| blob | 8283bb38eae91a6208a93eeaafd284148208482e |
1 ;;; cc-engine.el --- core syntax guessing engine for CC mode -*- coding: utf-8 -*-
3 ;; Copyright (C) 1985, 1987, 1992-2014 Free Software Foundation, Inc.
5 ;; Authors: 2001- Alan Mackenzie
6 ;; 1998- Martin Stjernholm
7 ;; 1992-1999 Barry A. Warsaw
8 ;; 1987 Dave Detlefs
9 ;; 1987 Stewart Clamen
10 ;; 1985 Richard M. Stallman
11 ;; Maintainer: bug-cc-mode@gnu.org
12 ;; Created: 22-Apr-1997 (split from cc-mode.el)
13 ;; Keywords: c languages
14 ;; Package: cc-mode
16 ;; This file is part of GNU Emacs.
18 ;; GNU Emacs is free software: you can redistribute it and/or modify
19 ;; it under the terms of the GNU General Public License as published by
20 ;; the Free Software Foundation, either version 3 of the License, or
21 ;; (at your option) any later version.
23 ;; GNU Emacs is distributed in the hope that it will be useful,
24 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
25 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
26 ;; GNU General Public License for more details.
28 ;; You should have received a copy of the GNU General Public License
29 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
31 ;;; Commentary:
33 ;; The functions which have docstring documentation can be considered
34 ;; part of an API which other packages can use in CC Mode buffers.
35 ;; Otoh, undocumented functions and functions with the documentation
36 ;; in comments are considered purely internal and can change semantics
37 ;; or even disappear in the future.
38 ;;
39 ;; (This policy applies to CC Mode as a whole, not just this file. It
40 ;; probably also applies to many other Emacs packages, but here it's
41 ;; clearly spelled out.)
43 ;; Hidden buffer changes
44 ;;
45 ;; Various functions in CC Mode use text properties for caching and
46 ;; syntactic markup purposes, and those of them that might modify such
47 ;; properties but still don't modify the buffer in a visible way are
48 ;; said to do "hidden buffer changes". They should be used within
49 ;; `c-save-buffer-state' or a similar function that saves and restores
50 ;; buffer modifiedness, disables buffer change hooks, etc.
51 ;;
52 ;; Interactive functions are assumed to not do hidden buffer changes,
53 ;; except in the specific parts of them that do real changes.
54 ;;
55 ;; Lineup functions are assumed to do hidden buffer changes. They
56 ;; must not do real changes, though.
57 ;;
58 ;; All other functions that do hidden buffer changes have that noted
59 ;; in their doc string or comment.
60 ;;
61 ;; The intention with this system is to avoid wrapping every leaf
62 ;; function that do hidden buffer changes inside
63 ;; `c-save-buffer-state'. It should be used as near the top of the
64 ;; interactive functions as possible.
65 ;;
66 ;; Functions called during font locking are allowed to do hidden
67 ;; buffer changes since the font-lock package run them in a context
68 ;; similar to `c-save-buffer-state' (in fact, that function is heavily
69 ;; inspired by `save-buffer-state' in the font-lock package).
71 ;; Use of text properties
72 ;;
73 ;; CC Mode uses several text properties internally to mark up various
74 ;; positions, e.g. to improve speed and to eliminate glitches in
75 ;; interactive refontification.
76 ;;
77 ;; Note: This doc is for internal use only. Other packages should not
78 ;; assume that these text properties are used as described here.
79 ;;
80 ;; 'category
81 ;; Used for "indirection". With its help, some other property can
82 ;; be cheaply and easily switched on or off everywhere it occurs.
83 ;;
84 ;; 'syntax-table
85 ;; Used to modify the syntax of some characters. It is used to
86 ;; mark the "<" and ">" of angle bracket parens with paren syntax, and
87 ;; to "hide" obtrusive characters in preprocessor lines.
88 ;;
89 ;; This property is used on single characters and is therefore
90 ;; always treated as front and rear nonsticky (or start and end open
91 ;; in XEmacs vocabulary). It's therefore installed on
92 ;; `text-property-default-nonsticky' if that variable exists (Emacs
93 ;; >= 21).
94 ;;
95 ;; 'c-is-sws and 'c-in-sws
96 ;; Used by `c-forward-syntactic-ws' and `c-backward-syntactic-ws' to
97 ;; speed them up. See the comment blurb before `c-put-is-sws'
98 ;; below for further details.
