1993 ;; narrowed out, and we can't risk marking the simple ws
1994 ;; at the end of it.
1996 t)))
1998 ;; We've searched over a piece of non-white syntactic ws. See if this
1999 ;; can be cached.
2004 ;; Cache if we started either from a marked rung or from a
2005 ;; completely uncached position.
2006 rung-is-marked
2009 ;; Cache if there's a marked rung in the encountered simple ws.
2017 "c-backward-sws caching [%s..%s] - [%s..%s] (min %s)"
2022 ;; Remove the properties for any nested ws that might be cached.
2023 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
2024 ;; anyway.
2029 ;; Remove any `c-in-sws' property from the last char of
2030 ;; the rung before we mark it with `c-is-sws', so that we
2031 ;; won't connect with the remains of a broken "ladder".
2034 rung-end-pos)
2037 last-put-in-sws-pos simple-ws-beg)
2038 rung-pos)
2043 "c-backward-sws not caching [%s..%s] - [%s..%s] (min %s)"
2049 ))
2051 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
2052 ;; another one before the point (which might occur when editing inside a
2053 ;; comment or macro).
2057 "c-backward-sws clearing at %s for cache separation"
2060 last-put-in-sws-pos))
2062 ;; If at bob and the buffer is narrowed, we have to clear the
2063 ;; character we're standing on instead since there might be a
2064 ;; `c-in-sws' before (point-min). In this case it's necessary
2065 ;; to clear both properties.
2067 "c-backward-sws clearing thoroughly at %s for cache separation"
2068 last-put-in-sws-pos)
2071 ))))
2073 \f
2074 ;; Other whitespace tools
2076 ;; Is the region (beg end) WS, and is there WS (or BOB/EOB) next to the
2077 ;; region? This is a "heuristic" function. .....
2078 ;;
2079 ;; The motivation for the second bit is to check whether removing this
2080 ;; region would coalesce two symbols.
2081 ;;
2082 ;; FIXME!!! This function doesn't check virtual semicolons in any way. Be
2083 ;; careful about using this function for, e.g. AWK. (2007/3/7)
2092 \f
2093 ;; A system for finding noteworthy parens before the point.
2096 ;; A maximum comfortable scanning distance, e.g. between
2097 ;; `c-state-cache-good-pos' and "HERE" (where we call c-parse-state). When
2098 ;; this distance is exceeded, we take "emergency measures", e.g. by clearing
2099 ;; the cache and starting again from point-min or a beginning of defun. This
2100 ;; value can be tuned for efficiency or set to a lower value for testing.
2104 ;; The state cache used by `c-parse-state' to cut down the amount of
2105 ;; searching. It's the result from some earlier `c-parse-state' call. See
2106 ;; `c-parse-state''s doc string for details of its structure.
2107 ;;
2108 ;; The use of the cached info is more effective if the next
2109 ;; `c-parse-state' call is on a line close by the one the cached state
2110 ;; was made at; the cache can actually slow down a little if the
2111 ;; cached state was made very far back in the buffer. The cache is
2112 ;; most effective if `c-parse-state' is used on each line while moving
2113 ;; forward.
2117 ;; This is a position where `c-state-cache' is known to be correct, or
2118 ;; nil (see below). It's a position inside one of the recorded unclosed
2119 ;; parens or the top level, but not further nested inside any literal or
2120 ;; subparen that is closed before the last recorded position.
2121 ;;
2122 ;; The exact position is chosen to try to be close to yet earlier than
2123 ;; the position where `c-state-cache' will be called next. Right now
2124 ;; the heuristic is to set it to the position after the last found
2125 ;; closing paren (of any type) before the line on which
2126 ;; `c-parse-state' was called. That is chosen primarily to work well
2127 ;; with refontification of the current line.
2128 ;;
2129 ;; 2009-07-28: When `c-state-point-min' and the last position where
2130 ;; `c-parse-state' or for which `c-invalidate-state-cache' was called, are
2131 ;; both in the same literal, there is no such "good position", and
2132 ;; c-state-cache-good-pos is then nil. This is the ONLY circumstance in which
2133 ;; it can be nil. In this case, `c-state-point-min-literal' will be non-nil.
2134 ;;
2135 ;; 2009-06-12: In a brace desert, c-state-cache-good-pos may also be in
2136 ;; the middle of the desert, as long as it is not within a brace pair
2137 ;; recorded in `c-state-cache' or a paren/bracket pair.
2140 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2141 ;; We maintain a simple cache of positions which aren't in a literal, so as to
2142 ;; speed up testing for non-literality.
2144 ;; The approximate interval between entries in `c-state-nonlit-pos-cache'.
2148 ;; A list of buffer positions which are known not to be in a literal or a cpp
2149 ;; construct. This is ordered with higher positions at the front of the list.
2150 ;; Only those which are less than `c-state-nonlit-pos-cache-limit' are valid.
2154 ;; An upper limit on valid entries in `c-state-nonlit-pos-cache'. This is
2155 ;; reduced by buffer changes, and increased by invocations of
2156 ;; `c-state-literal-at'.
2160 ;; A list of buffer positions which are known not to be in a literal. This is
2161 ;; ordered with higher positions at the front of the list. Only those which
2162 ;; are less than `c-state-semi-nonlit-pos-cache-limit' are valid.
2166 ;; An upper limit on valid entries in `c-state-semi-nonlit-pos-cache'. This is
2167 ;; reduced by buffer changes, and increased by invocations of
2168 ;; `c-state-literal-at'. FIXME!!!
2171 ;; Do a parse-partial-sexp from FROM to TO, returning either
2172 ;; (STATE TYPE (BEG . END)) if TO is in a literal; or
2173 ;; (STATE) otherwise,
2174 ;; where STATE is the parsing state at TO, TYPE is the type of the literal
2175 ;; (one of 'c, 'c++, 'string) and (BEG . END) is the boundaries of the literal.
2176 ;;
2177 ;; Only elements 3 (in a string), 4 (in a comment), 5 (following a quote),
2178 ;; 7 (comment type) and 8 (start of comment/string) (and possibly 9) of
2179 ;; STATE are valid.
2182 ty)
2189 nil ; TARGETDEPTH
2190 nil ; STOPBEFORE
2191 s ; OLDSTATE
2198 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2199 ;; string, comment, or macro.
2200 ;;
2201 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2202 ;; MAY NOT contain any positions within macros, since macros are frequently
2203 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2204 ;; We cannot rely on this mechanism whilst determining a cache pos since
2205 ;; this function is also called from outwith `c-parse-state'.
2210 pos npos high-pos lit macro-beg macro-end)
2211 ;; Trim the cache to take account of buffer changes.
2223 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2227 ;; Test for being in a literal. If so, go to after it.
2234 ;; Test for being in a macro. If so, go to after it.
2249 ;; Add one extra element above HERE so as to to avoid the previous
2250 ;; expensive calculation when the next call is close to the current
2251 ;; one. This is especially useful when inside a large macro.
2256 pos))))
2259 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2260 ;; string or comment. It may be in a macro.
2265 pos npos high-pos lit macro-beg macro-end)
2266 ;; Trim the cache to take account of buffer changes.
2278 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2282 ;; Test for being in a literal. If so, go to after it.
2295 pos))))
2298 ;; If position HERE is inside a literal, return (START . END), the
2299 ;; boundaries of the literal (which may be outside the accessible bit of the
2300 ;; buffer). Otherwise, return nil.
2301 ;;
2302 ;; This function is almost the same as `c-literal-limits'. Previously, it
2303 ;; differed in that it was a lower level function, and that it rigorously
2304 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2305 ;; virtually identical to this function.
2313 ;; Return the start of the literal containing POS, or POS itself.
2315 pos))
2318 ;; Return a position outside of a string/comment/macro at or before POS.
2319 ;; STATE is the parse-partial-sexp state at POS.
2323 pos)))
2328 res))))
2330 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2331 ;; Stuff to do with point-min, and coping with any literal there.
2334 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2335 ;; narrowing is likely to affect the parens that are visible before the point.
2341 ;; These two variables define the literal, if any, containing point-min.
2342 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2343 ;; literal. If there's no literal there, they're both nil.
2347 ;; This is the earliest buffer-pos from which scanning can be done. It is
2348 ;; either the end of the literal containing point-min, or point-min itself.
2349 ;; It becomes nil if the buffer is changed earlier than this point.
2351 ;; Return the lowest valid scanning pos. This will be the end of the
2352 ;; literal enclosing point-min, or point-min itself.
2364 ;; Determine the properties of any literal containing POINT-MIN, setting the
2365 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2366 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2368 lit)
2383 c-state-point-min-lit-start nil
2384 c-state-min-scan-pos p-min)))))
2387 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2388 ;; A variable which signals a brace dessert - helpful for reducing the number
2389 ;; of fruitless backward scans.
2392 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2393 ;; that defun has searched backwards for a brace pair and not found one. Its
2394 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2395 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2396 ;; nil when at top level) and FROM is where the backward search started. It
2397 ;; is reset to nil in `c-invalidate-state-cache'.
2400 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2401 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2402 ;; list of like structure.
2404 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2405 ;; (default `c-state-cache') (or nil).
2412 ;; Return the address of the latest brace/bracket/paren (whether left or
2413 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2420 ;; Return the position just after the latest brace/bracket/paren (whether
2421 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2429 ;; From the state-cache, determine the buffer position from which we might
2430 ;; scan forward to HERE to update this cache. This position will be just
2431 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2432 ;; return the earliest position in the accessible region which isn't within
2433 ;; a literal. If the visible portion of the buffer is entirely within a
2434 ;; literal, return NIL.
2436 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2452 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2453 ;; Variables which keep track of preprocessor constructs.
2458 ;; These are the limits of the macro containing point at the previous call of
2459 ;; `c-parse-state', or nil.
2461 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2462 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2464 ;; Return a start position for building `c-state-cache' from
2465 ;; scratch. This will be at the top level, 2 defuns back.
2467 ;; Go back 2 bods, but ignore any bogus positions returned by
2468 ;; beginning-of-defun (i.e. open paren in column zero).
2478 ;; Return the position of the opening paren/brace/bracket before HERE- which
2479 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2480 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2481 ;;
2482 ;; ............................................
2483 ;; | |
2484 ;; ( [ ( .........#macro.. ) ( ) ] )
2485 ;; ^ ^ ^ ^
2486 ;; | | | |
2487 ;; return HERE- HERE+ TOP
2488 ;;
2489 ;; If there aren't enough opening paren/brace/brackets, return the position
2490 ;; of the outermost one found, or HERE- if there are none. If there are no
2491 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2492 ;; must not be inside literals. Only the accessible portion of the buffer
2493 ;; will be scanned.
2495 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2496 ;; `here'. Go round the next loop each time we pass over such a ")". These
2497 ;; probably match "("s before `here-'.
2509 ;; PART 2: Scan back before `here-' searching for the "("s
2510 ;; matching/mismatching the ")"s found above. We only need to direct the
2511 ;; caller to scan when we've encountered unmatched right parens.
2520 pos))
2523 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2524 ;; to minimize the amount of scanning. HERE is the pertinent position in
2525 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2526 ;; its head trimmed) is known to be good, or nil if there is no such
2527 ;; position.
2528 ;;
2529 ;; The return value is a list, one of the following:
2530 ;;
2531 ;; o - ('forward CACHE-POS START-POINT) - scan forward from START-POINT,
2532 ;; which is not less than CACHE-POS.
2533 ;; o - ('backward CACHE-POS nil) - scan backwards (from HERE).
2534 ;; o - ('BOD nil START-POINT) - scan forwards from START-POINT, which is at the
2535 ;; top level.
2536 ;; o - ('IN-LIT nil nil) - point is inside the literal containing point-min.
2537 ;; , where CACHE-POS is the highest position recorded in `c-state-cache' at
2538 ;; or below HERE.
2540 BOD-pos ; position of 2nd BOD before HERE.
2541 strategy ; 'forward, 'backward, 'BOD, or 'IN-LIT.
2542 start-point
2548 start-point nil
2549 cache-pos nil
2561 start-point cache-pos)))
2563 ;; Might we be better off starting from the top level, two defuns back,
2564 ;; instead?
2569 start-point BOD-pos)))
2572 strategy
2577 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2578 ;; Routines which change `c-state-cache' and associated values.
2580 ;; The region (more precisely, point-min) has changed since we
2581 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2583 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2584 ;; It would be possible to do a better job here and recalculate the top
2585 ;; only.
2589 c-state-cache-good-pos c-state-min-scan-pos
2590 c-state-brace-pair-desert nil))
2592 ;; point-min has MOVED FORWARD.
2594 ;; Is the new point-min inside a (different) literal?
2599 ;; Cut off a bit of the tail from `c-state-cache'.
2601 pa)
2609 c-state-cache-good-pos c-state-min-scan-pos)
2612 )))
2617 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2618 ;; of nesting (not necessarily immediately preceding), push a cons onto
2619 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2620 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2621 ;; UPPER-LIM.
2622 ;;
2623 ;; Return non-nil when this has been done.
2624 ;;
2625 ;; The situation it copes with is this transformation:
2626 ;;
2627 ;; OLD: { (.) {...........}
2628 ;; ^ ^
2629 ;; FROM HERE
2630 ;;
2631 ;; NEW: { {....} (.) {.........
2632 ;; ^ ^ ^
2633 ;; LOWER BRACE PAIR HERE or HERE
2634 ;;
2635 ;; This routine should be fast. Since it can get called a LOT, we maintain
2636 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2637 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2641 new-cons
2649 ;; If we're essentially repeating a fruitless search, just give up.
2653 ;; DESERT-LIM. Only search what we absolutely need to,
2658 ;; CACHE-LIM. This limit will be necessary when an opening
2659 ;; paren at `cache-pos' has just had its matching close paren
2660 ;; inserted. `cache-pos' continues to be a search bound, even
2661 ;; though the algorithm below would skip over the new paren
2662 ;; pair.
2673 ;; In the next pair of nested loops, the inner one moves back past a
2674 ;; pair of (mis-)matching parens or brackets; the outer one moves
2675 ;; back over a sequence of unmatched close brace/paren/bracket each
2676 ;; time round.
2690 ; brace, ignore it.
2693 ;; We've found the desired brace-pair.
2705 ;; We haven't found a brace pair. Record this in the cache.
2709 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2710 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2711 ;; "push" "a" brace pair onto `c-state-cache'.
2712 ;;
2713 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2714 ;; otherwise push it normally.
2715 ;;
2716 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2717 ;; latter is inside a macro, not being a macro containing
2718 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2719 ;; base pair. This latter case is assumed to be rare.
2720 ;;
2721 ;; Note: POINT is not preserved in this routine.
2731 c-state-cache)))
2732 ;; N.B. This defsubst codes one method for the simple, normal case,
2733 ;; and a more sophisticated, slower way for the general case. Don't
2734 ;; eliminate this defsubst - it's a speed optimization.
2738 ;; Scan the buffer from FROM to (point-max), adding elements into
2739 ;; `c-state-cache' for braces etc. Return a candidate for
2740 ;; `c-state-cache-good-pos'.
2741 ;;
2742 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2743 ;; any. Typically, it is immediately after it. It must not be inside a
2744 ;; literal.
2753 ; Is actually the pos. to scan for a (/{/[ from,
2754 ; which sometimes is after a silly )/}/].
2757 ; good-pos.
2763 ;; Each time round the following loop, we enter a successively deeper
2764 ;; level of brace/paren nesting. (Except sometimes we "continue at
2765 ;; the existing level".) `pa+1' is a pos inside an opening
2766 ;; brace/paren/bracket, usually just after it.
2769 ;; Each time round the next loop moves forward over an opening then
2770 ;; a closing brace/bracket/paren. This loop is white hot, so it
2771 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2772 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2773 ;; call of `scan-lists' signals an error, which happens when there
2774 ;; are no more b/b/p's to scan.
2785 ;; We've just entered a deeper nesting level.
2787 ;; Insert the brace pair (if present) and the single open
2788 ;; paren/brace/bracket into `c-state-cache' It cannot be
2789 ;; inside a macro, except one around point, because of what
2790 ;; `c-neutralize-syntax-in-CPP' has done.
2792 ;; Insert the opening brace/bracket/paren position.
2794 ;; Clear admin stuff for the next more nested part of the scan.
2798 ;; All open p/b/b's at this nesting level, if any, have probably
2799 ;; been closed by matching/mismatching ones. We're probably
2800 ;; finished - we just need to check for having found an
2801 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2802 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2805 ;; Record the final, innermost, brace-pair if there is one.
2808 ;; Determine a good pos
2816 here-bol))
2825 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2826 ;; not be in it when it is amended for position (point-max).
2827 ;; Additionally, the "outermost" open-brace entry before (point-max)
2828 ;; will be converted to a cons if the matching close-brace is scanned.
2829 ;;
2830 ;; GOOD-POS is a "maximal" "safe position" - there must be no open
2831 ;; parens/braces/brackets between GOOD-POS and (point-max).
2832 ;;
2833 ;; As a second thing, calculate the result of parse-partial-sexp at
2834 ;; PPS-POINT, w.r.t. GOOD-POS. The motivation here is that
2835 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2836 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2837 ;; needs to be FAST).
2838 ;;
2839 ;; Return a list (GOOD-POS SCAN-BACK-POS PPS-STATE), where
2840 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2841 ;; to be good (we aim for this to be as high as possible);
2842 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2843 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2844 ;; position to scan backwards from.
2845 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2855 ; or nil
2862 ; (point-max)), or nil.
2868 pps-state ; Will be 9 or 10 elements long.
2869 pos
2870 upper-lim ; ,beyond which `c-state-cache' entries are removed
2871 scan-back-pos
2872 pair-beg pps-point-state target-depth)
2874 ;; Remove entries beyond (point-max). Also remove any entries inside
2875 ;; a macro, unless (point-max) is in the same macro.
2884 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2885 ;; RBrace and indicate we'll need to search backwards for a previous
2886 ;; brace pair.
2893 ;; The next loop jumps forward out of a nested level of parens each
2894 ;; time round; the corresponding elements in `c-state-cache' are
2895 ;; removed. `pos' is just after the brace-pair or the open paren at
2896 ;; (car c-state-cache). There can be no open parens/braces/brackets
2897 ;; between `good-pos'/`good-pos-actual-macro-start' and (point-max),
2898 ;; due to the interface spec to this function.
2901 in-macro-start)))
2903 ; marked by a `category' text property.
2904 good-pos))
2913 ;; Do nothing when we've merely reached pps-point.
2914 )
2916 ;; Scan!
2920 target-depth
2921 nil pps-state))
2931 ;; We've just moved out of the paren pair containing the brace-pair
2932 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2933 ;; and is potentially where the open brace of a cons in
2934 ;; c-state-cache will be.
2940 ; containing Lparen
2944 ; together with enclosed brace pair.
2945 ;; (t nil) ; Ignore an unmated Rparen.
2946 )))
2950 nil nil ; TARGETDEPTH, STOPBEFORE
2951 pps-state)))
2953 ;; If the last paren pair we moved out of was actually a brace pair,
2954 ;; insert it into `c-state-cache'.
2959 c-state-cache)))
2964 ;; Strip stale elements of `c-state-cache' by moving backwards through the
2965 ;; buffer, and inform the caller of the scenario detected.
2966 ;;
2967 ;; HERE is the position we're setting `c-state-cache' for.
2968 ;; CACHE-POS is just after the latest recorded position in `c-state-cache'
2969 ;; before HERE, or a position at or near point-min which isn't in a
2970 ;; literal.
2971 ;;
2972 ;; This function must only be called only when (> `c-state-cache-good-pos'
2973 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
2974 ;; optimized to eliminate (or minimize) scanning between these two
2975 ;; positions.
2976 ;;
2977 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
2978 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
2979 ;; could become so after missing elements are inserted into
2980 ;; `c-state-cache'. This is JUST AFTER an opening or closing
2981 ;; brace/paren/bracket which is already in `c-state-cache' or just before
2982 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
2983 ;; before `here''s line, or the start of the literal containing it.
2984 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
2985 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
2986 ;; to scan backwards from.
2987 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
2988 ;; POS and HERE which aren't recorded in `c-state-cache'.
2989 ;;
2990 ;; The comments in this defun use "paren" to mean parenthesis or square
2991 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
2992 ;;
2993 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
2994 ;; | | | | | |
2995 ;; CP E here D C good
2997 pa ren ; positions of "(" and ")"
2998 dropped-cons ; whether the last element dropped from `c-state-cache'
2999 ; was a cons (representing a brace-pair)
3000 good-pos ; see above.
3001 lit ; (START . END) of a literal containing some point.
3002 here-lit-start here-lit-end ; bounds of literal containing `here'
3003 ; or `here' itself.
3008 ;; Remove completely irrelevant entries from `c-state-cache'.
3014 ;; At this stage, (> pos here);
3015 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3020 ;; CASE 1: The top of the cache is a brace pair which now encloses
3021 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3022 ;; knowledge of what's inside these braces, we have no alternative but
3023 ;; to direct the caller to scan the buffer from the opening brace.
3027 ; entry into a { entry, so the caller needs to
3028 ; search for a brace pair before the {.
3030 ;; `here' might be inside a literal. Check for this.
3035 ;; Has `here' just "newly entered" a macro?
3048 ;; `here' might be nested inside any depth of parens (or brackets but
3049 ;; not braces). Scan backwards to find the outermost such opening
3050 ;; paren, if there is one. This will be the scan position to return.
3057 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3058 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3059 ;; a brace pair preceding this, it will already be in `c-state-cache',
3060 ;; unless there was a brace pair after it, i.e. there'll only be one to
3061 ;; scan for if we've just deleted one.
3064 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3065 ;; Further forward scanning isn't needed, but we still need to find a
3066 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3076 ;; CASE 3: After a }/)/] before `here''s BOL.
3080 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3081 ;; literal containing it.
3086 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3087 ;; Externally visible routines.
3091 c-state-cache-good-pos 1
3092 c-state-nonlit-pos-cache nil
3093 c-state-nonlit-pos-cache-limit 1
3094 c-state-brace-pair-desert nil
3095 c-state-point-min 1
3096 c-state-point-min-lit-type nil
3097 c-state-point-min-lit-start nil
3098 c-state-min-scan-pos 1
3099 c-state-old-cpp-beg nil
3100 c-state-old-cpp-end nil)
3103 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3104 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3105 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3106 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3107 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3108 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3109 ;; (defun c-state-dump ()
3110 ;; ;; For debugging.
3111 ;; ;(message
3112 ;; (concat
3113 ;; (c-sc-qde c-state-cache)
3114 ;; (c-sc-de c-state-cache-good-pos)
3115 ;; (c-sc-qde c-state-nonlit-pos-cache)
3116 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3117 ;; (c-sc-qde c-state-brace-pair-desert)
3118 ;; (c-sc-de c-state-point-min)
3119 ;; (c-sc-de c-state-point-min-lit-type)
3120 ;; (c-sc-de c-state-point-min-lit-start)
3121 ;; (c-sc-de c-state-min-scan-pos)
3122 ;; (c-sc-de c-state-old-cpp-beg)
3123 ;; (c-sc-de c-state-old-cpp-end)))
3124 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3127 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3128 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3129 ;; left in a consistent state.
3130 ;;
3131 ;; This is much like `c-whack-state-after', but it never changes a paren
3132 ;; pair element into an open paren element. Doing that would mean that the
3133 ;; new open paren wouldn't have the required preceding paren pair element.
3134 ;;
3135 ;; This function is called from c-after-change.
3137 ;; The caches of non-literals:
3143 ;; `c-state-cache':
3144 ;; Case 1: if `here' is in a literal containing point-min, everything
3145 ;; becomes (or is already) nil.
3149 c-state-cache-good-pos nil
3150 c-state-min-scan-pos nil)
3152 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3153 ;; below `here'. To maintain its consistency, we may need to insert a new
3154 ;; brace pair.
3156 too-high-pa ; recorded {/(/[ next above here, or nil.
3157 dropped-cons ; was the last removed element a brace pair?
3158 pa)
3159 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3166 ;; Do we need to add in an earlier brace pair, having lopped one off?
3173 ;; The brace-pair desert marker:
3181 ;; Find and record all noteworthy parens between some good point earlier in
3182 ;; the file and point. That good point is at least the beginning of the
3183 ;; top-level construct we are in, or the beginning of the preceding
3184 ;; top-level construct if we aren't in one.
3185 ;;
3186 ;; The returned value is a list of the noteworthy parens with the last one
3187 ;; first. If an element in the list is an integer, it's the position of an
3188 ;; open paren (of any type) which has not been closed before the point. If
3189 ;; an element is a cons, it gives the position of a closed BRACE paren
3190 ;; pair[*]; the car is the start brace position and the cdr is the position
3191 ;; following the closing brace. Only the last closed brace paren pair
3192 ;; before each open paren and before the point is recorded, and thus the
3193 ;; state never contains two cons elements in succession. When a close brace
3194 ;; has no matching open brace (e.g., the matching brace is outside the
3195 ;; visible region), it is not represented in the returned value.
3196 ;;
3197 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3198 ;; This defun explicitly treats mismatching parens/braces/brackets as
3199 ;; matching. It is the open brace which makes it a "brace" pair.
3200 ;;
3201 ;; If POINT is within a macro, open parens and brace pairs within
3202 ;; THIS macro MIGHT be recorded. This depends on whether their
3203 ;; syntactic properties have been suppressed by
3204 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3205 ;;
3206 ;; Currently no characters which are given paren syntax with the
3207 ;; syntax-table property are recorded, i.e. angle bracket arglist
3208 ;; parens are never present here. Note that this might change.
3209 ;;
3210 ;; BUG: This function doesn't cope entirely well with unbalanced
3211 ;; parens in macros. (2008-12-11: this has probably been resolved
3212 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3213 ;; following case the brace before the macro isn't balanced with the
3214 ;; one after it:
3215 ;;
3216 ;; {
3217 ;; #define X {
3218 ;; }
3219 ;;
3220 ;; Note to maintainers: this function DOES get called with point
3221 ;; within comments and strings, so don't assume it doesn't!
3222 ;;
3223 ;; This function might do hidden buffer changes.
3226 strategy ; 'forward, 'backward etc..
3227 ;; Candidate positions to start scanning from:
3228 cache-pos ; highest position below HERE already existing in
3229 ; cache (or 1).
3230 good-pos
3231 start-point
3232 bopl-state
3233 res
3235 ;; If POINT-MIN has changed, adjust the cache
3239 ;; Strategy?
3247 c-state-cache-good-pos start-point))
3249 ;; SCAN!
3258 ; start-point)
3267 good-pos)))
3276 scan-backward-pos))
3281 good-pos)))
3285 c-state-cache-good-pos nil)))))
3287 c-state-cache)
3290 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3291 ;;
3292 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3293 ;; of all parens in preprocessor constructs, except for any such construct
3294 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3295 ;; worrying further about macros and template delimiters.
3300 c-state-old-cpp-beg
3307 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3308 ;; description of the functionality and return value.
3309 ;;
3310 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3311 ;; of all parens in preprocessor constructs, except for any such construct
3312 ;; containing point. We can then call `c-parse-state-1' without worrying
3313 ;; further about macros and template delimiters.
3320 here-cpp-beg)
3322 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3323 ;; subsystem.
3328 here-cpp-beg here-cpp-end
3334 )))
3336 ;; Debug tool to catch cache inconsistencies. This is called from
3337 ;; 000tests.el.
3350 c-state-cache-good-pos
3351 c-state-nonlit-pos-cache
3352 c-state-nonlit-pos-cache-limit
3353 c-state-brace-pair-desert
3354 c-state-point-min
3355 c-state-point-min-lit-type
3356 c-state-point-min-lit-start
3357 c-state-min-scan-pos
3358 c-state-old-cpp-beg
3359 c-state-old-cpp-end))))
3384 ;; The cache can actually go further back due to the ad-hoc way
3385 ;; the first paren is found, so try to whack off a bit of its
3386 ;; start before complaining.
3387 ;; (save-excursion
3388 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3389 ;; (c-beginning-of-defun-1)
3390 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3391 ;; (c-beginning-of-defun-1))
3392 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3393 ;; (message (concat "c-parse-state inconsistency at %s: "
3394 ;; "using cache: %s, from scratch: %s")
3395 ;; here res1 res2)))
3398 here res1 res2)
3402 res1))
3408 'c-debug-parse-state
3414 \f
3416 ;; Whack off any state information from PAREN-STATE which lies
3417 ;; before BUFPOS. Not destructive on PAREN-STATE.
3420 car)
3431 ;; Whack off any state information from PAREN-STATE which lies at or
3432 ;; after BUFPOS. Not destructive on PAREN-STATE.
3437 ;; just check the car, because in a balanced brace
3438 ;; expression, it must be impossible for the corresponding
3439 ;; close brace to be before point, but the open brace to
3440 ;; be after.
3442 nil ; whack it off
3444 ;; its possible that the open brace is before
3445 ;; bufpos, but the close brace is after. In that
3446 ;; case, convert this to a non-cons element. The
3447 ;; rest of the state is before bufpos, so we're
3448 ;; done.
3450 ;; we know that both the open and close braces are
3451 ;; before bufpos, so we also know that everything else
3452 ;; on state is before bufpos.
3455 nil ; whack it off
3456 ;; it's before bufpos, so everything else should too.
3459 nil)))
3462 ;; Return the bufpos of the innermost enclosing open paren before
3463 ;; bufpos, or nil if none was found.
3473 enclosingp))
3476 ;; Return the bufpos of the outermost enclosing open paren, or nil
3477 ;; if none was found.
3484 pos))
3487 ;; Return the closest "safe" position recorded on PAREN-STATE that
3488 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3489 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3490 ;; find the closest limit before a given limit that might be nil.
3491 ;;
3492 ;; A "safe" position is a position at or after a recorded open
3493 ;; paren, or after a recorded close paren. The returned position is
3494 ;; thus either the first position after a close brace, or the first
3495 ;; position after an enclosing paren, or at the enclosing paren in
3496 ;; case BUFPOS is immediately after it.
3506 ;; See below.
3509 ;; elem is the position at and not after the opening paren, so
3510 ;; we can go forward one more step unless it's equal to
3511 ;; bufpos. This is useful in some cases avoid an extra paren
3512 ;; level between the safe position and bufpos.
3517 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3518 ;; goes to the closest previous point that is known to be outside
3519 ;; any string literal or comment. `c-state-cache' is used if it has
3520 ;; a position in the vicinity.
3522 elem
3525 ;; Note: Similar code in `c-safe-position'. The
3526 ;; difference is that we accept a safe position at
3527 ;; the point and don't bother to go forward past open
3528 ;; parens.
3543 ;; The position is far back. Try `c-beginning-of-defun-1'
3544 ;; (although we can't be entirely sure it will go to a position
3545 ;; outside a comment or string in current emacsen). FIXME:
3546 ;; Consult `syntax-ppss' here.
3551 \f
3552 ;; Tools for scanning identifiers and other tokens.
3555 "Return non-nil if the point is on or directly after an identifier.
3556 Keywords are recognized and not considered identifiers. If an
3557 identifier is detected, the returned value is its starting position.
3558 If an identifier ends at the point and another begins at it \(can only
3559 happen in Pike) then the point for the preceding one is returned.
3561 Note that this function might do hidden buffer changes. See the
3562 comment at the start of cc-engine.el for more info."
3564 ;; FIXME: Shouldn't this function handle "operator" in C++?
3571 ;; Check for a normal (non-keyword) identifier.
3577 ;; Handle the `<operator> syntax in Pike.
3581 t
3588 ;; Handle the "operator +" syntax in C++.
3605 )))
3608 ;; If the point is at the end of a symbol then skip backward to the
3609 ;; beginning of it. Don't move otherwise. Return non-nil if point
3610 ;; moved.
3611 ;;
3612 ;; This function might do hidden buffer changes.
3615 ;; Handle the `<operator> syntax in Pike.
3621 t
3623 nil)))))
3626 ;; Move to the beginning of the current token. Do not move if not
3627 ;; in the middle of one. BACK-LIMIT may be used to bound the
3628 ;; backward search; if given it's assumed to be at the boundary
3629 ;; between two tokens. Return non-nil if the point is moved, nil
3630 ;; otherwise.
3631 ;;
3632 ;; This function might do hidden buffer changes.
3639 ;; `c-nonsymbol-token-regexp' should always match
3640 ;; since we've skipped backward over punctuator
3641 ;; or paren syntax, but consume one char in case
3642 ;; it doesn't so that we don't leave point before
3643 ;; some earlier incorrect token.
3650 ;; Move to the end of the current token. Do not move if not in the
3651 ;; middle of one. BACK-LIMIT may be used to bound the backward
3652 ;; search; if given it's assumed to be at the boundary between two
3653 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3654 ;;
3655 ;; This function might do hidden buffer changes.
3663 ;; `c-nonsymbol-token-regexp' should always match since
3664 ;; we've skipped backward over punctuator or paren
3665 ;; syntax, but move forward in case it doesn't so that
3666 ;; we don't leave point earlier than we started with.
3682 "Move forward by tokens.
3683 A token is defined as all symbols and identifiers which aren't
3685 treated properly). Point is always either left at the beginning of a
3686 token or not moved at all. COUNT specifies the number of tokens to
3687 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3688 moves to the next token beginning only if not already at one. If
3689 BALANCED is true, move over balanced parens, otherwise move into them.
3690 Also, if BALANCED is true, never move out of an enclosing paren.
3692 LIMIT sets the limit for the movement and defaults to the point limit.
3693 The case when LIMIT is set in the middle of a token, comment or macro
3694 is handled correctly, i.e. the point won't be left there.
3696 Return the number of tokens left to move \(positive or negative). If
3697 BALANCED is true, a move over a balanced paren counts as one. Note
3698 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3699 be returned. Thus, a return value of 0 guarantees that point is at
3700 the requested position and a return value less \(without signs) than
3701 COUNT guarantees that point is at the beginning of some token.
3703 Note that this function might do hidden buffer changes. See the
3704 comment at the start of cc-engine.el for more info."
3711 c-jump-syntax-balanced
3712 c-jump-syntax-unbalanced))
3717 ;; If count is zero we should jump if in the middle of a token.
3724 ;; Skip whitespace. Count this as a move if we did in
3725 ;; fact move.
3729 ;; Moved out of bounds. Make sure the returned count isn't zero.
3734 ;; Use `condition-case' to avoid having the limit tests
3735 ;; inside the loop.
3743 t)
3746 t)
3747 ;; `c-nonsymbol-token-regexp' above should always
3748 ;; match if there are correct tokens. Try to
3749 ;; widen to see if the limit was set in the
3750 ;; middle of one, else fall back to treating
3751 ;; the offending thing as a one character token.
3756 nil)
3759 t))))
3769 count)))
3772 "Move backward by tokens.
3773 See `c-forward-token-2' for details."
3781 c-jump-syntax-balanced
3782 c-jump-syntax-unbalanced))
3786 ;; The count is zero so try to skip to the beginning of the
3787 ;; current token.
3791 ;; The limit is inside the same token, so return 1.
3794 ;; We're not in the middle of a token. If there's
3795 ;; whitespace after the point then we must move backward,
3796 ;; so set count to 1 in that case.
3798 ;; If we're looking at a '#' that might start a cpp
3799 ;; directive then we have to do a more elaborate check.
3812 ;; Use `condition-case' to avoid having to check for buffer
3813 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3822 ;; This can be very inefficient if there's a long
3823 ;; sequence of operator tokens without any separation.
3824 ;; That doesn't happen in practice, anyway.
3834 count)))
3837 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3839 characters are jumped over character by character. This function is
3840 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3845 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3847 characters are jumped over character by character. This function is
3848 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3852 \f
3853 ;; Tools for doing searches restricted to syntactically relevant text.
3856 paren-level not-inside-token
3857 lookbehind-submatch)
3858 "Like `re-search-forward', but only report matches that are found
3859 in syntactically significant text. I.e. matches in comments, macros
3860 or string literals are ignored. The start point is assumed to be
3861 outside any comment, macro or string literal, or else the content of
3862 that region is taken as syntactically significant text.
3864 If PAREN-LEVEL is non-nil, an additional restriction is added to
3865 ignore matches in nested paren sexps. The search will also not go
3866 outside the current list sexp, which has the effect that if the point
3867 should be moved to BOUND when no match is found \(i.e. NOERROR is
3868 neither nil nor t), then it will be at the closing paren if the end of
3869 the current list sexp is encountered first.
3871 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3872 ignored. Things like multicharacter operators and special symbols
3874 constants.
3876 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3877 subexpression in REGEXP. The end of that submatch is used as the
3878 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
3879 isn't used or if that subexpression didn't match then the start
3881 subexpression is never tested before the starting position, so it
3885 subexpression can match the end of nonwhite syntactic whitespace,
3886 i.e. the end of comments or cpp directives. This since the function
3887 skips over such things before resuming the search. It's on the other
3889 matches syntactic whitespace.
3891 Bug: Unbalanced parens inside cpp directives are currently not handled
3892 correctly \(i.e. they don't get ignored as they should) when
3893 PAREN-LEVEL is set.
3895 Note that this function might do hidden buffer changes. See the
3896 comment at the start of cc-engine.el for more info."
3901 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
3904 tmp
3905 ;; Start position for the last search.
3906 search-pos
3907 ;; The `parse-partial-sexp' state between the start position
3908 ;; and the point.
3909 state
3910 ;; The current position after the last state update. The next
3911 ;; `parse-partial-sexp' continues from here.
3913 ;; The position at which to check the state and the state
3914 ;; there. This is separate from `state-pos' since we might
3915 ;; need to back up before doing the next search round.
3916 check-pos check-state
3917 ;; Last position known to end a token.
3919 ;; Set when a valid match is found.
3920 found)
3938 state-pos check-pos paren-level nil state))
3940 check-state state))
3942 ;; NOTE: If we got a look behind subexpression and get
3943 ;; an insignificant match in something that isn't
3944 ;; syntactic whitespace (i.e. strings or in nested
3945 ;; parentheses), then we can never skip more than a
3946 ;; single character from the match start position
3947 ;; (i.e. `state-pos' here) before continuing the
3948 ;; search. That since the look behind subexpression
3949 ;; might match the end of the insignificant region in
3950 ;; the next search.
3954 ;; Match inside a line comment. Skip to eol. Use
3955 ;; `re-search-forward' instead of `skip-chars-forward' to get
3956 ;; the right bound behavior.
3960 ;; Match inside a block comment. Skip to the '*/'.
3967 ;; Match in the middle of the opener of a block or line
3968 ;; comment.
3973 ;; The last `parse-partial-sexp' above might have
3974 ;; stopped short of the real check position if the end
3975 ;; of the current sexp was encountered in paren-level
3976 ;; mode. The checks above are always false in that
3977 ;; case, and since they can do better skipping in
3978 ;; lookbehind-submatch mode, we do them before
3979 ;; checking the paren level.
3983 ;; Check the paren level first since we're short of the
3984 ;; syntactic checking position if the end of the
3985 ;; current sexp was encountered by `parse-partial-sexp'.
3988 ;; Inside a nested paren sexp.
3990 ;; See the NOTE above.
3992 ;; Skip out of the paren quickly.
3996 ;; Have exited the current paren sexp.
3999 ;; The last `parse-partial-sexp' call above
4000 ;; has left us just after the closing paren
4001 ;; in this case, so we can modify the bound
4002 ;; to leave the point at the right position
4003 ;; upon return.
4005 nil)
4009 ;; Match inside a string.
4012 ;; See the NOTE above.
4014 ;; Skip to the end of the string before continuing.
4023 continue)))
4028 ;; Match inside a macro. Skip to the end of it.
4036 (< check-pos
4042 ;; Inside a token.
4044 ;; See the NOTE above.
4049 ;; A real match.
4051 nil)))
4053 ;; Should loop to search again, but take care to avoid
4054 ;; looping on the same spot.
4058 nil
4061 t))))
4067 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4074 ;; Search failed. Set point as appropriate.
4078 nil)))
4083 ;; Return the start of the literal point is in, or nil.
4084 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4085 ;; bound in the caller.
4087 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4088 ;; if it's outside comments and strings.
4091 ;; Pick a safe position as close to the point as possible.
4092 ;;
4093 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4094 ;; position.
4107 ;; Cache positions along the way to use if we have to back up more. We
4108 ;; cache every closing paren on the same level. If the paren cache is
4109 ;; relevant in this region then we're typically already on the same
4110 ;; level as the target position. Note that we might cache positions
4111 ;; after opening parens in case safe-pos is in a nested list. That's
4112 ;; both uncommon and harmless.
4120 ;; If the state contains the start of the containing sexp we cache that
4121 ;; position too, so that parse-partial-sexp in the next run has a bigger
4122 ;; chance of starting at the same level as the target position and thus
4123 ;; will get more good safe positions into the list.