99 ;;
100 ;; 'c-type
101 ;; This property is used on single characters to mark positions with
102 ;; special syntactic relevance of various sorts. Its primary use is
103 ;; to avoid glitches when multiline constructs are refontified
104 ;; interactively (on font lock decoration level 3). It's cleared in
105 ;; a region before it's fontified and is then put on relevant chars
106 ;; in that region as they are encountered during the fontification.
107 ;; The value specifies the kind of position:
108 ;;
109 ;; 'c-decl-arg-start
110 ;; Put on the last char of the token preceding each declaration
111 ;; inside a declaration style arglist (typically in a function
112 ;; prototype).
113 ;;
114 ;; 'c-decl-end
115 ;; Put on the last char of the token preceding a declaration.
116 ;; This is used in cases where declaration boundaries can't be
117 ;; recognized simply by looking for a token like ";" or "}".
118 ;; `c-type-decl-end-used' must be set if this is used (see also
119 ;; `c-find-decl-spots').
120 ;;
121 ;; 'c-<>-arg-sep
122 ;; Put on the commas that separate arguments in angle bracket
123 ;; arglists like C++ template arglists.
124 ;;
125 ;; 'c-decl-id-start and 'c-decl-type-start
126 ;; Put on the last char of the token preceding each declarator
127 ;; in the declarator list of a declaration. They are also used
128 ;; between the identifiers cases like enum declarations.
129 ;; 'c-decl-type-start is used when the declarators are types,
130 ;; 'c-decl-id-start otherwise.
131 ;;
132 ;; 'c-awk-NL-prop
133 ;; Used in AWK mode to mark the various kinds of newlines. See
134 ;; cc-awk.el.
136 ;;; Code:
143 load-path)))
150 \f
151 ;; Make declarations for all the `c-lang-defvar' variables in cc-langs.
164 \f
165 ;;; Internal state variables.
167 ;; Internal state of hungry delete key feature
171 ;; The electric flag (toggled by `c-toggle-electric-state').
172 ;; If t, electric actions (like automatic reindentation, and (if
173 ;; c-auto-newline is also set) auto newlining) will happen when an electric
174 ;; key like `{' is pressed (or an electric keyword like `else').
178 ;; Internal state of auto newline feature.
182 ;; Included in the mode line to indicate the active submodes.
183 ;; (defvar c-submode-indicators nil)
184 ;; (make-variable-buffer-local 'c-submode-indicators)
187 ;; Calculate the new state of PREVSTATE, t or nil, based on arg. If
188 ;; arg is nil or zero, toggle the state. If arg is negative, turn
189 ;; the state off, and if arg is positive, turn the state on
195 \f
196 ;; Basic handling of preprocessor directives.
198 ;; This is a dynamically bound cache used together with
199 ;; `c-query-macro-start' and `c-query-and-set-macro-start'. It only
200 ;; works as long as point doesn't cross a macro boundary.
209 c-macro-start))
217 c-macro-start))
219 ;; One element macro cache to cope with continual movement within very large
220 ;; CPP macros.
223 ;; Nil or cons of the bounds of the most recent CPP form probed by
224 ;; `c-beginning-of-macro', `c-end-of-macro' or `c-syntactic-end-of-macro'.
225 ;; The cdr will be nil if we know only the start of the CPP form.
228 ;; The starting position from where we determined `c-macro-cache'.
231 ;; non-nil iff `c-macro-cache' has both elements set AND the cdr is at a
232 ;; syntactic end of macro, not merely an apparent one.
235 ;; Called from a before-change function. If the change region is before or
236 ;; in the macro characterized by `c-macro-cache' etc., nullify it
237 ;; appropriately. BEG and END are the standard before-change-functions
238 ;; parameters. END isn't used.
243 c-macro-cache-start-pos nil
244 c-macro-cache-syntactic nil))
249 c-macro-cache-syntactic nil))))
252 "Go to the beginning of a preprocessor directive.
253 Leave point at the beginning of the directive and return t if in one,
254 otherwise return nil and leave point unchanged.
256 Note that this function might do hidden buffer changes. See the
257 comment at the start of cc-engine.el for more info."
269 t))
271 c-macro-cache-start-pos nil
272 c-macro-cache-syntactic nil)
284 c-macro-cache-start-pos here)
285 t)
287 nil))))))
290 "Go to the end of a preprocessor directive.
291 More accurately, move the point to the end of the closest following
292 line that doesn't end with a line continuation backslash - no check is
293 done that the point is inside a cpp directive to begin with.