4129 ;; Inside string or comment. Continue search at the
4130 ;; beginning of it.
4134 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4135 i.e. don't stop at positions inside syntactic whitespace or string
4136 literals. Preprocessor directives are also ignored, with the exception
4137 of the one that the point starts within, if any. If LIMIT is given,
4138 it's assumed to be at a syntactically relevant position.
4140 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4141 sexps, and the search will also not go outside the current paren sexp.
4142 However, if LIMIT or the buffer limit is reached inside a nested paren
4143 then the point will be left at the limit.
4145 Non-nil is returned if the point moved, nil otherwise.
4147 Note that this function might do hidden buffer changes. See the
4148 comment at the start of cc-engine.el for more info."
4151 state-2
4152 ;; A list of syntactically relevant positions in descending
4153 ;; order. It's used to avoid scanning repeatedly over
4154 ;; potentially large regions with `parse-partial-sexp' to verify
4155 ;; each position. Used in `c-ssb-lit-begin'
4156 safe-pos-list
4157 ;; The result from `c-beginning-of-macro' at the start position or the
4158 ;; start position itself if it isn't within a macro. Evaluated on
4159 ;; demand.
4160 start-macro-beg
4161 ;; The earliest position after the current one with the same paren
4162 ;; level. Used only when `paren-level' is set.
4163 lit-beg
4168 ;; The next loop "tries" to find the end point each time round,
4169 ;; loops when it hasn't succeeded.
4177 ;; Don't stop inside a literal
4180 t)
4188 ;; Not at the right level.
4191 ;; We stop above if we go out of a paren.
4192 ;; Now check whether it precedes or is
4193 ;; nested in the starting sexp.
4197 pps-end-pos paren-level-pos
4198 nil nil state-2))
4201 ;; We've stopped short of the starting position
4202 ;; so the hit was inside a nested list. Go up
4203 ;; until we are at the right level.
4212 nil)
4213 t))
4216 nil))
4218 ;; The hit was outside the list at the start
4219 ;; position. Go to the start of the list and exit.
4221 nil))
4224 ;; Inside a macro.
4232 t
4234 ;; It's inside the same macro we started in so it's
4235 ;; a relevant match.
4237 nil))))))
4241 ;; Skip syntactic ws afterwards so that we don't stop at the
4242 ;; end of a comment if `skip-chars' is something like "^/".
4246 ;; We might want to extend this with more useful return values in
4247 ;; the future.
4250 ;; The following is an alternative implementation of
4251 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4252 ;; track of the syntactic context. It turned out to be generally
4253 ;; slower than the one above which uses forward checks from earlier
4254 ;; safe positions.
4255 ;;
4256 ;;(defconst c-ssb-stop-re
4257 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4258 ;; ;; stop at to avoid going into comments and literals.
4259 ;; (concat
4260 ;; ;; Match comment end syntax and string literal syntax. Also match
4261 ;; ;; '/' for block comment endings (not covered by comment end
4262 ;; ;; syntax).
4263 ;; "\\s>\\|/\\|\\s\""
4264 ;; (if (memq 'gen-string-delim c-emacs-features)
4265 ;; "\\|\\s|"
4266 ;; "")
4267 ;; (if (memq 'gen-comment-delim c-emacs-features)
4268 ;; "\\|\\s!"
4269 ;; "")))
4270 ;;
4271 ;;(defconst c-ssb-stop-paren-re
4272 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4273 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4274 ;;
4275 ;;(defconst c-ssb-sexp-end-re
4276 ;; ;; Regexp matching the ending syntax of a complex sexp.
4277 ;; (concat c-string-limit-regexp "\\|\\s)"))
4278 ;;
4279 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4280 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4281 ;;i.e. don't stop at positions inside syntactic whitespace or string
4282 ;;literals. Preprocessor directives are also ignored. However, if the
4283 ;;point is within a comment, string literal or preprocessor directory to
4284 ;;begin with, its contents is treated as syntactically relevant chars.
4285 ;;If LIMIT is given, it limits the backward search and the point will be
4286 ;;left there if no earlier position is found.
4287 ;;
4288 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4289 ;;sexps, and the search will also not go outside the current paren sexp.
4290 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4291 ;;then the point will be left at the limit.
4292 ;;
4293 ;;Non-nil is returned if the point moved, nil otherwise.
4294 ;;
4295 ;;Note that this function might do hidden buffer changes. See the
4296 ;;comment at the start of cc-engine.el for more info."
4297 ;;
4298 ;; (save-restriction
4299 ;; (when limit
4300 ;; (narrow-to-region limit (point-max)))
4301 ;;
4302 ;; (let ((start (point)))
4303 ;; (catch 'done
4304 ;; (while (let ((last-pos (point))
4305 ;; (stop-pos (progn
4306 ;; (skip-chars-backward skip-chars)
4307 ;; (point))))
4308 ;;
4309 ;; ;; Skip back over the same region as
4310 ;; ;; `skip-chars-backward' above, but keep to
4311 ;; ;; syntactically relevant positions.
4312 ;; (goto-char last-pos)
4313 ;; (while (and
4314 ;; ;; `re-search-backward' with a single char regexp
4315 ;; ;; should be fast.
4316 ;; (re-search-backward
4317 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4318 ;; stop-pos 'move)
4319 ;;
4320 ;; (progn
4321 ;; (cond
4322 ;; ((looking-at "\\s(")
4323 ;; ;; `paren-level' is set and we've found the
4324 ;; ;; start of the containing paren.
4325 ;; (forward-char)
4326 ;; (throw 'done t))
4327 ;;
4328 ;; ((looking-at c-ssb-sexp-end-re)
4329 ;; ;; We're at the end of a string literal or paren
4330 ;; ;; sexp (if `paren-level' is set).
4331 ;; (forward-char)
4332 ;; (condition-case nil
4333 ;; (c-backward-sexp)
4334 ;; (error
4335 ;; (goto-char limit)
4336 ;; (throw 'done t))))
4337 ;;
4338 ;; (t
4339 ;; (forward-char)
4340 ;; ;; At the end of some syntactic ws or possibly
4341 ;; ;; after a plain '/' operator.
4342 ;; (let ((pos (point)))
4343 ;; (c-backward-syntactic-ws)
4344 ;; (if (= pos (point))
4345 ;; ;; Was a plain '/' operator. Go past it.
4346 ;; (backward-char)))))
4347 ;;
4348 ;; (> (point) stop-pos))))
4349 ;;
4350 ;; ;; Now the point is either at `stop-pos' or at some
4351 ;; ;; position further back if `stop-pos' was at a
4352 ;; ;; syntactically irrelevant place.
4353 ;;
4354 ;; ;; Skip additional syntactic ws so that we don't stop
4355 ;; ;; at the end of a comment if `skip-chars' is
4356 ;; ;; something like "^/".
4357 ;; (c-backward-syntactic-ws)
4358 ;;
4359 ;; (< (point) stop-pos))))
4360 ;;
4361 ;; ;; We might want to extend this with more useful return values
4362 ;; ;; in the future.
4363 ;; (/= (point) start))))
4365 \f
4366 ;; Tools for handling comments and string literals.
4369 "Return the type of literal point is in, if any.
4370 The return value is `c' if in a C-style comment, `c++' if in a C++
4371 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4372 is non-nil and in a preprocessor line, or nil if somewhere else.
4373 Optional LIM is used as the backward limit of the search. If omitted,
4374 or nil, `c-beginning-of-defun' is used.
4376 The last point calculated is cached if the cache is enabled, i.e. if
4377 `c-in-literal-cache' is bound to a two element vector.
4379 Note that this function might do hidden buffer changes. See the
4380 comment at the start of cc-engine.el for more info."
4391 "Return a cons of the beginning and end positions of the comment or
4392 string surrounding point (including both delimiters), or nil if point
4394 to start parsing from. If NEAR is non-nil, then the limits of any
4396 spaces and tabs between point and the literal. The search for such a
4397 literal is done first in forward direction. If NOT-IN-DELIMITER is
4398 non-nil, the case when point is inside a starting delimiter won't be
4399 recognized. This only has effect for comments which have starting
4400 delimiters with more than one character.
4402 Note that this function might do hidden buffer changes. See the
4403 comment at the start of cc-engine.el for more info."
4420 ;; We're standing in a comment starter.
4426 ;; Search forward for a literal.
4437 ;; Search backward.
4448 ;; Block comment. Due to the nature of line
4449 ;; comments, they will always be covered by the
4450 ;; normal case above.
4453 ;; If LIM is bogus, beg will be bogus.
4457 ))))
4459 ;; In case external callers use this; it did have a docstring.
4463 "If the argument is a cons of two buffer positions (such as returned by
4464 `c-literal-limits'), and that range contains a C++ style line comment,
4465 then an extended range is returned that contains all adjacent line
4466 comments (i.e. all comments that starts in the same column with no
4467 empty lines or non-whitespace characters between them). Otherwise the
4468 argument is returned.
4470 Note that this function might do hidden buffer changes. See the
4471 comment at the start of cc-engine.el for more info."
4482 ;; Got to take care in the backward direction to handle
4483 ;; comments which are preceded by code.
4497 range)
4501 "Convenience function that given the result of `c-literal-limits',
4502 returns nil or the type of literal that the range surrounds, one
4503 of the symbols 'c, 'c++ or 'string. It's much faster than using
4504 `c-in-literal' and is intended to be used when you need both the
4505 type of a literal and its limits.
4507 Note that this function might do hidden buffer changes. See the
4508 comment at the start of cc-engine.el for more info."
4519 range))
4522 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4523 ;; This doesn't preserve point.
4532 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4533 ;; (default point). This is done by going back further in the buffer then
4534 ;; searching forward for literals. The position found won't be in a
4535 ;; literal. We start searching for the sought position TRY-SIZE (default
4536 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4537 ;; :-)
4545 stack elt size
4548 ;; Move forward one literal each time round this loop.
4549 ;; Move forward to the start of a comment or string.
4551 pos
4552 start
4553 nil ; target-depth
4554 nil ; stop-before
4555 s ; state
4558 ;; Gather details of the non-literal-bit - starting pos and size.
4562 pos))
4566 ;; Move forward to the end of the comment/string.
4570 start
4571 nil ; target-depth
4572 nil ; stop-before
4573 s ; state
4577 ;; Now try and find enough non-literal characters recorded on the stack.
4578 ;; Go back one recorded literal each time round this loop.
4580 stack)
4585 ;; Have we found enough yet?
4595 ;; Return a buffer position about HOW-FAR non-literal characters forward
4596 ;; from START-POS (default point), which must not be inside a literal.
4603 ;; Scan over counted characters.
4605 pos
4607 nil ; target-depth
4608 nil ; stop-before
4609 s ; state
4613 ;; Scan over literal characters.
4616 pos
4618 nil ; target-depth
4619 nil ; stop-before
4620 s ; state
4625 \f
4626 ;; `c-find-decl-spots' and accompanying stuff.
4628 ;; Variables used in `c-find-decl-spots' to cache the search done for
4629 ;; the first declaration in the last call. When that function starts,
4630 ;; it needs to back up over syntactic whitespace to look at the last
4631 ;; token before the region being searched. That can sometimes cause
4632 ;; moves back and forth over a quite large region of comments and
4633 ;; macros, which would be repeated for each changed character when
4634 ;; we're called during fontification, since font-lock refontifies the
4635 ;; current line for each change. Thus it's worthwhile to cache the
4636 ;; first match.
4637 ;;
4638 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4639 ;; the syntactic whitespace less or equal to some start position.
4640 ;; There's no cached value if it's nil.
4641 ;;
4642 ;; `c-find-decl-match-pos' is the match position if
4643 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4644 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4655 ; (defface c-debug-decl-spot-face
4656 ; '((t (:background "Turquoise")))
4657 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4658 ; (defface c-debug-decl-sws-face
4659 ; '((t (:background "Khaki")))
4660 ; "Debug face to mark the syntactic whitespace between the declaration
4661 ; spots and the preceding token end.")
4677 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4678 ;; but it contains lots of free variables that refer to things
4679 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4680 ;; if there is a match, otherwise at `cfd-limit'.
4681 ;;
4682 ;; This macro might do hidden buffer changes.
4685 ;; Find the next property match position if we haven't got one already.
4696 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4697 ;; got one already.
4707 ;; Match. Check if it's inside a comment or string literal.
4710 ;; Matched the end of a token preceding a decl spot.
4714 ;; Matched a token that start a decl spot.
4717 c-literal-faces)
4719 ;; No match. Finish up and exit the loop.
4721 nil)
4723 ;; Skip out of comments and string literals.
4730 ;; If we matched at the decl start, we have to back up over the
4731 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4732 ;; any decl spots in the syntactic ws.
4737 ;; Choose whichever match is closer to the start.
4740 cfd-re-match nil)
4742 cfd-prop-match nil))
4747 ;; Skip forward past comments only so we don't skip macros.
4749 ;; Set the position to continue at. We can avoid going over
4750 ;; the comments skipped above a second time, but it's possible
4751 ;; that the comment skipping has taken us past `cfd-prop-match'
4752 ;; since the property might be used inside comments.
4758 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4759 ;; label from the point to CFD-LIMIT.
4760 ;;
4761 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4762 ;; arguments: The first is the end position of the token preceding the spot,
4763 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4764 ;; the match is inside a macro. Point should be moved forward by at least
4765 ;; one token.
4766 ;;
4767 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4768 ;; it should return non-nil to ensure that the next search will find them.
4769 ;;
4770 ;; Such a spot is:
4771 ;; o The first token after bob.
4772 ;; o The first token after the end of submatch 1 in
4773 ;; `c-decl-prefix-or-start-re' when that submatch matches.
4774 ;; o The start of each `c-decl-prefix-or-start-re' match when
4775 ;; submatch 1 doesn't match.
4776 ;; o The first token after the end of each occurrence of the
4777 ;; `c-type' text property with the value `c-decl-end', provided
4778 ;; `c-type-decl-end-used' is set.
4779 ;;
4780 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4781 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4782 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4783 ;;
4784 ;; If the match is inside a macro then the buffer is narrowed to the
4785 ;; end of it, so that CFD-FUN can investigate the following tokens
4786 ;; without matching something that begins inside a macro and ends
4787 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4788 ;; CFD-FACE-CHECKLIST checks exist.
4789 ;;
4790 ;; The spots are visited approximately in order from top to bottom.
4791 ;; It's however the positions where `c-decl-prefix-or-start-re'
4792 ;; matches and where `c-decl-end' properties are found that are in
4793 ;; order. Since the spots often are at the following token, they
4794 ;; might be visited out of order insofar as more spots are reported
4795 ;; later on within the syntactic whitespace between the match
4796 ;; positions and their spots.
4797 ;;
4798 ;; It's assumed that comments and strings are fontified in the
4799 ;; searched range.
4800 ;;
4801 ;; This is mainly used in fontification, and so has an elaborate
4802 ;; cache to handle repeated calls from the same start position; see
4803 ;; the variables above.
4804 ;;
4805 ;; All variables in this function begin with `cfd-' to avoid name
4806 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4807 ;;
4808 ;; This function might do hidden buffer changes.
4812 ;; The end of the token preceding the decl spot last found
4813 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4814 ;; no match.
4815 cfd-re-match
4816 ;; The end position of the last `c-decl-prefix-or-start-re'
4817 ;; match. If this is greater than `cfd-continue-pos', the
4818 ;; next regexp search is started here instead.
4820 ;; The end of the last `c-decl-end' found by
4821 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4822 ;; match. If searching for the property isn't needed then we
4823 ;; disable it by setting it to `cfd-limit' directly.
4825 ;; The end of the token preceding the decl spot last found by
4826 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4827 ;; bob. `cfd-limit' if there's no match. In other words,
4828 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4830 ;; The position to continue searching at.
4831 cfd-continue-pos
4832 ;; The position of the last "real" token we've stopped at.
4833 ;; This can be greater than `cfd-continue-pos' when we get
4834 ;; hits inside macros or at `c-decl-end' positions inside
4835 ;; comments.
4837 ;; The end position of the last entered macro.
4840 ;; Initialize by finding a syntactically relevant start position
4841 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4842 ;; search unless we're at bob.
4845 ;; Must back up a bit since we look for the end of the previous
4846 ;; statement or declaration, which is earlier than the first
4847 ;; returned match.
4850 ;; First we need to move to a syntactically relevant position.
4851 ;; Begin by backing out of comment or string literals.
4854 ;; Try to use the faces to back up to the start of the
4855 ;; literal. FIXME: What if the point is on a declaration
4856 ;; inside a comment?
4862 ;; XEmacs doesn't fontify the quotes surrounding string
4863 ;; literals.
4872 ;; Don't trust the literal to contain only literal faces
4873 ;; (the font lock package might not have fontified the
4874 ;; start of it at all, for instance) so check that we have
4875 ;; arrived at something that looks like a start or else
4876 ;; resort to `c-literal-limits'.
4883 ;; The start is in a literal. If the limit is in the same
4884 ;; one we don't have to find a syntactic position etc. We
4885 ;; only check that if the limit is at or before bonl to save
4886 ;; time; it covers the by far most common case when font-lock
4887 ;; refontifies the current line only.
4898 ;; Completely inside a literal. Set up variables to trig the
4899 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
4900 ;; find a suitable start position.
4903 ;; Check if the region might be completely inside a macro, to
4904 ;; optimize that like the completely-inside-literal above.
4911 ;; (Maybe) completely inside a macro. Only need to trig the
4912 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
4913 ;; set things up.
4915 start-in-macro t))
4918 ;; Back out of any macro so we don't miss any declaration
4919 ;; that could follow after it.
4923 ;; Now we're at a proper syntactically relevant position so we
4924 ;; can use the cache. But first clear it if it applied
4925 ;; further down.
4930 ;; Don't have to do this if the cache is relevant here,
4931 ;; typically if the same line is refontified again. If
4932 ;; we're just some syntactic whitespace further down we can
4933 ;; still use the cache to limit the skipping.
4936 ;; If we hit `c-find-decl-syntactic-pos' and
4937 ;; `c-find-decl-match-pos' is set then we install the cached
4938 ;; values. If we hit `c-find-decl-syntactic-pos' and
4939 ;; `c-find-decl-match-pos' is nil then we know there's no decl
4940 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
4941 ;; and so we can continue the search from this point. If we
4942 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
4943 ;; the right spot to begin searching anyway.
4945 c-find-decl-match-pos)
4947 cfd-continue-pos syntactic-pos)
4952 ;; Always consider bob a match to get the first
4953 ;; declaration in the file. Do this separately instead of
4954 ;; letting `c-decl-prefix-or-start-re' match bob, so that
4955 ;; regexp always can consume at least one character to
4956 ;; ensure that we won't get stuck in an infinite loop.
4961 ;; Do an initial search now. In the bob case above it's
4962 ;; only done to search for a `c-decl-end' spot.
4966 cfd-match-pos)))))
4968 ;; Advance `cfd-continue-pos' if it's before the start position.
4969 ;; The closest continue position that might have effect at or
4970 ;; after the start depends on what we started in. This also
4971 ;; finds a suitable start position in the special cases when the
4972 ;; region is completely within a literal or macro.
4977 ;; If we're in a macro then it's the closest preceding token
4978 ;; in the macro. Check this before `start-in-literal',
4979 ;; since if we're inside a literal in a macro, the preceding
4980 ;; token is earlier than any `c-decl-end' spot inside the
4981 ;; literal (comment).
4983 ;; The only syntactic ws in macros are comments.
4989 ;; If we're in a comment it can only be the closest
4990 ;; preceding `c-decl-end' position within that comment, if
4991 ;; any. Go back to the beginning of such a property so that
4992 ;; `c-find-decl-prefix-search' will find the end of it.
4993 ;; (Can't stop at the end and install it directly on
4994 ;; `cfd-prop-match' since that variable might be cleared
4995 ;; after `cfd-fun' below.)
4996 ;;
4997 ;; Note that if the literal is a string then the property
4998 ;; search will simply skip to the beginning of it right
4999 ;; away.
5011 ;; Didn't find any property inside the comment, so we can
5012 ;; skip it entirely. (This won't skip past a string, but
5013 ;; that'll be handled quickly by the next
5014 ;; `c-find-decl-prefix-search' anyway.)
5020 ;; If we started in normal code, the only match that might
5021 ;; apply before the start is what we already got in
5022 ;; `cfd-match-pos' so we can continue at the start position.
5023 ;; (Note that we don't get here if the first match is below
5024 ;; it.)
5027 ;; Delete found matches if they are before our new continue
5028 ;; position, so that `c-find-decl-prefix-search' won't back up
5029 ;; to them later on.
5037 ;; This is the normal case and we got a proper syntactic
5038 ;; position. If there's a match then it's always outside
5039 ;; macros and comments, so advance to the next token and set
5040 ;; `cfd-token-pos'. The loop below will later go back using
5041 ;; `cfd-continue-pos' to fix declarations inside the
5042 ;; syntactic ws.
5050 ;; Have one of the special cases when the region is completely
5051 ;; within a literal or macro. `cfd-continue-pos' is set to a
5052 ;; good start position for the search, so do it.
5055 ;; Now loop. Round what? (ACM, 2006/7/5). We already got the first match.
5062 ;; Kludge to filter out matches on the "<" that
5063 ;; aren't open parens, for the sake of languages
5064 ;; that got `c-recognize-<>-arglists' set.
5069 ;; If `cfd-continue-pos' is less or equal to
5070 ;; `cfd-token-pos', we've got a hit inside a macro
5071 ;; that's in the syntactic whitespace before the last
5072 ;; "real" declaration we've checked. If they're equal
5073 ;; we've arrived at the declaration a second time, so
5074 ;; there's nothing to do.
5078 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5079 ;; we're still searching for declarations embedded in
5080 ;; the syntactic whitespace. In that case we need
5081 ;; only to skip comments and not macros, since they
5082 ;; can't be nested, and that's already been done in
5083 ;; `c-find-decl-prefix-search'.
5088 ;; Continue if the following token fails the
5089 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5096 t)))
5107 ;; Narrow to the end of the macro if we got a hit inside
5108 ;; one, to avoid recognizing things that start inside the
5109 ;; macro and end outside it.
5111 ;; Not in the same macro as in the previous round.
5122 t
5125 t)
5126 ;; The matched token was the last thing in the macro,
5127 ;; so the whole match is bogus.
5129 nil))))
5136 ;; Restore limits if we did macro narrowing above.
5143 ; cfd-match-pos, etc.
5145 \f
5146 ;; A cache for found types.
5148 ;; Buffer local variable that contains an obarray with the types we've
5149 ;; found. If a declaration is recognized somewhere we record the
5150 ;; fully qualified identifier in it to recognize it as a type
5151 ;; elsewhere in the file too. This is not accurate since we do not
5152 ;; bother with the scoping rules of the languages, but in practice the
5153 ;; same name is seldom used as both a type and something else in a
5154 ;; file, and we only use this as a last resort in ambiguous cases (see
5155 ;; `c-forward-decl-or-cast-1').
5156 ;;
5157 ;; Not every type need be in this cache. However, things which have
5158 ;; ceased to be types must be removed from it.
5159 ;;
5160 ;; Template types in C++ are added here too but with the template
5161 ;; arglist replaced with "<>" in references or "<" for the one in the
5162 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5163 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5164 ;; template specs can be fairly sized programs in themselves) and
5165 ;; improves the hit ratio (it's a type regardless of the template
5166 ;; args; it's just not the same type, but we're only interested in
5167 ;; recognizing types, not telling distinct types apart). Note that
5168 ;; template types in references are added here too; from the example
5169 ;; above there will also be an entry "Foo<".
5174 ;; Clears `c-found-types'.
5178 ;; Add the given region as a type in `c-found-types'. If the region
5179 ;; doesn't match an existing type but there is a type which is equal
5180 ;; to the given one except that the last character is missing, then
5181 ;; the shorter type is removed. That's done to avoid adding all
5182 ;; prefixes of a type as it's being entered and font locked. This
5183 ;; doesn't cover cases like when characters are removed from a type
5184 ;; or added in the middle. We'd need the position of point when the
5185 ;; font locking is invoked to solve this well.
5186 ;;
5187 ;; This function might do hidden buffer changes.
5194 ;; Remove the "NAME" from c-found-types, if present.
5198 ;; Return non-nil if the given region contains a type in
5199 ;; `c-found-types'.
5200 ;;
5201 ;; This function might do hidden buffer changes.
5203 c-found-types))
5206 ;; Return all the types in `c-found-types' as a sorted list of
5207 ;; strings.
5211 type-list)))
5212 c-found-types)
5215 ;; Shut up the byte compiler.
5219 ;; An after change function which, in conjunction with the info in
5220 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5221 ;; from `c-found-types', should this type have become stale. For
5222 ;; example, this happens to "foo" when "foo \n bar();" becomes
5223 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5224 ;; the fontification.
5225 ;;
5226 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5227 ;; type?
5247 ;; Changing the amount of (already existing) whitespace - don't do anything.
5252 ;; The syntactic relationship which defined a "found type" has been
5253 ;; destroyed.
5256 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5257 )))
5259 \f
5260 ;; Setting and removing syntax properties on < and > in languages (C++
5261 ;; and Java) where they can be template/generic delimiters as well as
5262 ;; their normal meaning of "less/greater than".
5264 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5265 ;; be delimiters, they are marked as such with the category properties
5266 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5268 ;; STRATEGY:
5269 ;;
5270 ;; It is impossible to determine with certainty whether a <..> pair in
5271 ;; C++ is two comparison operators or is template delimiters, unless
5272 ;; one duplicates a lot of a C++ compiler. For example, the following
5273 ;; code fragment:
5274 ;;
5275 ;; foo (a < b, c > d) ;
5276 ;;
5277 ;; could be a function call with two integer parameters (each a
5278 ;; relational expression), or it could be a constructor for class foo
5279 ;; taking one parameter d of templated type "a < b, c >". They are
5280 ;; somewhat easier to distinguish in Java.
5281 ;;
5282 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5283 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5284 ;; individually when their context so indicated. This gave rise to
5285 ;; intractable problems when one of a matching pair was deleted, or
5286 ;; pulled into a literal.]
5287 ;;
5288 ;; At each buffer change, the syntax-table properties are removed in a
5289 ;; before-change function and reapplied, when needed, in an
5290 ;; after-change function. It is far more important that the
5291 ;; properties get removed when they they are spurious than that they
5292 ;; be present when wanted.
5293 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5295 ;; POS (default point) is at a < character. If it is marked with
5296 ;; open paren syntax-table text property, remove the property,
5297 ;; together with the close paren property on the matching > (if
5298 ;; any).
5313 ;; POS (default point) is at a > character. If it is marked with
5314 ;; close paren syntax-table property, remove the property, together
5315 ;; with the open paren property on the matching < (if any).
5330 ;; POS (default point) is at a < or > character. If it has an
5331 ;; open/close paren syntax-table property, remove this property both
5332 ;; from the current character and its partner (which will also be
5333 ;; thusly marked).
5343 ;; POS (default point) is at a < character. If it is both marked
5344 ;; with open/close paren syntax-table property, and has a matching >
5345 ;; (also marked) which is after LIM, remove the property both from
5346 ;; the current > and its partner. Return t when this happens, nil
5347 ;; when it doesn't.
5361 t)))
5364 ;; POS (default point) is at a > character. If it is both marked
5365 ;; with open/close paren syntax-table property, and has a matching <
5366 ;; (also marked) which is before LIM, remove the property both from
5367 ;; the current < and its partner. Return t when this happens, nil
5368 ;; when it doesn't.
5382 t)))
5384 ;; Set by c-common-init in cc-mode.el.
5389 ;; Unmark certain pairs of "< .... >" which are currently marked as
5390 ;; template/generic delimiters. (This marking is via syntax-table
5391 ;; text properties).
5392 ;;
5393 ;; These pairs are those which are in the current "statement" (i.e.,
5394 ;; the region between the {, }, or ; before BEG and the one after
5395 ;; END), and which enclose any part of the interval (BEG END).
5396 ;;
5397 ;; Note that in C++ (?and Java), template/generic parens cannot
5398 ;; enclose a brace or semicolon, so we use these as bounds on the
5399 ;; region we must work on.
5400 ;;
5401 ;; This function is called from before-change-functions (via
5402 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5403 ;; and point is undefined, both at entry and exit.
5404 ;;
5405 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5406 ;; 2010-01-29.
5410 new-beg new-end need-new-beg need-new-end)
5411 ;; Locate the barrier before the changed region
5416 ;; Remove the syntax-table properties from each pertinent <...> pair.
5417 ;; Firsly, the ones with the < before beg and > after beg.
5422 ;; Locate the barrier after END.
5427 ;; Remove syntax-table properties from the remaining pertinent <...>
5428 ;; pairs, those with a > after end and < before end.
5433 ;; Extend the fontification region, if needed.
5445 ;; This is called from `after-change-functions' when
5446 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5447 ;; chars with paren syntax become part of another operator like "<<"
5448 ;; or ">=".
5449 ;;
5450 ;; This function might do hidden buffer changes.
5483 \f
5484 ;; Handling of small scale constructs like types and names.
5486 ;; Dynamically bound variable that instructs `c-forward-type' to also
5487 ;; treat possible types (i.e. those that it normally returns 'maybe or
5488 ;; 'found for) as actual types (and always return 'found for them).
5489 ;; This means that it records them in `c-record-type-identifiers' if
5490 ;; that is set, and that it adds them to `c-found-types'.
5493 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5494 ;; mark up successfully parsed arglists with paren syntax properties on
5495 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5496 ;; `c-type' property of each argument separating comma.
5497 ;;
5498 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5499 ;; all arglists for side effects (i.e. recording types), otherwise it
5500 ;; exploits any existing paren syntax properties to quickly jump to the
5501 ;; end of already parsed arglists.
5502 ;;
5503 ;; Marking up the arglists is not the default since doing that correctly
5504 ;; depends on a proper value for `c-restricted-<>-arglists'.
5507 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5508 ;; not accept arglists that contain binary operators.
5509 ;;
5510 ;; This is primarily used to handle C++ template arglists. C++
5511 ;; disambiguates them by checking whether the preceding name is a
5512 ;; template or not. We can't do that, so we assume it is a template
5513 ;; if it can be parsed as one. That usually works well since
5514 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5515 ;; in almost all cases would be pointless.
5516 ;;
5517 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5518 ;; should let the comma separate the function arguments instead. And
5519 ;; in a context where the value of the expression is taken, e.g. in
5520 ;; "if (a < b || c > d)", it's probably not a template.
5523 ;; Dynamically bound variables that instructs
5524 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5525 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5526 ;; `c-forward-label' to record the ranges of all the type and
5527 ;; reference identifiers they encounter. They will build lists on
5528 ;; these variables where each element is a cons of the buffer
5529 ;; positions surrounding each identifier. This recording is only
5530 ;; activated when `c-record-type-identifiers' is non-nil.
5531 ;;
5532 ;; All known types that can't be identifiers are recorded, and also
5533 ;; other possible types if `c-promote-possible-types' is set.
5534 ;; Recording is however disabled inside angle bracket arglists that
5535 ;; are encountered inside names and other angle bracket arglists.
5536 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5537 ;; instead.
5538 ;;
5539 ;; Only the names in C++ template style references (e.g. "tmpl" in
5540 ;; "tmpl<a,b>::foo") are recorded as references, other references
5541 ;; aren't handled here.
5542 ;;
5543 ;; `c-forward-label' records the label identifier(s) on
5544 ;; `c-record-ref-identifiers'.
5548 ;; This variable will receive a cons cell of the range of the last
5549 ;; single identifier symbol stepped over by `c-forward-name' if it's
5550 ;; successful. This is the range that should be put on one of the
5551 ;; record lists above by the caller. It's assigned nil if there's no
5552 ;; such symbol in the name.
5557 ;; Always true.
5567 ;; Always true.
5575 ;; Dynamically bound variable that instructs `c-forward-type' to
5576 ;; record the ranges of types that only are found. Behaves otherwise
5577 ;; like `c-record-type-identifiers'.
5581 ;; Used internally in `c-forward-keyword-clause' to move forward
5582 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5583 ;; possibly is prefixed by keywords and their associated clauses.
5584 ;; Try with a type/name first to not trip up on those that begin
5585 ;; with a keyword. Return t if a known or found type is moved
5586 ;; over. The point is clobbered if nil is returned. If range
5587 ;; recording is enabled, the identifier is recorded on as a type
5588 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5589 ;;
5590 ;; This macro might do hidden buffer changes.
5595 nil
5602 t)))
5605 ;; Used internally in `c-forward-keyword-clause' to move forward
5606 ;; over a comma separated list of types or names using
5607 ;; `c-forward-keyword-prefixed-id'.
5608 ;;
5609 ;; This macro might do hidden buffer changes.
5620 ;; Submatch MATCH in the current match data is assumed to surround a
5621 ;; token. If it's a keyword, move over it and any immediately
5622 ;; following clauses associated with it, stopping at the start of
5623 ;; the next token. t is returned in that case, otherwise the point
5624 ;; stays and nil is returned. The kind of clauses that are
5625 ;; recognized are those specified by `c-type-list-kwds',
5626 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5627 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5628 ;; and `c-<>-arglist-kwds'.
5629 ;;
5630 ;; This function records identifier ranges on
5631 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5632 ;; `c-record-type-identifiers' is non-nil.
5633 ;;
5634 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5635 ;; apply directly after the keyword, the type list is moved over
5636 ;; only when there is no unaccounted token before it (i.e. a token
5637 ;; that isn't moved over due to some other keyword list). The
5638 ;; identifier ranges in the list are still recorded if that should
5639 ;; be done, though.
5640 ;;
5641 ;; This function might do hidden buffer changes.
5644 ;; The call to `c-forward-<>-arglist' below is made after
5645 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5646 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5647 ;; should therefore be nil.
5659 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5664 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5669 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5676 ;; Use `c-forward-type' on every identifier we can find
5677 ;; inside the paren, to record the types.
5702 ;; If we are at the colon already, we move over the type
5703 ;; list after it.
5709 ;; Not at the colon, so stop here. But the identifier
5710 ;; ranges in the type list later on should still be
5711 ;; recorded.
5714 ;; If a keyword matched both one of the types above and
5715 ;; this one, we match `c-colon-type-list-re' after the
5716 ;; clause matched above.
5723 ;; There's a type after the `c-colon-type-list-re' match
5724 ;; after a keyword in `c-colon-type-list-kwds'.
5728 t)))
5730 ;; cc-mode requires cc-fonts.
5734 ;; The point is assumed to be at a "<". Try to treat it as the open
5735 ;; paren of an angle bracket arglist and move forward to the
5736 ;; corresponding ">". If successful, the point is left after the
5737 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5738 ;; returned. If ALL-TYPES is t then all encountered arguments in
5739 ;; the arglist that might be types are treated as found types.
5740 ;;
5741 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5742 ;; function handles text properties on the angle brackets and argument
5743 ;; separating commas.
5744 ;;
5745 ;; `c-restricted-<>-arglists' controls how lenient the template
5746 ;; arglist recognition should be.
5747 ;;
5748 ;; This function records identifier ranges on
5749 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5750 ;; `c-record-type-identifiers' is non-nil.
5751 ;;
5752 ;; This function might do hidden buffer changes.
5755 ;; If `c-record-type-identifiers' is set then activate
5756 ;; recording of any found types that constitute an argument in
5757 ;; the arglist.
5765 ;; `nconc' doesn't mind that the tail of
5766 ;; `c-record-found-types' is t.
5769 t)
5772 nil)))
5775 ;; Recursive part of `c-forward-<>-arglist'.
5776 ;;
5777 ;; This function might do hidden buffer changes.
5780 ;; Cover this so that any recorded found type ranges are
5781 ;; automatically lost if it turns out to not be an angle
5782 ;; bracket arglist. It's propagated through the return value
5783 ;; on successful completion.
5785 ;; List that collects the positions after the argument
5786 ;; separating ',' in the arglist.
5787 arg-start-pos)
5788 ;; If the '<' has paren open syntax then we've marked it as an angle
5789 ;; bracket arglist before, so skip to the end.
5797 t
5798 ;; Got unmatched paren angle brackets. We don't clear the paren
5799 ;; syntax properties and retry, on the basis that it's very
5800 ;; unlikely that paren angle brackets become operators by code
5801 ;; manipulation. It's far more likely that it doesn't match due
5802 ;; to narrowing or some temporary change.
5804 nil))
5809 ;; go forward one non-alphanumeric character (group) per iteration of
5810 ;; this loop.
5817 ;; All encountered identifiers are types, so set the
5818 ;; promote flag and parse the type.
5841 ;; Note: These regexps exploit the match order in \| so
5842 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5844 ;; Stop on ',', '|', '&', '+' and '-' to catch
5845 ;; common binary operators that could be between
5846 ;; two comparison expressions "a<b" and "c>d".
5848 nil t t))
5852 ;; Either an operator starting with '>' or the end of
5853 ;; the angle bracket arglist.
5860 ;; The angle bracket arglist is finished.
5872 ;; Either an operator starting with '<' or a nested arglist.
5876 ;; The '<' begins a multi-char operator.
5880 ;; We're at a nested <.....>
5886 ;; There's always an identifier before an angle
5887 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
5888 ;; or `c-<>-arglist-kwds'.
5905 ;; It was an angle bracket arglist.
5908 ;; Record the identifier before the template as a type
5909 ;; or reference depending on whether the arglist is last
5910 ;; in a qualified identifier.
5920 ;; At a "less than" operator.
5923 )))
5930 ;; Just another argument. Record the position. The
5931 ;; type check stuff that made us stop at it is at
5932 ;; the top of the loop.
5936 ;; Got a character that can't be in an angle bracket
5937 ;; arglist argument. Abort using `throw', since
5938 ;; it's useless to try to find a surrounding arglist
5939 ;; if we're nested.
5945 ;; The point is assumed to be directly after a ">". Try to treat it
5946 ;; as the close paren of an angle bracket arglist and move back to
5947 ;; the corresponding "<". If successful, the point is left at
5948 ;; the "<" and t is returned, otherwise the point isn't moved and
5949 ;; nil is returned. ALL-TYPES is passed on to
5950 ;; `c-forward-<>-arglist'.
5951 ;;
5952 ;; If the optional LIMIT is given, it bounds the backward search.
5953 ;; It's then assumed to be at a syntactically relevant position.
5954 ;;
5955 ;; This is a wrapper around `c-forward-<>-arglist'. See that
5956 ;; function for more details.
5965 t
5966 ;; See corresponding note in `c-forward-<>-arglist'.
5968 nil)
5975 t
5976 ;; Stopped at bob or a char that isn't allowed in an
5977 ;; arglist, so we've failed.
5979 nil)
5984 ;; If we moved then the "<" was part of some
5985 ;; multicharacter token.
5986 t
5992 ;; Matched the arglist. Break the while.
5994 nil)
5996 ;; We started from a non-paren ">" inside an
5997 ;; arglist.
5999 nil)
6001 ;; Matched a shorter arglist. Can be a nested
6002 ;; one so continue looking.
6004 t))
6005 t))))))
6010 ;; Move forward over a complete name if at the beginning of one,
6011 ;; stopping at the next following token. A keyword, as such,
6012 ;; doesn't count as a name. If the point is not at something that
6013 ;; is recognized as a name then it stays put.
6014 ;;
6015 ;; A name could be something as simple as "foo" in C or something as
6016 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6017 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6018 ;; int>::*volatile const" in C++ (this function is actually little
6019 ;; more than a `looking-at' call in all modes except those that,
6020 ;; like C++, have `c-recognize-<>-arglists' set).
6021 ;;
6022 ;; Return
6023 ;; o - nil if no name is found;
6024 ;; o - 'template if it's an identifier ending with an angle bracket
6025 ;; arglist;
6026 ;; o - 'operator of it's an operator identifier;
6027 ;; o - t if it's some other kind of name.
6028 ;;
6029 ;; This function records identifier ranges on
6030 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6031 ;; `c-record-type-identifiers' is non-nil.
6032 ;;
6033 ;; This function might do hidden buffer changes.
6036 ;; Turn off `c-promote-possible-types' here since we might
6037 ;; call `c-forward-<>-arglist' and we don't want it to promote
6038 ;; every suspect thing in the arglist to a type. We're
6039 ;; typically called from `c-forward-type' in this case, and
6040 ;; the caller only wants the top level type that it finds to
6041 ;; be promoted.
6042 c-promote-possible-types)
6048 ;; Check for keyword. We go to the last symbol in
6049 ;; `c-identifier-key' first.
6062 ;; "template" is only valid inside an
6063 ;; identifier if preceded by "::".