295 Note that this function might do hidden buffer changes. See the
296 comment at the start of cc-engine.el for more info."
305 c-macro-cache-start-pos nil
306 c-macro-cache-syntactic nil))
312 t)))
317 ;; Go to the end of a CPP directive, or a "safe" pos just before.
318 ;;
319 ;; This is normally the end of the next non-escaped line. A "safe"
320 ;; position is one not within a string or comment. (The EOL on a line
321 ;; comment is NOT "safe").
322 ;;
323 ;; This function must only be called from the beginning of a CPP construct.
324 ;;
325 ;; Note that this function might do hidden buffer changes. See the comment
326 ;; at the start of cc-engine.el for more info.
329 s)
341 ;; Assuming point is at the "#" that introduces a preprocessor
342 ;; directive, it's moved forward to the end of the identifier which is
343 ;; "#define"d (or whatever c-opt-cpp-macro-define specifies). Non-nil
344 ;; is returned in this case, in all other cases nil is returned and
345 ;; point isn't moved.
346 ;;
347 ;; This function might do hidden buffer changes.
353 ;; Assuming point is at the "#" that introduces a preprocessor
354 ;; directive, it's moved forward to the start of the definition body
355 ;; if it's a "#define" (or whatever c-opt-cpp-macro-define
356 ;; specifies). Non-nil is returned in this case, in all other cases
357 ;; nil is returned and point isn't moved.
358 ;;
359 ;; This function might do hidden buffer changes.
365 \f
366 ;;; Basic utility functions.
369 ;; Return the given region as a string where all syntactic
370 ;; whitespace is removed or, where necessary, replaced with a single
371 ;; space. If PAREN-LEVEL is given then all parens in the region are
372 ;; collapsed to "()", "[]" etc.
373 ;;
374 ;; This function might do hidden buffer changes.
418 ;; Shift the indentation of the current line with the specified
419 ;; amount (positive inwards). The buffer is modified only if
420 ;; SHIFT-AMT isn't equal to zero.
423 tmp-char-inserted)
425 nil
426 ;; If we're on an empty line inside a macro, we take the point
427 ;; to be at the current indentation and shift it to the
428 ;; appropriate column. This way we don't treat the extra
429 ;; whitespace out to the line continuation as indentation.
445 ;; If initial point was within line's indentation and we're not on
446 ;; a line with a line continuation in a macro, position after the
447 ;; indentation. Else stay at same point in text.
455 ;; Return non-nil if the string KEYWORD is a known keyword. More
456 ;; precisely, the value is the symbol for the keyword in
457 ;; `c-keywords-obarray'.
461 ;; Return non-nil if the symbol KEYWORD-SYM, as returned by
462 ;; `c-keyword-sym', is a member of LANG-CONSTANT, which is the name
463 ;; of a language constant that ends with "-kwds". If KEYWORD-SYM is
464 ;; nil then the result is nil.
467 ;; String syntax chars, suitable for skip-syntax-(forward|backward).
472 ;; Regexp matching string limit syntax.
477 ;; Regexp matching WS followed by string limit syntax.
481 ;; Holds formatted error strings for the few cases where parse errors
482 ;; are reported.
489 c-parsing-error)
491 ;; Faces given to comments and string literals. This is used in some
492 ;; situations to speed up recognition; it isn't mandatory that font
493 ;; locking is in use. This variable is extended with the face in
494 ;; `c-doc-face-name' when fontification is activated in cc-fonts.el.
498 ;; New in Emacs 22.
502 ;; Put a c-type property with the given value at POS.
506 ;; Remove all occurrences of the c-type property that has the given
507 ;; value in the region between FROM and TO. VALUE is assumed to not
508 ;; be nil.
509 ;;
510 ;; Note: This assumes that c-type is put on single chars only; it's
511 ;; very inefficient if matching properties cover large regions.
520 \f
521 ;; Some debug tools to visualize various special positions. This
522 ;; debug code isn't as portable as the rest of CC Mode.
558 \f
559 ;; `c-beginning-of-statement-1' and accompanying stuff.
561 ;; KLUDGE ALERT: c-maybe-labelp is used to pass information between
562 ;; c-crosses-statement-barrier-p and c-beginning-of-statement-1. A
563 ;; better way should be implemented, but this will at least shut up
564 ;; the byte compiler.
567 ;; New awk-compatible version of c-beginning-of-statement-1, ACM 2002/6/22
569 ;; Macros used internally in c-beginning-of-statement-1 for the
570 ;; automaton actions.