6068 t))
6070 ;; Handle a C++ operator or template identifier.
6074 ;; Got "... ::template".
6078 res subres))))
6081 ;; Got a cast operator.
6085 ;; Now we should match a sequence of either
6086 ;; '*', '&' or a name followed by ":: *",
6087 ;; where each can be followed by a sequence
6088 ;; of `c-opt-type-modifier-key'.
6091 t)
6101 t))))
6109 ;; Got some other operator.
6117 nil)
6119 ;; `id-start' is equal to `id-end' if we've jumped over
6120 ;; an identifier that doesn't end with a symbol token.
6121 ;; That can occur e.g. for Java import directives on the
6122 ;; form "foo.bar.*".
6129 res t)))
6139 ;; Got a concatenated identifier. This handles the
6140 ;; cases with tricky syntactic whitespace that aren't
6141 ;; covered in `c-identifier-key'.
6144 t)
6148 ;; Maybe an angle bracket arglist.
6156 c-last-identifier-range nil)
6161 ;; Continue if there's an identifier concatenation
6162 ;; operator after the template argument.
6168 t)
6173 nil)))
6174 )))))
6177 res))
6180 ;; Move forward over a type spec if at the beginning of one,
6181 ;; stopping at the next following token. The keyword "typedef"
6182 ;; isn't part of a type spec here.
6183 ;;
6184 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6185 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6186 ;; The current (2009-03-10) intention is to convert all uses of
6187 ;; `c-forward-type' to call with this parameter set, then to
6188 ;; eliminate it.
6189 ;;
6190 ;; Return
6191 ;; o - t if it's a known type that can't be a name or other
6192 ;; expression;
6193 ;; o - 'known if it's an otherwise known type (according to
6194 ;; `*-font-lock-extra-types');
6195 ;; o - 'prefix if it's a known prefix of a type;
6196 ;; o - 'found if it's a type that matches one in `c-found-types';
6197 ;; o - 'maybe if it's an identifier that might be a type; or
6198 ;; o - nil if it can't be a type (the point isn't moved then).
6199 ;;
6200 ;; The point is assumed to be at the beginning of a token.
6201 ;;
6202 ;; Note that this function doesn't skip past the brace definition
6203 ;; that might be considered part of the type, e.g.
6204 ;; "enum {a, b, c} foo".
6205 ;;
6206 ;; This function records identifier ranges on
6207 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6208 ;; `c-record-type-identifiers' is non-nil.
6209 ;;
6210 ;; This function might do hidden buffer changes.
6218 ;; Skip leading type modifiers. If any are found we know it's a
6219 ;; prefix of a type.
6228 ; "typedef".
6236 ;; In many languages the name can be used without the
6237 ;; prefix, so we add it to `c-found-types'.
6240 c-last-identifier-range)
6262 id-range c-last-identifier-range))))
6274 id-end)
6276 ;; Looking at a primitive or known type identifier. We've
6277 ;; checked for a name first so that we don't go here if the
6278 ;; known type match only is a prefix of another name.
6289 ;; There might be more keywords for the type.
6317 ;; A normal identifier.
6327 ;; It's an identifier that has been used as
6328 ;; a type somewhere else.
6329 'found
6330 ;; It's an identifier that might be a type.
6333 ;; A template is a type.
6337 ;; Otherwise it's an operator identifier, which is not a type.
6342 ;; Skip trailing type modifiers. If any are found we know it's
6343 ;; a type.
6349 ;; Step over any type suffix operator. Do not let the existence
6350 ;; of these alter the classification of the found type, since
6351 ;; these operators typically are allowed in normal expressions
6352 ;; too.
6359 ;; Look for a trailing operator that concatenates the type
6360 ;; with a following one, and if so step past that one through
6361 ;; a recursive call. Note that we don't record concatenated
6362 ;; types in `c-found-types' - it's the component types that
6363 ;; are recorded when appropriate.
6366 c-promote-possible-types))
6367 ;; If we can't promote then set `c-record-found-types' so that
6368 ;; we can merge in the types from the second part afterwards if
6369 ;; it turns out to be a known type there.
6372 subres)
6381 ;; If either operand certainly is a type then both are, but we
6382 ;; don't let the existence of the operator itself promote two
6383 ;; uncertain types to a certain one.
6402 ;; Merge in the ranges of any types found by the second
6403 ;; `c-forward-type'.
6405 ;; `nconc' doesn't mind that the tail of
6406 ;; `c-record-found-types' is t.
6408 c-record-type-identifiers))))
6416 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6418 res))
6421 ;; Used for Java code only at the moment. Assumes point is on the
6422 ;; @, moves forward an annotation. returns nil if there is no
6423 ;; annotation at point.
6430 t)))
6432 \f
6433 ;; Handling of large scale constructs like statements and declarations.
6435 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6436 ;; defsubst or perhaps even a defun, but it contains lots of free
6437 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6439 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6440 ;; of types when parsing a declaration, which means that it
6441 ;; sometimes consumes the identifier in the declaration as a type.
6442 ;; This is used to "backtrack" and make the last type be treated as
6443 ;; an identifier instead.
6446 ;; These identifiers are bound only in the inner let.
6448 identifier-start type-start
6449 got-parens nil
6450 got-identifier t
6451 got-suffix t
6452 got-suffix-after-parens id-start
6456 t
6457 backup-at-type))
6459 id-start backup-id-start)
6461 id-start start-pos))
6463 ;; When these flags already are set we've found specifiers that
6464 ;; unconditionally signal these attributes - backtracking doesn't
6465 ;; change that. So keep them set in that case.
6472 ;; This identifier is bound only in the inner let.
6476 ;; Move forward over a declaration or a cast if at the start of one.
6477 ;; The point is assumed to be at the start of some token. Nil is
6478 ;; returned if no declaration or cast is recognized, and the point
6479 ;; is clobbered in that case.
6480 ;;
6481 ;; If a declaration is parsed:
6482 ;;
6483 ;; The point is left at the first token after the first complete
6484 ;; declarator, if there is one. The return value is a cons where
6485 ;; the car is the position of the first token in the declarator. (See
6486 ;; below for the cdr.)
6487 ;; Some examples:
6488 ;;
6489 ;; void foo (int a, char *b) stuff ...
6490 ;; car ^ ^ point
6491 ;; float (*a)[], b;
6492 ;; car ^ ^ point
6493 ;; unsigned int a = c_style_initializer, b;
6494 ;; car ^ ^ point
6495 ;; unsigned int a (cplusplus_style_initializer), b;
6496 ;; car ^ ^ point (might change)
6497 ;; class Foo : public Bar {}
6498 ;; car ^ ^ point
6499 ;; class PikeClass (int a, string b) stuff ...
6500 ;; car ^ ^ point
6501 ;; enum bool;
6502 ;; car ^ ^ point
6503 ;; enum bool flag;
6504 ;; car ^ ^ point
6505 ;; void cplusplus_function (int x) throw (Bad);
6506 ;; car ^ ^ point
6507 ;; Foo::Foo (int b) : Base (b) {}
6508 ;; car ^ ^ point
6509 ;;
6510 ;; The cdr of the return value is non-nil when a
6511 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6512 ;; Specifically it is a dotted pair (A . B) where B is t when a
6513 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6514 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6515 ;; specifier is present. I.e., (some of) the declared
6516 ;; identifier(s) are types.
6517 ;;
6518 ;; If a cast is parsed:
6519 ;;
6520 ;; The point is left at the first token after the closing paren of
6521 ;; the cast. The return value is `cast'. Note that the start
6522 ;; position must be at the first token inside the cast parenthesis
6523 ;; to recognize it.
6524 ;;
6525 ;; PRECEDING-TOKEN-END is the first position after the preceding
6526 ;; token, i.e. on the other side of the syntactic ws from the point.
6527 ;; Use a value less than or equal to (point-min) if the point is at
6528 ;; the first token in (the visible part of) the buffer.
6529 ;;
6530 ;; CONTEXT is a symbol that describes the context at the point:
6531 ;; 'decl In a comma-separated declaration context (typically
6532 ;; inside a function declaration arglist).
6533 ;; '<> In an angle bracket arglist.
6534 ;; 'arglist Some other type of arglist.
6535 ;; nil Some other context or unknown context. Includes
6536 ;; within the parens of an if, for, ... construct.
6537 ;;
6538 ;; LAST-CAST-END is the first token after the closing paren of a
6539 ;; preceding cast, or nil if none is known. If
6540 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6541 ;; the position after the closest preceding call where a cast was
6542 ;; matched. In that case it's used to discover chains of casts like
6543 ;; "(a) (b) c".
6544 ;;
6545 ;; This function records identifier ranges on
6546 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6547 ;; `c-record-type-identifiers' is non-nil.
6548 ;;
6549 ;; This function might do hidden buffer changes.
6552 ;; first type, i.e. after any leading specifiers. It might
6553 ;; also point at the beginning of the preceding syntactic
6554 ;; whitespace.
6556 ;; Set to the result of `c-forward-type'.
6557 at-type
6558 ;; The position of the first token in what we currently
6559 ;; believe is the type in the declaration or cast, after any
6560 ;; specifiers and their associated clauses.
6561 type-start
6562 ;; The position of the first token in what we currently
6563 ;; believe is the declarator for the first identifier. Set
6564 ;; when the type is found, and moved forward over any
6565 ;; `c-decl-hangon-kwds' and their associated clauses that
6566 ;; occurs after the type.
6567 id-start
6568 ;; These store `at-type', `type-start' and `id-start' of the
6569 ;; identifier before the one in those variables. The previous
6570 ;; identifier might turn out to be the real type in a
6571 ;; declaration if the last one has to be the declarator in it.
6572 ;; If `backup-at-type' is nil then the other variables have
6573 ;; undefined values.
6574 backup-at-type backup-type-start backup-id-start
6575 ;; Set if we've found a specifier (apart from "typedef") that makes
6576 ;; the defined identifier(s) types.
6577 at-type-decl
6578 ;; Set if we've a "typedef" keyword.
6579 at-typedef
6580 ;; Set if we've found a specifier that can start a declaration
6581 ;; where there's no type.
6582 maybe-typeless
6583 ;; If a specifier is found that also can be a type prefix,
6584 ;; these flags are set instead of those above. If we need to
6585 ;; back up an identifier, they are copied to the real flag
6586 ;; variables. Thus they only take effect if we fail to
6587 ;; interpret it as a type.
6588 backup-at-type-decl backup-maybe-typeless
6589 ;; Whether we've found a declaration or a cast. We might know
6590 ;; this before we've found the type in it. It's 'ids if we've
6591 ;; found two consecutive identifiers (usually a sure sign, but
6592 ;; we should allow that in labels too), and t if we've found a
6593 ;; specifier keyword (a 100% sure sign).
6594 at-decl-or-cast
6595 ;; Set when we need to back up to parse this as a declaration
6596 ;; but not as a cast.
6597 backup-if-not-cast
6598 ;; For casts, the return position.
6599 cast-end
6600 ;; Save `c-record-type-identifiers' and
6601 ;; `c-record-ref-identifiers' since ranges are recorded
6602 ;; speculatively and should be thrown away if it turns out
6603 ;; that it isn't a declaration or cast.
6610 ;; Check for a type. Unknown symbols are treated as possible
6611 ;; types, but they could also be specifiers disguised through
6612 ;; macros like __INLINE__, so we recognize both types and known
6613 ;; specifiers after them too.
6617 ;; Look for a specifier keyword clause.
6629 ;; Found a known or possible type or a prefix of a known type.
6632 ;; Got two identifiers with nothing but whitespace
6633 ;; between them. That can only happen in declarations.
6637 ;; If the previous identifier is a found type we
6638 ;; record it as a real one; it might be some sort of
6639 ;; alias for a prefix like "unsigned".
6646 backup-type-start type-start
6647 backup-id-start id-start
6648 at-type found-type
6649 type-start start
6651 ;; The previous ambiguous specifier/type turned out
6652 ;; to be a type since we've parsed another one after
6653 ;; it, so clear these backup flags.
6654 backup-at-type-decl nil
6655 backup-maybe-typeless nil))
6659 ;; Handle known specifier keywords and
6660 ;; `c-decl-hangon-kwds' which can occur after known
6661 ;; types.
6664 ;; It's a hang-on keyword that can occur anywhere.
6668 ;; Move the identifier start position if
6669 ;; we've passed a type.
6671 ;; Otherwise treat this as a specifier and
6672 ;; move the fallback position.
6676 ;; It's an ordinary specifier so we know that
6677 ;; anything before this can't be the type.
6679 start-pos kwd-clause-end)
6682 ;; It's ambiguous whether this keyword is a
6683 ;; specifier or a type prefix, so set the backup
6684 ;; flags. (It's assumed that `c-forward-type'
6685 ;; moved further than `c-forward-keyword-clause'.)
6693 ;; This test only happens after we've scanned a type.
6694 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6699 ;; Haven't matched a type so it's an unambiguous
6700 ;; specifier keyword and we know we're in a
6701 ;; declaration.
6706 ;; If the type isn't known we continue so that we'll jump
6707 ;; over all specifiers and type identifiers. The reason
6708 ;; to do this for a known type prefix is to make things
6709 ;; like "unsigned INT16" work.
6714 ;; If a known type was found, we still need to skip over any
6715 ;; hangon keyword clauses after it. Otherwise it has already
6716 ;; been done in the loop above.
6722 ;; A prefix type is itself a primitive type when it's not
6723 ;; followed by another type.
6727 ;; Got no type but set things up to continue anyway to handle
6728 ;; the various cases when a declaration doesn't start with a
6729 ;; type.
6734 ;; If it's C++ then check if the last "type" ends on the form
6735 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
6736 ;; (con|de)structor.
6747 ;; Cheating in the handling of syntactic ws below.
6754 name))
6755 ;; It is a (con|de)structor name. In that case the
6756 ;; declaration is typeless so zap out any preceding
6757 ;; identifier(s) that we might have taken as types.
6760 backup-at-type nil
6761 id-start type-start))))))
6763 ;; Check for and step over a type decl expression after the thing
6764 ;; that is or might be a type. This can't be skipped since we
6765 ;; need the correct end position of the declarator for
6766 ;; `max-type-decl-end-*'.
6768 ;; True if there's a non-open-paren match of
6769 ;; `c-type-decl-prefix-key'.
6770 got-prefix
6771 ;; True if the declarator is surrounded by a parenthesis pair.
6772 got-parens
6773 ;; True if there is an identifier in the declarator.
6774 got-identifier
6775 ;; True if there's a non-close-paren match of
6776 ;; `c-type-decl-suffix-key'.
6777 got-suffix
6778 ;; True if there's a prefix match outside the outermost
6779 ;; paren pair that surrounds the declarator.
6780 got-prefix-before-parens
6781 ;; True if there's a suffix match outside the outermost
6782 ;; paren pair that surrounds the declarator. The value is
6783 ;; the position of the first suffix match.
6784 got-suffix-after-parens
6785 ;; True if we've parsed the type decl to a token that is
6786 ;; known to end declarations in this context.
6787 at-decl-end
6788 ;; The earlier values of `at-type' and `type-start' if we've
6789 ;; shifted the type backwards.
6790 identifier-type identifier-start
6791 ;; If `c-parse-and-markup-<>-arglists' is set we need to
6792 ;; turn it off during the name skipping below to avoid
6793 ;; getting `c-type' properties that might be bogus. That
6794 ;; can happen since we don't know if
6795 ;; `c-restricted-<>-arglists' will be correct inside the
6796 ;; arglist paren that gets entered.
6801 ;; Skip over type decl prefix operators. (Note similar code in
6802 ;; `c-font-lock-declarators'.)
6806 ;; If the second submatch matches in C++ then
6807 ;; we're looking at an identifier that's a
6808 ;; prefix only if it specifies a member pointer.
6811 ;; We only check for a trailing "::" and
6812 ;; let the "*" that should follow be
6813 ;; matched in the next round.
6815 ;; It turned out to be the real identifier,
6816 ;; so stop.
6817 nil))
6818 t))
6832 ;; Skip over an identifier.
6837 ;; Skip over type decl suffix operators.
6844 t)
6848 t)
6854 ;; No suffix matched. We might have matched the
6855 ;; identifier as a type and the open paren of a
6856 ;; function arglist as a type decl prefix. In that
6857 ;; case we should "backtrack": Reinterpret the last
6858 ;; type as the identifier, move out of the arglist and
6859 ;; continue searching for suffix operators.
6860 ;;
6861 ;; Do this even if there's no preceding type, to cope
6862 ;; with old style function declarations in K&R C,
6863 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
6864 ;; style declarations. That isn't applicable in an
6865 ;; arglist context, though.
6870 maybe-typeless
6871 backup-maybe-typeless
6878 t))
6885 at-type)
6886 ;; Have found no identifier but `c-typeless-decl-kwds' has
6887 ;; matched so we know we're inside a declaration. The
6888 ;; preceding type must be the identifier instead.
6892 at-decl-or-cast
6895 ;; CASE 1
6897 ;; Encountered something inside parens that isn't matched by
6898 ;; the `c-type-decl-*' regexps, so it's not a type decl
6899 ;; expression. Try to skip out to the same paren depth to
6900 ;; not confuse the cast check below.
6902 ;; If we've found a specifier keyword then it's a
6903 ;; declaration regardless.
6911 ;; Now we've collected info about various characteristics of
6912 ;; the construct we're looking at. Below follows a decision
6913 ;; tree based on that. It's ordered to check more certain
6914 ;; signs before less certain ones.
6919 ;; CASE 2
6922 ;; Got another identifier directly after the type, so it's a
6923 ;; declaration.
6930 maybe-typeless
6931 backup-maybe-typeless))
6932 ;; Got a declaration of the form "foo bar (gnu);" where we've
6933 ;; recognized "bar" as the type and "gnu" as the declarator.
6934 ;; In this case it's however more likely that "bar" is the
6935 ;; declarator and "gnu" a function argument or initializer (if
6936 ;; `c-recognize-paren-inits' is set), since the parens around
6937 ;; "gnu" would be superfluous if it's a declarator. Shift the
6938 ;; type one step backward.
6941 ;; Found no identifier.
6947 ;; CASE 3
6949 ;; Got a plain list of identifiers. If a colon follows it's
6950 ;; a valid label, or maybe a bitfield. Otherwise the last
6951 ;; one probably is the declared identifier and we should
6952 ;; back up to the previous type, providing it isn't a cast.
6956 ;; If we've found a specifier keyword then it's a
6957 ;; declaration regardless.
6968 ;; CASE 4
6972 ;; Got a plain list of identifiers followed by some suffix.
6973 ;; If this isn't a cast then the last identifier probably is
6974 ;; the declared one and we should back up to the previous
6975 ;; type.
6979 ;; CASE 5
6981 ;; If the type is known we know that there can't be any
6982 ;; identifier somewhere else, and it's only in declarations in
6983 ;; e.g. function prototypes and in casts that the identifier may
6984 ;; be left out.
6988 ;; Only got a single identifier (parsed as a type so far).
6989 ;; CASE 6
6991 ;; Check that the identifier isn't at the start of an
6992 ;; expression.
6993 at-decl-end
6996 ;; Inside an arglist that contains declarations. If K&R
6997 ;; style declarations and parenthesis style initializers
6998 ;; aren't allowed then the single identifier must be a
6999 ;; type, else we require that it's known or found
7000 ;; (primitive types are handled above).
7005 ;; Inside a template arglist. Accept known and found
7006 ;; types; other identifiers could just as well be
7007 ;; constants in C++.
7010 ;; CASE 7
7011 ;; Can't be a valid declaration or cast, but if we've found a
7012 ;; specifier it can't be anything else either, so treat it as
7013 ;; an invalid/unfinished declaration or cast.
7021 maybe-typeless
7022 backup-maybe-typeless
7026 c-after-suffixed-type-maybe-decl-key))))))
7027 ;; Got an empty paren pair and a preceding type that probably
7028 ;; really is the identifier. Shift the type backwards to make
7029 ;; the last one the identifier. This is analogous to the
7030 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7031 ;; above.
7032 ;;
7033 ;; Exception: In addition to the conditions in that
7034 ;; "backtracking" code, do not shift backward if we're not
7035 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7036 ;; Since there's no preceding type, the shift would mean that
7037 ;; the declaration is typeless. But if the regexp doesn't match
7038 ;; then we will simply fall through in the tests below and not
7039 ;; recognize it at all, so it's better to try it as an abstract
7040 ;; declarator instead.
7043 ;; Still no identifier.