573 stack)))
578 t
579 ,do-if-done
584 ;; Throw nil if stack is empty, else throw non-nil.
608 noerror comma-delim)
609 "Move to the start of the current statement or declaration, or to
610 the previous one if already at the beginning of one. Only
611 statements/declarations on the same level are considered, i.e. don't
612 move into or out of sexps (not even normal expression parentheses).
614 If point is already at the earliest statement within braces or parens,
615 this function doesn't move back into any whitespace preceding it; it
616 returns 'same in this case.
620 is within the continuation. If starting at such a token, move to the
621 corresponding statement start. If at the beginning of a statement,
622 move to the closest containing statement if there is any. This might
623 also stop at a continuation clause.
625 Labels are treated as part of the following statements if
626 IGNORE-LABELS is non-nil. (FIXME: Doesn't work if we stop at a known
627 statement start keyword.) Otherwise, each label is treated as a
628 separate statement.
630 Macros are ignored \(i.e. skipped over) unless point is within one, in
631 which case the content of the macro is treated as normal code. Aside
632 from any normal statement starts found in it, stop at the first token
635 macros before leaving them.
637 Return:
639 'same if stopped at the beginning of the current statement;
640 'up if stepped to a containing statement;
641 'previous if stepped to a preceding statement;
642 'beginning if stepped from a statement continuation clause to
643 its start clause; or
644 'macro if stepped to a macro start.
645 Note that 'same and not 'label is returned if stopped at the same
646 label without crossing the colon character.
648 LIM may be given to limit the search. If the search hits the limit,
649 point will be left at the closest following token, or at the start
650 position if that is less ('same is returned in this case).
652 NOERROR turns off error logging to `c-parsing-error'.
654 Normally only ';' and virtual semicolons are considered to delimit
655 statements, but if COMMA-DELIM is non-nil then ',' is treated
656 as a delimiter too.
658 Note that this function might do hidden buffer changes. See the
659 comment at the start of cc-engine.el for more info."
661 ;; The bulk of this function is a pushdown automaton that looks at statement
662 ;; boundaries and the tokens (such as "while") in c-opt-block-stmt-key. Its
663 ;; purpose is to keep track of nested statements, ensuring that such
664 ;; statements are skipped over in their entirety (somewhat akin to what C-M-p
665 ;; does with nested braces/brackets/parentheses).
666 ;;
667 ;; Note: The position of a boundary is the following token.
668 ;;
669 ;; Beginning with the current token (the one following point), move back one
670 ;; sexp at a time (where a sexp is, more or less, either a token or the
671 ;; entire contents of a brace/bracket/paren pair). Each time a statement
672 ;; boundary is crossed or a "while"-like token is found, update the state of
673 ;; the PDA. Stop at the beginning of a statement when the stack (holding
674 ;; nested statement info) is empty and the position has been moved.
675 ;;
676 ;; The following variables constitute the PDA:
677 ;;
678 ;; sym: This is either the "while"-like token (e.g. 'for) we've just
679 ;; scanned back over, 'boundary if we've just gone back over a
680 ;; statement boundary, or nil otherwise.
681 ;; state: takes one of the values (nil else else-boundary while
682 ;; while-boundary catch catch-boundary).
683 ;; nil means "no "while"-like token yet scanned".
684 ;; 'else, for example, means "just gone back over an else".
685 ;; 'else-boundary means "just gone back over a statement boundary
686 ;; immediately after having gone back over an else".
687 ;; saved-pos: A vector of either saved positions (tok ptok pptok, etc.) or
688 ;; of error reporting information.
689 ;; stack: The stack onto which the PDA pushes its state. Each entry
690 ;; consists of a saved value of state and saved-pos. An entry is
691 ;; pushed when we move back over a "continuation" token (e.g. else)
692 ;; and popped when we encounter the corresponding opening token
693 ;; (e.g. if).
694 ;;
695 ;;
696 ;; The following diagram briefly outlines the PDA.
697 ;;
698 ;; Common state:
699 ;; "else": Push state, goto state `else'.
700 ;; "while": Push state, goto state `while'.
701 ;; "catch" or "finally": Push state, goto state `catch'.
702 ;; boundary: Pop state.
703 ;; other: Do nothing special.
704 ;;
705 ;; State `else':
706 ;; boundary: Goto state `else-boundary'.
707 ;; other: Error, pop state, retry token.
708 ;;
709 ;; State `else-boundary':
710 ;; "if": Pop state.