7044 ;; CASE 8
7046 ;; Require `got-prefix' together with either `got-parens' or
7047 ;; `got-suffix' to recognize it as an abstract declarator:
7048 ;; `got-parens' only is probably an empty function call.
7049 ;; `got-suffix' only can build an ordinary expression together
7050 ;; with the preceding identifier which we've taken as a type.
7051 ;; We could actually accept on `got-prefix' only, but that can
7052 ;; easily occur temporarily while writing an expression so we
7053 ;; avoid that case anyway. We could do a better job if we knew
7054 ;; the point when the fontification was invoked.
7057 ;; CASE 9
7061 got-suffix-after-parens
7063 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7064 ;; normal function call after all (or perhaps a C++ style object
7065 ;; instantiation expression).
7068 ;; CASE 10
7070 ;; By now we've located the type in the declaration that we know
7071 ;; we're in.
7074 ;; CASE 11
7082 ;; Shift the type backward in the case that there's a
7083 ;; single identifier inside parens. That can only
7084 ;; occur in K&R style function declarations so it's
7085 ;; more likely that it really is a function call.
7086 ;; Therefore we only do this after
7087 ;; `c-after-suffixed-type-decl-key' has matched.
7089 got-suffix-after-parens))
7090 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7093 ;; CASE 12
7096 ;; It's a declaration if a known type precedes it and it can't be a
7097 ;; function call.
7100 ;; If we get here we can't tell if this is a type decl or a normal
7101 ;; expression by looking at it alone. (That's under the assumption
7102 ;; that normal expressions always can look like type decl expressions,
7103 ;; which isn't really true but the cases where it doesn't hold are so
7104 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7105 ;; the effort to look for them.)
7108 ;; If this is a declaration it should end here or its initializer(*)
7109 ;; should start here, so check for allowed separation tokens. Note
7110 ;; that this rule doesn't work e.g. with a K&R arglist after a
7111 ;; function header.
7112 ;;
7113 ;; *) Don't check for C++ style initializers using parens
7114 ;; since those already have been matched as suffixes.
7115 ;;
7116 ;; If `at-decl-or-cast' is then we've found some other sign that
7117 ;; it's a declaration or cast, so then it's probably an
7118 ;; invalid/unfinished one.
7121 ;; Below are tests that only should be applied when we're certain to
7122 ;; not have parsed halfway through an expression.
7124 ;; CASE 14
7126 ;; The expression starts with a known type so treat it as a
7127 ;; declaration.
7130 ;; CASE 15
7132 ;; In C++ we check if the identifier is a known type, since
7133 ;; (con|de)structors use the class name as identifier.
7134 ;; We've always shifted over the identifier as a type and
7135 ;; then backed up again in this case.
7136 identifier-type
7139 ;; `at-type' probably won't be 'found for
7140 ;; destructors since the "~" is then part of the
7141 ;; type name being checked against the list of
7142 ;; known types, so do a check without that
7143 ;; operator.
7148 c-identifier-syntax-table
7152 ;; We have already parsed the type earlier,
7153 ;; so it'd be possible to cache the end
7154 ;; position instead of redoing it here, but
7155 ;; then we'd need to keep track of another
7156 ;; position everywhere.
7164 ;; CASE 16
7166 at-type
7170 ;; Got something like "foo * bar;". Since we're not inside an
7171 ;; arglist it would be a meaningless expression because the
7172 ;; result isn't used. We therefore choose to recognize it as
7173 ;; a declaration. Do not allow a suffix since it could then
7174 ;; be a function call.
7177 ;; CASE 17
7182 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7183 ;; be an odd expression or it could be a declaration. Treat
7184 ;; it as a declaration if "a" has been used as a type
7185 ;; somewhere else (if it's a known type we won't get here).
7188 ;; CASE 18
7194 ;; Got a type followed by an abstract declarator. If `got-prefix'
7195 ;; is set it's something like "a *" without anything after it. If
7196 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7197 ;; or similar, which we accept only if the context rules out
7198 ;; expressions.
7201 ;; If we had a complete symbol table here (which rules out
7202 ;; `c-found-types') we should return t due to the disambiguation rule
7203 ;; (in at least C++) that anything that can be parsed as a declaration
7204 ;; is a declaration. Now we're being more defensive and prefer to
7205 ;; highlight things like "foo (bar);" as a declaration only if we're
7206 ;; inside an arglist that contains declarations.
7209 ;; The point is now after the type decl expression.
7212 ;; Check for a cast.
7215 c-cast-parens
7217 ;; Should be the first type/identifier in a cast paren.
7221 ;; The closing paren should follow.
7226 ;; There should be a primary expression after it.
7235 ;; Check if the expression begins with a prefix keyword.
7238 ;; Expression begins with an ambiguous operator. Treat
7239 ;; it as a cast if it's a type decl or if we've
7240 ;; recognized the type somewhere else.
7243 ;; Unless it's a keyword, it's the beginning of a primary
7244 ;; expression.
7246 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7247 ;; that it matched a whole one so that we don't e.g. confuse
7248 ;; the operator '-' with '->'. It's ok if it matches further,
7249 ;; though, since it e.g. can match the float '.5' while the
7250 ;; operator regexp only matches '.'.
7254 ;; There should either be a cast before it or something that isn't an
7255 ;; identifier or close paren.
7263 ;; It's a symbol. Accept it only if it's one of the
7264 ;; keywords that can precede an expression (without
7265 ;; surrounding parens).
7268 ;; Check that it isn't a close paren (block close is ok,
7269 ;; though).
7271 ;; Check that it isn't a nonsymbol identifier.
7274 ;; Handle the cast.
7284 ;; We're at a declaration. Highlight the type and the following
7285 ;; declarators.
7291 ;; Make sure to propagate the `c-decl-arg-start' property to
7292 ;; the next argument if it's set in this one, to cope with
7293 ;; interactive refontification.
7307 ;; False alarm. Restore the recorded ranges.
7309 c-record-ref-identifiers save-rec-ref-ids)
7310 nil))))
7313 ;; Assuming that point is at the beginning of a token, check if it starts a
7314 ;; label and if so move over it and return non-nil (t in default situations,
7315 ;; specific symbols (see below) for interesting situations), otherwise don't
7316 ;; move and return nil. "Label" here means "most things with a colon".
7317 ;;
7318 ;; More precisely, a "label" is regarded as one of:
7319 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7320 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7321 ;; bare "case", should the colon be missing. We return t;
7322 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7323 ;; return t;
7324 ;; (iv) One of QT's "extended" C++ variants of
7325 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7326 ;; Returns the symbol `qt-2kwds-colon'.
7327 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7328 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7329 ;; colon). Currently (2006-03), this applies only to Objective C's
7330 ;; keywords "@private", "@protected", and "@public". Returns t.
7331 ;;
7332 ;; One of the things which will NOT be recognized as a label is a bit-field
7333 ;; element of a struct, something like "int foo:5".
7334 ;;
7335 ;; The end of the label is taken to be just after the colon, or the end of
7336 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7337 ;; after the end on return. The terminating char gets marked with
7338 ;; `c-decl-end' to improve recognition of the following declaration or
7339 ;; statement.
7340 ;;
7341 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7342 ;; label, if any, has already been marked up like that.
7343 ;;
7344 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7345 ;; after the preceding token, i.e. on the other side of the
7346 ;; syntactic ws from the point. Use a value less than or equal to
7347 ;; (point-min) if the point is at the first token in (the visible
7348 ;; part of) the buffer.
7349 ;;
7350 ;; The optional LIMIT limits the forward scan for the colon.
7351 ;;
7352 ;; This function records the ranges of the label symbols on
7353 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7354 ;; non-nil.
7355 ;;
7356 ;; This function might do hidden buffer changes.
7359 label-end
7360 qt-symbol-idx
7361 macro-start ; if we're in one.
7362 label-type
7363 kwd)
7365 ;; "case" or "default" (Doesn't apply to AWK).
7368 ;; Record only the keyword itself for fontification, since in
7369 ;; case labels the following is a constant expression and not
7370 ;; a label.
7374 ;; Find the label end.
7378 ;; Stop on chars that aren't allowed in expressions,
7379 ;; and on operator chars that would be meaningless
7380 ;; there. FIXME: This doesn't cope with ?: operators.
7388 t)
7390 ;; It's an unfinished label. We consider the keyword enough
7391 ;; to recognize it as a label, so that it gets fontified.
7392 ;; Leave the point at the end of it, but don't put any
7393 ;; `c-decl-end' marker.
7395 t))))
7397 ;; @private, @protected, @public, in Objective C, or similar.
7400 ;; For a `c-opt-extra-label-key' match, we record the whole
7401 ;; thing for fontification. That's to get the leading '@' in
7402 ;; Objective-C protection labels fontified.
7409 ;; All other cases of labels.
7412 ;; A colon label must have something before the colon.
7415 ;; Check that we're not after a token that can't precede a label.
7417 ;; Trivially succeeds when there's no preceding token.
7418 ;; Succeeds when we're at a virtual semicolon.
7427 ;; Check if we're after a label, if we're after a closing
7428 ;; paren that belong to statement, and with
7429 ;; `c-label-prefix-re'. It's done in different order
7430 ;; depending on `assume-markup' since the checks have
7431 ;; different expensiveness.
7458 ;; Might be after another label, so check it recursively.
7462 ;; Essentially the same as the
7463 ;; `c-syntactic-re-search-forward' regexp below.
7469 ;; Note: the following should work instead of the
7470 ;; narrow-to-region above. Investigate why not,
7471 ;; sometime. ACM, 2006-03-31.
7472 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7473 ;; macro-start t)
7475 ;; If the caller turned on recording for us,
7476 ;; it shouldn't apply when we check the
7477 ;; preceding label.
7478 c-record-type-identifiers)
7479 ;; A label can't start at a cpp directive. Check for
7480 ;; this, since c-forward-syntactic-ws would foul up on it.
7485 ;; Point is still at the beginning of the possible label construct.
7486 ;;
7487 ;; Check that the next nonsymbol token is ":", or that we're in one
7488 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7489 ;; arguments. FIXME: Should build this regexp from the language
7490 ;; constants.
7492 ;; public: protected: private:
7502 ;; QT double keyword like "protected slots:" or goto target.
7521 'qt-1kwd-colon
7533 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7543 t)
7546 ;; Record the identifiers in the label for fontification, unless
7547 ;; it begins with `c-label-kwds' in which case the following
7548 ;; identifiers are part of a (constant) expression that
7549 ;; shouldn't be fontified.
7561 ;; Not a label.
7563 label-type))
7566 ;; Assuming the point is at the beginning of a token, try to move
7567 ;; forward to the end of the Objective-C directive that starts
7568 ;; there. Return t if a directive was fully recognized, otherwise
7569 ;; the point is moved as far as one could be successfully parsed and
7570 ;; nil is returned.
7571 ;;
7572 ;; This function records identifier ranges on
7573 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7574 ;; `c-record-type-identifiers' is non-nil.
7575 ;;
7576 ;; This function might do hidden buffer changes.
7579 start-char
7581 lim
7582 ;; Turn off recognition of angle bracket arglists while parsing
7583 ;; types here since the protocol reference list might then be
7584 ;; considered part of the preceding name or superclass-name.
7595 t)
7604 ;; Handle the name of the class itself.
7606 ; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7607 ; at EOB.
7614 ;; Look for ": superclass-name" or "( category-name )".
7625 ;; Look for a protocol reference list.
7631 t))))
7637 t)
7640 nil)))
7643 ;; Go to the first non-whitespace after the colon that starts a
7644 ;; multiple inheritance introduction. Optional LIM is the farthest
7645 ;; back we should search.
7646 ;;
7647 ;; This function might do hidden buffer changes.
7655 ;; Return nil if we aren't in a method definition, otherwise the
7656 ;; position of the initial [+-].
7657 ;;
7658 ;; This function might do hidden buffer changes.
7664 ))
7666 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7668 ;; Return non-nil if point is within a gcc \"asm\" block.
7669 ;;
7670 ;; This should be called with point inside an argument list.
7671 ;;
7672 ;; Only one level of enclosing parentheses is considered, so for
7673 ;; instance `nil' is returned when in a function call within an asm
7674 ;; operand.
7675 ;;
7676 ;; This function might do hidden buffer changes.
7688 a function.
7690 More precisely, being at the top-level means that point is either
7691 outside any enclosing block (such as a function definition), or
7692 directly inside a class, namespace or other block that contains
7693 another declaration level.
7695 If point is not at the top-level (e.g. it is inside a method
7696 definition), then nil is returned. Otherwise, if point is at a
7697 top-level not enclosed within a class definition, t is returned.
7698 Otherwise, a 2-vector is returned where the zeroth element is the
7699 buffer position of the start of the class declaration, and the first
7700 element is the buffer position of the enclosing class's opening
7701 brace.
7703 Note that this function might do hidden buffer changes. See the
7704 comment at the start of cc-engine.el for more info."
7710 ;; Return non-nil if the point is in the region after the argument
7711 ;; list of a function and its opening brace (or semicolon in case it
7712 ;; got no body). If there are K&R style argument declarations in
7713 ;; that region, the point has to be inside the first one for this
7714 ;; function to recognize it.
7715 ;;
7716 ;; If successful, the point is moved to the first token after the
7717 ;; function header (see `c-forward-decl-or-cast-1' for details) and
7718 ;; the position of the opening paren of the function arglist is
7719 ;; returned.
7720 ;;
7721 ;; The point is clobbered if not successful.
7722 ;;
7723 ;; LIM is used as bound for backward buffer searches.
7724 ;;
7725 ;; This function might do hidden buffer changes.
7738 ;; There should not be a '=' or ',' between beg and the
7739 ;; start of the declaration since that means we were in the
7740 ;; "expression part" of the declaration.
7745 ;; Check that there's an arglist paren in the
7746 ;; declaration.
7749 ;; The declarator is a paren expression, so skip past it
7750 ;; so that we don't get stuck on that instead of the
7751 ;; function arglist.
7755 ;; Don't trip up on "operator ()".
7762 ;; Return the position of the first argument declaration if point is
7763 ;; inside a K&R style argument declaration list, nil otherwise.
7764 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
7765 ;; position that bounds the backward search for the argument list.
7766 ;;
7767 ;; Point must be within a possible K&R region, e.g. just before a top-level
7768 ;; "{". It must be outside of parens and brackets. The test can return
7769 ;; false positives otherwise.
7770 ;;
7771 ;; This function might do hidden buffer changes.
7775 ;; If we're in a macro, our search range is restricted to it. Narrow to
7776 ;; the searchable range.
7780 before-lparen after-rparen
7782 ; we give up
7785 ;; Search backwards for the defun's argument list. We give up if we
7786 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
7787 ;; a knr region) or BOB.
7788 ;;
7789 ;; The criterion for a paren structure being the arg list is:
7790 ;; o - there is non-WS stuff after it but before any "{"; AND
7791 ;; o - the token after it isn't a ";" AND
7792 ;; o - it is preceded by either an identifier (the function name) or
7793 ;; a macro expansion like "DEFUN (...)"; AND
7794 ;; o - its content is a non-empty comma-separated list of identifiers
7795 ;; (an empty arg list won't have a knr region).
7796 ;;
7797 ;; The following snippet illustrates these rules:
7798 ;; int foo (bar, baz, yuk)
7799 ;; int bar [] ;
7800 ;; int (*baz) (my_type) ;
7801 ;; int (*) (void) (*yuk) (void) ;
7802 ;; {
7813 ; parens/brackets.
7817 ;; We're inside a paren. Could it be our argument list....?
7823 ;; FIXME!!! What about macros between the parens? 2007/01/20
7826 ;; It can't be the arg list if next token is ; or {
7831 ;; Is the thing preceding the list an identifier (the
7832 ;; function name), or a macro expansion?
7842 ;; Have we got a non-empty list of comma-separated
7843 ;; identifiers?
7857 ;; ...Yes. We've identified the function's argument list.
7863 ;; ...No. The current parens aren't the function's arg list.
7870 ;; skip forward over conditional at point, including any predicate
7871 ;; statements in parentheses. No error checking is performed.
7872 ;;
7873 ;; This function might do hidden buffer changes.
7875 ;; else if()
7880 ;; do, else, try, finally
7885 ;; for, if, while, switch, catch, synchronized, foreach
7889 ;; If looking at the token after a conditional then return the
7890 ;; position of its start, otherwise return nil.
7891 ;;
7892 ;; This function might do hidden buffer changes.
7902 ;; If the point is after an operator identifier that isn't handled
7903 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
7904 ;; position of the start of that identifier is returned. nil is
7905 ;; returned otherwise. The point may be anywhere in the syntactic
7906 ;; whitespace after the last token of the operator identifier.
7907 ;;
7908 ;; This function might do hidden buffer changes.
7920 ;; Assuming point is at a brace that opens a statement block of some
7921 ;; kind, move to the proper anchor point for that block. It might
7922 ;; need to be adjusted further by c-add-stmt-syntax, but the
7923 ;; position at return is suitable as start position for that
7924 ;; function.
7925 ;;
7926 ;; This function might do hidden buffer changes.
7933 ;; Assuming point is at a brace that opens the block of a top level
7934 ;; declaration of some kind, move to the proper anchor point for
7935 ;; that block.
7936 ;;
7937 ;; This function might do hidden buffer changes.
7942 ;; Search forward for the end of the "header" of the current
7943 ;; declaration. That's the position where the definition body
7944 ;; starts, or the first variable initializer, or the ending
7945 ;; semicolon. I.e. search forward for the closest following
7946 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
7947 ;; _after_ the first found token, or at point-max if none is found.
7948 ;;
7949 ;; This function might do hidden buffer changes.
7954 ;; In C++ we need to take special care to handle operator
7955 ;; tokens and those pesky template brackets.
7960 ;; Handle operator identifiers, i.e. ignore any
7961 ;; operator token preceded by "operator".
7968 t
7970 nil)))))
7979 ;; Go to the beginning of the current declaration, or the beginning
7980 ;; of the previous one if already at the start of it. Point won't
7981 ;; be moved out of any surrounding paren. Return a cons cell of the
7982 ;; form (MOVE . KNR-POS). MOVE is like the return value from
7983 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
7984 ;; style argument declarations (and they are to be recognized) then
7985 ;; KNR-POS is set to the start of the first such argument
7986 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
7987 ;; position that bounds the backward search.
7988 ;;
7989 ;; NB: Cases where the declaration continues after the block, as in
7990 ;; "struct foo { ... } bar;", are currently recognized as two
7991 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
7992 ;;
7993 ;; This function might do hidden buffer changes.
7999 ;; `c-beginning-of-statement-1' stops at a block start, but we
8000 ;; want to continue if the block doesn't begin a top level
8001 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8002 ;; or an open paren.
8004 ;; Go back one "statement" each time round the loop until we're just
8005 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8006 ;; an ObjC method. This will move over a multiple declaration whose
8007 ;; components are comma separated.
8009 ;; Must check with c-opt-method-key in ObjC mode.
8019 ;; Check that we don't move from the first thing in a
8020 ;; macro to its header.
8026 move tentative-move))
8031 ;; Handle K&R argdecls. Back up after the "statement" jumped
8032 ;; over by `c-beginning-of-statement-1', unless it was the
8033 ;; function body, in which case we're sitting on the opening
8034 ;; brace now. Then test if we're in a K&R argdecl region and
8035 ;; that we started at the other side of the first argdecl in
8036 ;; it.
8046 'previous
8048 knr-argdecl-start))
8051 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8052 ;; statement, so the result will be 'previous if we've moved over any.
8053 ;; So change our result back to 'same if necessary.
8054 ;;
8055 ;; If they were brace list initializers we might not have moved over a
8056 ;; declaration boundary though, so change it to 'same if we've moved
8057 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8058 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8059 ;; potentially can search over a large amount of text.). Take special
8060 ;; pains not to get mislead by C++'s "operator=", and the like.
8063 c++-template-syntax-table
8069 ; no more, or get to one which isn't an "operator ="
8072 c-overloadable-operators-regexp
8073 c-opt-op-identifier-prefix
8088 ;; Assuming point is at the start of a declaration (as detected by
8089 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8090 ;; `c-beginning-of-decl-1', this function handles the case when a
8091 ;; block is followed by identifiers in e.g. struct declarations in C
8092 ;; or C++. If a proper end was found then t is returned, otherwise
8093 ;; point is moved as far as possible within the current sexp and nil
8094 ;; is returned. This function doesn't handle macros; use
8095 ;; `c-end-of-macro' instead in those cases.
8096 ;;
8097 ;; This function might do hidden buffer changes.
8100 c++-template-syntax-table
8108 ;; Stopped at the ';' in a K&R argdecl section which is
8109 ;; detected using the same criteria as in
8110 ;; `c-beginning-of-decl-1'. Move to the following block
8111 ;; start.
8115 ;; Encountered a block in the declaration. Jump over it.
8126 ;; Check for `c-opt-block-decls-with-vars-key'
8127 ;; before the first paren.
8130 c-opt-block-decls-with-vars-key
8132 lim t t t)
8135 ;; Check that the first following paren is
8136 ;; the block.
8138 lim t t t)
8140 ;; The declaration doesn't have any of the
8141 ;; `c-opt-block-decls-with-vars' keywords in the
8142 ;; beginning, so it ends here at the end of the block.
8150 nil)))
8153 ;; Assuming the point is at an open brace, check if it starts a
8154 ;; block that contains another declaration level, i.e. that isn't a
8155 ;; statement block or a brace list, and if so return non-nil.
8156 ;;
8157 ;; If the check is successful, the return value is the start of the
8158 ;; keyword that tells what kind of construct it is, i.e. typically
8159 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8160 ;; the point will be at the start of the construct, before any
8161 ;; leading specifiers, otherwise it's at the returned position.
8162 ;;
8163 ;; The point is clobbered if the check is unsuccessful.
8164 ;;
8165 ;; CONTAINING-SEXP is the position of the open of the surrounding
8166 ;; paren, or nil if none.
8167 ;;
8168 ;; The optional LIMIT limits the backward search for the start of
8169 ;; the construct. It's assumed to be at a syntactically relevant
8170 ;; position.
8171 ;;
8172 ;; If any template arglists are found in the searched region before
8173 ;; the open brace, they get marked with paren syntax.
8174 ;;
8175 ;; This function might do hidden buffer changes.
8182 ;; Could be at the end of a template arglist.
8193 ;; Note: Can't get bogus hits inside template arglists below since they
8194 ;; have gotten paren syntax above.
8196 ;; If `goto-start' is set we begin by searching for the
8197 ;; first possible position of a leading specifier list.
8198 ;; The `c-decl-block-key' search continues from there since
8199 ;; we know it can't match earlier.
8202 open-brace t t)
8204 t)
8205 t)
8212 ;; Found a keyword that can't be a type?
8215 ;; Can be a type too, in which case it's the return type of a
8216 ;; function (under the assumption that no declaration level
8217 ;; block construct starts with a type).
8220 ;; Jumped over a type, but it could be a declaration keyword
8221 ;; followed by the declared identifier that we've jumped over
8222 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8223 ;; then we should be at the declarator now, so check for a
8224 ;; valid declarator start.
8225 ;;
8226 ;; Note: This doesn't cope with the case when a declared
8227 ;; identifier is followed by e.g. '(' in a language where '('
8228 ;; also might be part of a declarator expression. Currently
8229 ;; there's no such language.
8233 ;; In Pike a list of modifiers may be followed by a brace
8234 ;; to make them apply to many identifiers. Note that the
8235 ;; match data will be empty on return in this case.
8241 ;; Use this variant to avoid yet another special regexp.
8245 t)))
8247 ;; Got a match.
8250 ;; Back up over any preceding specifiers and their clauses
8251 ;; by going forward from `first-specifier-pos', which is the
8252 ;; earliest possible position where the specifier list can
8253 ;; start.
8259 ;; Accept any plain symbol token on the ground that
8260 ;; it's a specifier masked through a macro (just
8261 ;; like `c-forward-decl-or-cast-1' skip forward over
8262 ;; such tokens).
8263 ;;
8264 ;; Could be more restrictive wrt invalid keywords,
8265 ;; but that'd only occur in invalid code so there's
8266 ;; no use spending effort on it.
8272 ;; Can't parse a declaration preamble and is still
8273 ;; before `kwd-start'. That means `first-specifier-pos'
8274 ;; was in some earlier construct. Search again.
8278 ;; Got no preamble before the block declaration keyword.
8284 kwd-start)))
8287 ;; Check if the closest containing paren sexp is a declaration
8288 ;; block, returning a 2 element vector in that case. Aref 0
8289 ;; contains the bufpos at boi of the class key line, and aref 1
8290 ;; contains the bufpos of the open brace. This function is an
8291 ;; obsolete wrapper for `c-looking-at-decl-block'.
8292 ;;
8293 ;; This function might do hidden buffer changes.
8301 nil))
8306 ;; Pull the next open brace from PS (which has the form of paren-state),
8307 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
8315 ;; Return the buffer position of the most enclosing decl-block brace (in the
8316 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8317 ;; none was found.
8326 open-brace))
8329 ;; Return the position of the L-brace if point is inside a brace list
8330 ;; initialization of an array, etc. This is an approximate function,
8331 ;; designed for speed over accuracy. It will not find every bracelist, but
8332 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8333 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8334 ;; is everywhere else.
8343 b-pos)))
8346 ;; return the buffer position of the beginning of the brace list
8347 ;; statement if we're inside a brace list, otherwise return nil.
8348 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8349 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8350 ;; braces
8351 ;;
8352 ;; N.B.: This algorithm can potentially get confused by cpp macros
8353 ;; placed in inconvenient locations. It's a trade-off we make for
8354 ;; speed.
8355 ;;
8356 ;; This function might do hidden buffer changes.
8358 ;; This will pick up brace list declarations.
8371 ;; this will pick up array/aggregate init lists, even if they are nested.
8374 ;; Pike can have class definitions anywhere, so we must
8375 ;; check for the class key here.
8377 c-decl-block-key))
8378 bufpos braceassignp lim next-containing)
8380 containing-sexp)
8391 ;; We're in an in-expression block of some kind. Do not
8392 ;; check nesting. We deliberately set the limit to the
8393 ;; containing sexp, so that c-looking-at-inexpr-block
8394 ;; doesn't check for an identifier before it.
8396 ;; see if the open brace is preceded by = or [...] in
8397 ;; this statement, but watch out for operator=
8400 ;; Checks to do only on the first sexp before the brace.
8403 ;; In Java, an initialization brace list may follow
8404 ;; directly after "new Foo[]", so check for a "new"
8405 ;; earlier.
8411 ;; Carry on looking if this is an
8412 ;; identifier (may contain "." in Java)
8413 ;; or another "[]" sexp.
8416 ;; Checks to do on all sexps before the brace, up to the
8417 ;; beginning of the statement.
8425 ;; We've seen a =, but must check earlier tokens so
8426 ;; that it isn't something that should be ignored.
8432 ;; Check for operator =
8435 nil)
8436 ;; Check for `<opchar>= in Pike.
8439 ;; Special case for Pikes
8440 ;; `[]=, since '[' is not in
8441 ;; the punctuation class.
8444 nil)
8446 ;; make sure we're not in a C++ template
8447 ;; argument assignment
8459 ))))
8460 nil)
8467 ;; Brace lists can't contain a semicolon, so we're done.
8469 ;; Go up one level.
8471 lim nil
8472 next-containing nil))
8473 ;; we've hit the beginning of the aggregate list
8477 )
8478 bufpos))
8479 ))
8482 ;; If we're looking at the start of a pike-style list, ie `({Â })',
8483 ;; `([Â ])', `(<Â >)' etc, a cons of a cons of its starting and ending
8484 ;; positions and its entry in c-special-brace-lists is returned, nil
8485 ;; otherwise. The ending position is nil if the list is still open.
8486 ;; LIM is the limit for forward search. The point may either be at
8487 ;; the `(' or at the following paren character. Tries to check the
8488 ;; matching closer, but assumes it's correct if no balanced paren is
8489 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8490 ;; a special brace list).
8491 ;;
8492 ;; This function might do hidden buffer changes.
8497 inner-beg end type)
8512 nil)))))
8522 ;; Check balancing of the inner paren
8523 ;; below.
8526 t)
8527 ;; If the inner char isn't a paren then
8528 ;; we can't check balancing, so just
8529 ;; check the char before the outer
8530 ;; closing paren.
8545 ;; Return non-nil if between two statements or declarations, assuming
8546 ;; point is not inside a literal or comment.
8547 ;;
8548 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8549 ;; are recommended instead.
8550 ;;
8551 ;; This function might do hidden buffer changes.
8556 ;; Return non-nil if we're looking at the beginning of a block
8557 ;; inside an expression. The value returned is actually a cons of
8558 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8559 ;; position of the beginning of the construct.
8560 ;;
8561 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8562 ;; position of the closest containing list. If it's nil, the
8563 ;; containing paren isn't used to decide whether we're inside an
8564 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8565 ;; needs to be farther back.
8566 ;;
8567 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8568 ;; brace block might be done. It should only be used when the
8569 ;; construct can be assumed to be complete, i.e. when the original
8570 ;; starting position was further down than that.
8571 ;;
8572 ;; This function might do hidden buffer changes.
8577 ;; Look at the character after point only as a last resort
8578 ;; when we can't disambiguate.
8598 ;; If the class definition is at the start of
8599 ;; a statement, we don't consider it an
8600 ;; in-expression class.
8608 ;; Also, in Pike we treat it as an
8609 ;; in-expression class if it's used in an
8610 ;; object clone expression.
8626 nil)
8634 ;; Accept several square bracket sexps for
8635 ;; Java array initializations.
8643 block-follows
8644 containing-sexp
8653 nil
8656 res))))
8659 ;; Returns non-nil if we're looking at the end of an in-expression
8660 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
8661 ;; PAREN-STATE is the paren state relevant at the current position.
8662 ;;
8663 ;; This function might do hidden buffer changes.
8665 ;; We currently only recognize a block.
8668 containing-sexp)
8677 paren-state)
8678 containing-sexp)))))
8681 ;; Is there a "virtual semicolon" at POS or point?
8682 ;; (See cc-defs.el for full details of "virtual semicolons".)
8683 ;;
8684 ;; This is true when point is at the last non syntactic WS position on the
8685 ;; line, there is a macro call last on the line, and this particular macro's
8686 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
8687 ;; semicolon.
8695 c-macro-with-semi-re
8698 ;; Check we've got nothing after this except comments and empty lines
8699 ;; joined by escaped EOLs.
8712 nil)
8720 ;; Check for one of the listed macros being before point.
8724 t))
8732 \f
8733 ;; `c-guess-basic-syntax' and the functions that precedes it below
8734 ;; implements the main decision tree for determining the syntactic
8735 ;; analysis of the current line of code.
8737 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
8738 ;; auto newline analysis.
8742 ;; BRACEPOS is the position of a brace in a construct like "namespace
8743 ;; Bar {". Return the anchor point in this construct; this is the
8744 ;; earliest symbol on the brace's line which isn't earlier than
8745 ;; "namespace".
8746 ;;
8747 ;; Currently (2007-08-17), "like namespace" means "matches
8748 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
8749 ;; or anything like that.
8759 ;; A simple function to prepend a new syntax element to
8760 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
8761 ;; should always be dynamically bound but since we read it first
8762 ;; we'll fail properly anyway if this function is misused.
8764 c-syntactic-context)))
8767 ;; Like `c-add-syntax' but appends to the end of the syntax list.
8768 ;; (Normally not necessary.)
8773 syntax-extra-args
8774 stop-at-boi-only
8775 containing-sexp
8776 paren-state)
8777 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
8778 ;; needed with further syntax elements of the types `substatement',
8779 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
8780 ;; `defun-block-intro'.
8781 ;;
8782 ;; Do the generic processing to anchor the given syntax symbol on
8783 ;; the preceding statement: Skip over any labels and containing
8784 ;; statements on the same line, and then search backward until we
8785 ;; find a statement or block start that begins at boi without a
8786 ;; label or comment.
8787 ;;
8788 ;; Point is assumed to be at the prospective anchor point for the
8789 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
8790 ;; skip past open parens and containing statements. Most of the added
8791 ;; syntax elements will get the same anchor point - the exception is
8792 ;; for an anchor in a construct like "namespace"[*] - this is as early
8793 ;; as possible in the construct but on the same line as the {.
8794 ;;
8795 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
8796 ;;
8797 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
8798 ;; syntax symbol. They are appended after the anchor point.
8799 ;;
8800 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
8801 ;; if the current statement starts there.
8802 ;;
8803 ;; Note: It's not a problem if PAREN-STATE "overshoots"
8804 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
8805 ;;
8806 ;; This function might do hidden buffer changes.
8809 ;; This is by far the most common case, so let's give it special
8810 ;; treatment.
8815 ;; Set when we're on a label, so that we don't stop there.
8816 ;; FIXME: To be complete we should check if we're on a label
8817 ;; now at the start.
8818 on-label)
8820 ;; Use point as the anchor point for "namespace", "extern", etc.
8824 syntax-extra-args)
8826 ;; Loop while we have to back out of containing blocks.
8831 ;; Loop while we have to back up statements.
8833 on-label
8836 ;; Skip past any comments that stands between the
8837 ;; statement start and boi.
8845 ;; Skip to the beginning of this statement or backward
8846 ;; another one.
8854 ;; If we didn't move we're at the start of a block and
8855 ;; have to continue outside it.
8864 ;; Special case to avoid deeper and deeper indentation
8865 ;; of "else if" clauses.
8866 )
8871 ;; If stop-at-boi-only is nil, we shouldn't back up
8872 ;; over previous or containing statements to try to
8873 ;; reach boi, so go back to the last position and
8874 ;; exit.
8881 ;; If we've moved into another statement then we
8882 ;; should no longer try to stop in the middle of a
8883 ;; line.
8886 ;; Record this as a substatement if we skipped up one
8887 ;; level.
8890 )))
8892 containing-sexp)
8894 ;; Now we have to go out of this block.
8897 ;; Don't stop in the middle of a special brace list opener
8898 ;; like "({".
8910 ;; Analyze the construct in front of the block we've stepped out
8911 ;; from and add the right syntactic element for it.
8914 step-type)
8917 ;; Stepped out of a parenthesis block, so we're in an
8918 ;; expression now.
8930 ;; Stepped out of an in-expression statement. This
8931 ;; syntactic element won't get an anchor pos.
8934 ;; A parenthesis normally belongs to an arglist.
8942 on-label nil))
8944 ;; Stepped out of a brace block.
8956 containing-sexp)))
8958 'defun-block-intro
8960 nil))
8964 c-other-decl-block-key-in-symbols-alist))
8971 ;; Always done if the open brace was at boi. The
8972 ;; c-beginning-of-statement-1 call above is necessary
8973 ;; anyway, to decide the type of block-intro to add.
8976 ))
8978 ;; Fill in the current point as the anchor for all the symbols
8979 ;; added above.
8988 )))
8991 containing-decl-open
8992 containing-decl-start
8993 containing-decl-kwd
8994 paren-state)
8995 ;; The inclass and class-close syntactic symbols are added in
8996 ;; several places and some work is needed to fix everything.
8997 ;; Therefore it's collected here.
8998 ;;
8999 ;; This function might do hidden buffer changes.
9004 containing-decl-open)
9006 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9007 ;; here, but we have to do like this for compatibility.
9017 char-after-ip
9018 beg-of-same-or-containing-stmt
9019 containing-sexp
9020 paren-state)
9021 ;; This function contains the decision tree reached through both
9022 ;; cases 18 and 10. It's a continued statement or top level
9023 ;; construct of some kind.
9024 ;;
9025 ;; This function might do hidden buffer changes.
9032 ;; (CASE A removed.)
9033 ;; CASE B: open braces for class or brace-lists
9040 ;; CASE B.1: class-open
9047 ;; CASE B.2: brace-list-open
9052 indent-point t t t)))
9053 ;; The most semantically accurate symbol here is
9054 ;; brace-list-open, but we normally report it simply as a
9055 ;; statement-cont. The reason is that one normally adjusts
9056 ;; brace-list-open for brace lists as top-level constructs,
9057 ;; and brace lists inside statements is a completely different
9058 ;; context. C.f. case 5A.3.
9061 ;; Turn off the dwim above when we're
9062 ;; analyzing the nature of the brace
9063 ;; for the auto newline feature.
9064 'brace-list-open
9066 nil nil
9067 containing-sexp paren-state))
9069 ;; CASE B.3: The body of a function declared inside a normal
9070 ;; block. Can occur e.g. in Pike and when using gcc
9071 ;; extensions, but watch out for macros followed by blocks.
9072 ;; C.f. cases E, 16F and 17G.
9078 ;; Turn off recognition of constructs that lacks a
9079 ;; type in this case, since that's more likely to be
9080 ;; a macro followed by a block.
9083 containing-sexp paren-state))
9085 ;; CASE B.4: Continued statement with block open. The most
9086 ;; accurate analysis is perhaps `statement-cont' together with
9087 ;; `block-open' but we play DWIM and use `substatement-open'
9088 ;; instead. The rationale is that this typically is a macro
9089 ;; followed by a block which makes it very similar to a
9090 ;; statement with a substatement block.
9093 containing-sexp paren-state))
9094 ))
9096 ;; CASE C: iostream insertion or extraction operator
9100 ;; If there is no preceding streamop in the statement
9101 ;; then indent this line as a normal statement-cont.
9105 t))))
9107 ;; CASE E: In the "K&R region" of a function declared inside a
9108 ;; normal block. C.f. case B.3.
9110 ;; Check that the next token is a '{'. This works as
9111 ;; long as no language that allows nested function
9112 ;; definitions allows stuff like member init lists, K&R
9113 ;; declarations or throws clauses there.
9114 ;;
9115 ;; Note that we do a forward search for something ahead
9116 ;; of the indentation line here. That's not good since
9117 ;; the user might not have typed it yet. Unfortunately
9118 ;; it's exceedingly tricky to recognize a function
9119 ;; prototype in a code block without resorting to this.
9127 ;; Turn off recognition of constructs that lacks a
9128 ;; type in this case, since that's more likely to be
9129 ;; a macro followed by a block.
9132 containing-sexp paren-state))
9134 ;;CASE F: continued statement and the only preceding items are
9135 ;;annotations.
9143 t)
9150 ;; CASE G: a template list continuation?
9151 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9169 ;; In a nested template arglist.
9175 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9176 ;; template aware.
9179 ;; CASE D: continued statement.
9183 containing-sexp paren-state))
9184 )))
9186 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9187 ;; 2005/11/29).
9188 ;;;###autoload
9190 "Return the syntactic context of the current line."
9196 ;; A whole ugly bunch of various temporary variables. Have
9197 ;; to declare them here since it's not possible to declare
9198 ;; a variable with only the scope of a cond test and the
9199 ;; following result clauses, and most of this function is a
9200 ;; single gigantic cond. :P
9201 literal char-before-ip before-ws-ip char-after-ip macro-start
9202 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9203 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9204 containing-<
9205 ;; The following record some positions for the containing
9206 ;; declaration block if we're directly within one:
9207 ;; `containing-decl-open' is the position of the open
9208 ;; brace. `containing-decl-start' is the start of the
9209 ;; declaration. `containing-decl-kwd' is the keyword
9210 ;; symbol of the keyword that tells what kind of block it
9211 ;; is.
9212 containing-decl-open
9213 containing-decl-start
9214 containing-decl-kwd
9215 ;; The open paren of the closest surrounding sexp or nil if
9216 ;; there is none.
9217 containing-sexp
9218 ;; The position after the closest preceding brace sexp
9219 ;; (nested sexps are ignored), or the position after
9220 ;; `containing-sexp' if there is none, or (point-min) if
9221 ;; `containing-sexp' is nil.
9222 lim
9223 ;; The paren state outside `containing-sexp', or at
9224 ;; `indent-point' if `containing-sexp' is nil.
9226 ;; There's always at most one syntactic element which got
9227 ;; an anchor pos. It's stored in syntactic-relpos.
9228 syntactic-relpos
9231 ;; Check if we're directly inside an enclosing declaration
9232 ;; level block.
9241 containing-sexp)
9242 t)))
9245 containing-sexp nil)
9250 ;; Init some position variables.
9259 ;; Ignore balanced paren. The next entry
9260 ;; can't be another one.
9263 ;; If there is no surrounding open paren then
9264 ;; put the last balanced pair back on paren-state.
9266 containing-sexp nil)))
9274 ;; If we're in a parenthesis list then ',' delimits the
9275 ;; "statements" rather than being an operator (with the
9276 ;; exception of the "for" clause). This difference is
9277 ;; typically only noticeable when statements are used in macro
9278 ;; arglists.
9282 ;; cache char before and after indent point, and move point to
9283 ;; the most likely position to perform the majority of tests
9292 ;; are we in a literal?
9295 ;; now figure out syntactic qualities of the current line
9298 ;; CASE 1: in a string.
9302 ;; CASE 2: in a C or C++ style comment.
9304 ;; This is a kludge for XEmacs where we use
9305 ;; `buffer-syntactic-context', which doesn't correctly
9306 ;; recognize "\*/" to end a block comment.
9307 ;; `parse-partial-sexp' which is used by
9308 ;; `c-literal-limits' will however do that in most
9309 ;; versions, which results in that we get nil from
9310 ;; `c-literal-limits' even when `c-in-literal' claims
9311 ;; we're inside a comment.
9315 ;; CASE 3: in a cpp preprocessor macro continuation.
9325 ;; If at the beginning of the body of a #define
9326 ;; directive then analyze as cpp-define-intro
9327 ;; only. Go on with the syntactic analysis
9328 ;; otherwise. in-macro-expr is set if we're in a
9329 ;; cpp expression, i.e. before the #define body
9330 ;; or anywhere in a non-#define directive.
9337 nil)))))
9341 ;; CASE 11: an else clause?
9345 containing-sexp paren-state))
9347 ;; CASE 12: while closure of a do/while construct?
9355 containing-sexp paren-state))
9357 ;; CASE 13: A catch or finally clause? This case is simpler
9358 ;; than if-else and do-while, because a block is required
9359 ;; after every try, catch and finally.
9367 t)
9371 t)
9374 t))
9379 containing-sexp paren-state))
9381 ;; CASE 18: A substatement we can recognize by keyword.
9391 ;; Ought to cache the result from the
9392 ;; c-beginning-of-statement-1 calls here.
9402 ;; Step up to the containing statement if we
9403 ;; stayed in the same one.
9411 ;; There was no containing statement after all.
9413 placeholder))
9415 ;; Require a parenthesis after these keywords.
9416 ;; Necessary to catch e.g. synchronized in Java,
9417 ;; which can be used both as statement and
9418 ;; modifier.
9424 ;; CASE 18A: Simple substatement.
9430 containing-sexp paren-state))
9436 containing-sexp paren-state))
9439 containing-sexp paren-state))))
9441 ;; CASE 18B: Some other substatement. This is shared
9442 ;; with case 10.
9444 char-after-ip
9445 placeholder
9446 lim
9447 paren-state)))
9449 ;; CASE 14: A case or default label
9455 containing-sexp))
9458 ;; Got a bogus label at the top level. In lack of better
9459 ;; alternatives, anchor it on (point-min).
9462 ;; CASE 15: any other label
9469 containing-decl-open
9470 containing-decl-start
9471 containing-decl-kwd
9472 paren-state))
9473 ;; Append access-label with the same anchor point as
9474 ;; inclass gets.
9480 containing-sexp))
9485 ;; If the surrounding statement is a switch then
9486 ;; let's analyze all labels as switch labels, so
9487 ;; that they get lined up consistently.
9488 'case-label
9494 ;; A label on the top level. Treat it as a class
9495 ;; context. (point-min) is the closest we get to the
9496 ;; class open brace.
9499 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9500 ;; 17E.
9503 containing-sexp
9504 ;; Have to turn on the heuristics after
9505 ;; the point even though it doesn't work
9506 ;; very well. C.f. test case class-16.pike.
9507 t))
9512 ;; It's a statement block or an anonymous class.
9514 ;; It's a Pike lambda. Check whether we are between the
9515 ;; lambda keyword and the argument list or at the defun
9516 ;; opener.
9518 'inline-open
9524 paren-state)
9528 ;; CASE 5: Line is inside a declaration level block or at top level.
9532 ;; CASE 5A: we are looking at a defun, brace list, class,
9533 ;; or inline-inclass method opening brace
9540 ;; CASE 5A.1: Non-class declaration block open.
9555 ;; Special case for extern-lang-open.
9556 'extern-lang-open
9558 nil t containing-sexp paren-state))
9560 ;; CASE 5A.2: we are looking at a class opening brace
9569 ;; CASE 5A.3: brace list open
9594 ))))
9598 ;; We're in Java and have found that the open brace
9599 ;; belongs to a "new Foo[]" initialization list,
9600 ;; which means the brace list is part of an
9601 ;; expression and not a top level definition. We
9602 ;; therefore treat it as any topmost continuation
9603 ;; even though the semantically correct symbol still
9604 ;; is brace-list-open, on the same grounds as in
9605 ;; case B.2.
9611 ;; CASE 5A.4: inline defun open
9617 containing-decl-open
9618 containing-decl-start
9619 containing-decl-kwd
9620 paren-state))
9622 ;; CASE 5A.5: ordinary defun open
9630 ;; Bogus to use bol here, but it's the legacy. (Resolved,
9631 ;; 2007-11-09)
9632 ))))
9634 ;; CASE 5B: After a function header but before the body (or
9635 ;; the ending semicolon if there's no body).
9642 ;; CASE 5B.1: Member init list.
9647 ;; There is no preceding member init clause.
9648 ;; Indent relative to the beginning of indentation
9649 ;; for the topmost-intro line that contains the
9650 ;; prototype's open paren.
9653 ;; Indent relative to the first member init clause.
9658 ;; CASE 5B.2: K&R arg decl intro
9665 containing-decl-open
9666 containing-decl-start
9667 containing-decl-kwd
9668 paren-state)))
9670 ;; CASE 5B.4: Nether region after a C++ or Java func
9671 ;; decl, which could include a `throws' declaration.
9675 )))
9677 ;; CASE 5C: inheritance line. could be first inheritance
9678 ;; line, or continuation of a multiple inheritance
9686 ;; watch out for scope operator
9691 )))
9700 ;; for Java
9705 cont done)
9710 done t))
9714 )
9716 injava-inher)
9719 ))
9722 ;; CASE 5C.1: non-hanging colon on an inher intro
9726 ;; don't add inclass symbol since relative point already
9727 ;; contains any class offset
9728 )
9730 ;; CASE 5C.2: hanging colon on an inher intro
9736 containing-decl-open
9737 containing-decl-start
9738 containing-decl-kwd
9739 paren-state)))
9741 ;; CASE 5C.3: in a Java implements/extends
9755 )))
9757 ;; CASE 5C.4: a continued inheritance line
9761 ;; don't add inclass symbol since relative point already
9762 ;; contains any class offset
9763 )))
9765 ;; CASE 5D: this could be a top-level initialization, a
9766 ;; member init list continuation, or a template argument
9767 ;; list continuation.
9769 ;; Note: We use the fact that lim is always after any
9770 ;; preceding brace sexp.
9780 t))
9785 t)
9791 nil)
9792 t))
9794 ;; NB: No c-after-special-operator-id stuff in this
9795 ;; clause - we assume only C++ needs it.
9800 ;; CASE 5D.3: perhaps a template list continuation?
9815 ;; In a nested template arglist.
9821 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9822 ;; template aware.
9825 ;; CASE 5D.4: perhaps a multiple inheritance line?
9837 t)
9842 ;; CASE 5D.5: Continuation of the "expression part" of a
9843 ;; top level construct. Or, perhaps, an unrecognized construct.
9855 ;; A preceding comma at the top level means that a
9856 ;; new variable declaration starts here. Use
9857 ;; topmost-intro-cont for it, for consistency with
9858 ;; the first variable declaration. C.f. case 5N.
9861 nil nil containing-sexp paren-state))
9862 ))
9864 ;; CASE 5F: Close of a non-class declaration level block.
9868 ;; This is inconsistent: Should use `containing-decl-open'
9869 ;; here if it's at boi, like in case 5J.
9873 ;; Special case for compatibility with the
9874 ;; extern-lang syntactic symbols.
9875 'extern-lang-close
9878 nil t
9880 paren-state))
9882 ;; CASE 5G: we are looking at the brace which closes the
9883 ;; enclosing nested class decl
9888 containing-decl-open
9889 containing-decl-start
9890 containing-decl-kwd
9891 paren-state))
9893 ;; CASE 5H: we could be looking at subsequent knr-argdecls
9900 ;; Do an extra check to avoid tripping up on
9901 ;; statements that occur in invalid contexts
9902 ;; (e.g. in macro bodies where we don't really
9903 ;; know the context of what we're looking at).
9910 ;; CASE 5I: ObjC method definition.
9915 ;; Handle the case when it's the first (non-comment)
9916 ;; thing in the buffer. Can't look for a 'same return
9917 ;; value from cbos1 since ObjC directives currently
9918 ;; aren't recognized fully, so that we get 'same
9919 ;; instead of 'previous if it moved over a preceding
9920 ;; directive.
9924 ;; CASE 5P: AWK pattern or function or continuation
9925 ;; thereof.
9931 'topmost-intro-cont
9933 nil nil
9934 containing-sexp paren-state))
9936 ;; CASE 5N: At a variable declaration that follows a class
9937 ;; definition or some other block declaration that doesn't
9938 ;; end at the closing '}'. C.f. case 5D.5.
9947 ;; Speed up the backward search a bit.
9952 ;; The '}' is unbalanced.
9953 nil
9958 containing-sexp paren-state))
9960 ;; NOTE: The point is at the end of the previous token here.
9962 ;; CASE 5J: we are at the topmost level, make
9963 ;; sure we skip back past any access specifiers
9965 ;; A macro continuation line is never at top level.
9987 nil)))))
9988 ;; For historic reasons we anchor at bol of the last
9989 ;; line of the previous declaration. That's clearly
9990 ;; highly bogus and useless, and it makes our lives hard
9991 ;; to remain compatible. :P
10002 ;; Special case for compatibility with the
10003 ;; extern-lang syntactic symbols.
10004 'inextern-lang
10007 nil t
10009 paren-state))
10011 containing-decl-open
10012 containing-decl-start
10013 containing-decl-kwd
10014 paren-state)))
10016 macro-start
10021 ;; CASE 5K: we are at an ObjC method definition
10022 ;; continuation line.
10031 ;; CASE 5L: we are at the first argument of a template
10032 ;; arglist that begins on the previous line.
10040 ;; CASE 5Q: we are at a statement within a macro.
10045 ;;CASE 5N: We are at a topmost continuation line and the only
10046 ;;preceding items are annotations.
10053 t)
10059 ;; CASE 5M: we are at a topmost continuation line
10070 ))
10073 ;; (CASE 6 has been removed.)
10075 ;; CASE 7: line is an expression, not a statement. Most
10076 ;; likely we are either in a function prototype or a function
10077 ;; call argument list
10085 ;; CASE 7A: we are looking at the arglist closing paren.
10086 ;; C.f. case 7F.
10098 paren-state))
10100 ;; CASE 7B: Looking at the opening brace of an
10101 ;; in-expression block or brace list. C.f. cases 4, 16A
10102 ;; and 17E.
10106 paren-state))
10110 placeholder
10113 paren-state)
10114 containing-sexp)))
10115 ;; placeholder is nil if it's a block directly in
10116 ;; a function arglist. That makes us skip out of
10117 ;; this case.
10118 )))
10123 paren-state)
10127 ;; CASE 7C: we are looking at the first argument in an empty
10128 ;; argument list. Use arglist-close if we're actually
10129 ;; looking at a close paren or bracket.
10141 paren-state))
10143 ;; CASE 7D: we are inside a conditional test clause. treat
10144 ;; these things as statements
10154 ))
10156 ;; CASE 7E: maybe a continued ObjC method call. This is the
10157 ;; case when we are inside a [] bracketed exp, and what
10158 ;; precede the opening bracket is not an identifier.
10166 )))
10168 ;; CASE 7F: we are looking at an arglist continuation line,
10169 ;; but the preceding argument is on the same line as the
10170 ;; opening paren. This case includes multi-line
10171 ;; mathematical paren groupings, but we could be on a
10172 ;; for-list continuation line. C.f. case 7A.
10189 paren-state))
10191 ;; CASE 7G: we are looking at just a normal arglist
10192 ;; continuation line
10195 ))
10197 ;; CASE 8: func-local multi-inheritance line
10207 ;; CASE 8A: non-hanging colon on an inher intro
10212 ;; CASE 8B: hanging colon on an inher intro
10216 ;; CASE 8C: a continued inheritance line
10220 )))
10222 ;; CASE 9: we are inside a brace-list
10232 ;; CASE 9A: In the middle of a special brace list opener.
10252 ;; CASE 9B: brace-list-close brace
10254 ;; Check special brace list closer.
10261 ;; We were between the special close char and the `)'.
10264 ;; We were before the special close char.
10268 ;; Normal brace list check.
10279 ;; Prepare for the rest of the cases below by going to the
10280 ;; token following the opening brace
10293 ;; CASE 9C: we're looking at the first line in a brace-list
10304 ;; CASE 9D: this is just a later brace-list-entry or
10305 ;; brace-entry-open
10314 ))
10315 ))))
10317 ;; CASE 10: A continued statement or top level construct.
10327 ;; This is shared with case 18.
10329 char-after-ip
10330 placeholder
10331 containing-sexp
10332 paren-state))
10334 ;; CASE 16: block close brace, possibly closing the defun or
10335 ;; the class
10337 ;; From here on we have the next containing sexp in lim.
10342 ;; CASE 16E: Closing a statement block? This catches
10343 ;; cases where it's preceded by a statement keyword,
10344 ;; which works even when used in an "invalid" context,
10345 ;; e.g. a macro argument.
10350 ;; CASE 16A: closing a lambda defun or an in-expression
10351 ;; block? C.f. cases 4, 7B and 17E.
10354 nil))
10356 'inline-close
10366 paren-state)
10370 ;; CASE 16B: does this close an inline or a function in
10371 ;; a non-class declaration level block?
10378 nil))
10388 ;; CASE 16F: Can be a defun-close of a function declared
10389 ;; in a statement block, e.g. in Pike or when using gcc
10390 ;; extensions, but watch out for macros followed by
10391 ;; blocks. Let it through to be handled below.
10392 ;; C.f. cases B.3 and 17G.
10398 ;; Turn off recognition of constructs that
10399 ;; lacks a type in this case, since that's more
10400 ;; likely to be a macro followed by a block.
10407 ;; CASE 16C: If there is an enclosing brace then this is
10408 ;; a block close since defun closes inside declaration
10409 ;; level blocks have been handled above.
10411 ;; If the block is preceded by a case/switch label on
10412 ;; the same line, we anchor at the first preceding label
10413 ;; at boi. The default handling in c-add-stmt-syntax
10414 ;; really fixes it better, but we do like this to keep
10415 ;; the indentation compatible with version 5.28 and
10416 ;; earlier. C.f. case 17H.
10423 ;; c-backward-to-block-anchor not necessary here; those
10424 ;; situations are handled in case 16E above.
10427 ;; CASE 16D: Only top level defun close left.
10433 paren-state))
10434 ))
10436 ;; CASE 19: line is an expression, not a statement, and is directly
10437 ;; contained by a template delimiter. Most likely, we are in a
10438 ;; template arglist within a statement. This case is based on CASE
10439 ;; 7. At some point in the future, we may wish to create more
10440 ;; syntactic symbols such as `template-intro',
10441 ;; `template-cont-nonempty', etc., and distinguish between them as we
10442 ;; do for `arglist-intro' etc. (2009-12-07).
10448 ; '<') before indent-point.
10456 paren-state))
10458 ;; CASE 17: Statement or defun catchall.
10461 ;; Back up statements until we find one that starts at boi.
10464 containing-sexp)))
10468 nil)
10473 ;; CASE 17B: continued statement
10477 containing-sexp paren-state))
10479 ;; CASE 17A: After a case/default label?
10487 'statement-case-open
10489 nil t containing-sexp paren-state))
10491 ;; CASE 17D: any old statement
10498 containing-sexp paren-state)
10502 ;; CASE 17I: Inside a substatement block.
10504 ;; The following tests are all based on containing-sexp.
10506 ;; From here on we have the next containing sexp in lim.
10511 lim paren-state)
10515 ;; CASE 17E: first statement in an in-expression block.
10516 ;; C.f. cases 4, 7B and 16A.
10519 nil))
10521 'defun-block-intro
10530 paren-state)
10536 ;; CASE 17F: first statement in an inline, or first
10537 ;; statement in a top-level defun. we can tell this is it
10538 ;; if there are no enclosing braces that haven't been
10539 ;; narrowed out by a class (i.e. don't use bod here).
10542 paren-state)))
10548 nil))))
10553 ;; CASE 17G: First statement in a function declared inside
10554 ;; a normal block. This can occur in Pike and with
10555 ;; e.g. the gcc extensions, but watch out for macros
10556 ;; followed by blocks. C.f. cases B.3 and 16F.
10562 ;; Turn off recognition of constructs that lacks
10563 ;; a type in this case, since that's more likely
10564 ;; to be a macro followed by a block.
10570 lim paren-state))
10572 ;; CASE 17H: First statement in a block.
10574 ;; If the block is preceded by a case/switch label on the
10575 ;; same line, we anchor at the first preceding label at
10576 ;; boi. The default handling in c-add-stmt-syntax is
10577 ;; really fixes it better, but we do like this to keep the
10578 ;; indentation compatible with version 5.28 and earlier.
10579 ;; C.f. case 16C.
10586 ;; c-backward-to-block-anchor not necessary here; those
10587 ;; situations are handled in case 17I above.
10589 lim paren-state))
10592 ))
10593 )
10595 ;; now we need to look at any modifiers
10599 ;; are we looking at a comment only line?
10604 ;; we might want to give additional offset to friends (in C++).
10609 ;; Set syntactic-relpos.
10615 nil)
10616 t))
10619 ;; Start of or a continuation of a preprocessor directive?
10634 ;; If inside a cpp expression, i.e. anywhere in a
10635 ;; cpp directive except a #define body, we only let
10636 ;; through the syntactic analysis that is internal
10637 ;; in the expression. That means the arglist
10638 ;; elements, if they are anchored inside the cpp
10639 ;; expression.
10648 ;; Inside a #define body and the syntactic analysis is
10649 ;; anchored on the start of the #define. In this case
10650 ;; we add cpp-define-intro to get the extra
10651 ;; indentation of the #define body.
10654 ;; return the syntax
10655 c-syntactic-context)))
10657 \f
10658 ;; Indentation calculation.
10661 ;; offset can be a number, a function, a variable, a list, or one of
10662 ;; the symbols + or -
10663 ;;
10664 ;; This function might do hidden buffer changes.
10683 langelem symbol))
10689 offset symbol)
10690 nil)
10704 Error evaluating offset %S for %s: \
10705 Cannot combine absolute offset %S with relative %S in `%s' method"
10711 Error evaluating offset %S for %s: \
10712 Cannot combine relative offset %S with absolute %S in `%s' method"
10717 res))
10735 Error evaluating offset %S for %s: \
10736 Cannot combine absolute offsets %S and %S in `add' method"
10740 res))
10749 res))))
10756 nil))))
10760 res
10762 offset symbol res)
10763 nil)))
10766 ;; Get offset from LANGELEM which is a list beginning with the
10767 ;; syntactic symbol and followed by any analysis data it provides.
10768 ;; That data may be zero or more elements, but if at least one is
10769 ;; given then the first is the anchor position (or nil). The symbol
10770 ;; is matched against `c-offsets-alist' and the offset calculated
10771 ;; from that is returned.
10772 ;;
10773 ;; This function might do hidden buffer changes.
10783 offset
10787 ))
10790 ;; This is a compatibility wrapper for `c-calc-offset' in case
10791 ;; someone is calling it directly. It takes an old style syntactic
10792 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
10793 ;; new list form.
10794 ;;
10795 ;; This function might do hidden buffer changes.
10802 ;; Calculate the syntactic indentation from a syntactic description
10803 ;; as returned by `c-guess-syntax'.
10804 ;;
10805 ;; Note that topmost-intro always has an anchor position at bol, for
10806 ;; historical reasons. It's often used together with other symbols
10807 ;; that has more sane positions. Since we always use the first
10808 ;; found anchor position, we rely on that these other symbols always
10809 ;; precede topmost-intro in the LANGELEMS list.
10810 ;;
10811 ;; This function might do hidden buffer changes.
10819 ;; Got an absolute column that overrides any indentation
10820 ;; we've collected so far, but not the relative
10821 ;; indentation we might get for the nested structures
10822 ;; further down the langelems list.
10826 ;; Got a relative change of the current calculated
10827 ;; indentation.
10830 ;; Use the anchor position from the first syntactic
10831 ;; element with one.
10841 indent)))
10843 \f
10846 ;;; cc-engine.el ends here