711 ;; boundary: Error, pop state.
712 ;; other: See common state.
713 ;;
714 ;; State `while':
715 ;; boundary: Save position, goto state `while-boundary'.
716 ;; other: Pop state, retry token.
717 ;;
718 ;; State `while-boundary':
719 ;; "do": Pop state.
720 ;; boundary: Restore position if it's not at start, pop state. [*see below]
721 ;; other: See common state.
722 ;;
723 ;; State `catch':
724 ;; boundary: Goto state `catch-boundary'.
725 ;; other: Error, pop state, retry token.
726 ;;
727 ;; State `catch-boundary':
728 ;; "try": Pop state.
729 ;; "catch": Goto state `catch'.
730 ;; boundary: Error, pop state.
731 ;; other: See common state.
732 ;;
733 ;; [*] In the `while-boundary' state, we had pushed a 'while state, and were
734 ;; searching for a "do" which would have opened a do-while. If we didn't
735 ;; find it, we discard the analysis done since the "while", go back to this
736 ;; token in the buffer and restart the scanning there, this time WITHOUT
737 ;; pushing the 'while state onto the stack.
738 ;;
739 ;; In addition to the above there is some special handling of labels
740 ;; and macros.
744 macro-start
747 c-stmt-delim-chars-with-comma
748 c-stmt-delim-chars))
749 c-in-literal-cache c-maybe-labelp after-case:-pos saved
750 ;; Current position.
751 pos
752 ;; Position of last stmt boundary character (e.g. ;).
753 boundary-pos
754 ;; The position of the last sexp or bound that follows the
755 ;; first found colon, i.e. the start of the nonlabel part of
756 ;; the statement. It's `start' if a colon is found just after
757 ;; the start.
758 after-labels-pos
759 ;; Like `after-labels-pos', but the first such position inside
760 ;; a label, i.e. the start of the last label before the start
761 ;; of the nonlabel part of the statement.
762 last-label-pos
763 ;; The last position where a label is possible provided the
764 ;; statement started there. It's nil as long as no invalid
765 ;; label content has been found (according to
766 ;; `c-nonlabel-token-key'). It's `start' if no valid label
767 ;; content was found in the label. Note that we might still
768 ;; regard it a label if it starts with `c-label-kwds'.
769 label-good-pos
770 ;; Putative positions of the components of a bitfield declaration,
771 ;; e.g. "int foo : NUM_FOO_BITS ;"
772 bitfield-type-pos bitfield-id-pos bitfield-size-pos
773 ;; Symbol just scanned back over (e.g. 'while or 'boundary).
774 ;; See above.
775 sym
776 ;; Current state in the automaton. See above.
777 state
778 ;; Current saved positions. See above.
779 saved-pos
780 ;; Stack of conses (state . saved-pos).
781 stack
782 ;; Regexp which matches "for", "if", etc.
785 ;; Return value.
787 ;; Positions of the last three sexps or bounds we've stopped at.
788 tok ptok pptok)
798 ;; Try to skip back over unary operator characters, to register
799 ;; that we've moved.
803 ;; Protect post-++/-- operators just before a virtual semicolon.
807 ;; Skip back over any semicolon here. If it was a bare semicolon, we're
808 ;; done. Later on we ignore the boundaries for statements that don't
809 ;; contain any sexp. The only thing that is affected is that the error
810 ;; checking is a little less strict, and we really don't bother.
820 pos saved)
822 ;; Begin at start and not pos to detect macros if we stand
823 ;; directly after the #.
826 ;; Record this as the first token if not starting inside it.
830 ;; The following while loop goes back one sexp (balanced parens,
831 ;; etc. with contents, or symbol or suchlike) each iteration. This
832 ;; movement is accomplished with a call to c-backward-sexp approx 170
833 ;; lines below.
834 ;;
835 ;; The loop is exited only by throwing nil to the (catch 'loop ...):
836 ;; 1. On reaching the start of a macro;
837 ;; 2. On having passed a stmt boundary with the PDA stack empty;
838 ;; 3. On reaching the start of an Objective C method def;
839 ;; 4. From macro `c-bos-pop-state'; when the stack is empty;
840 ;; 5. From macro `c-bos-pop-state-and-retry' when the stack is empty.
844 ;; Are we in a macro, just after the opening #?
859 ;; Stop at the first token in the content of the macro.
863 ret 'macro
864 ignore-labels t))
867 ;; Do a round through the automaton if we've just passed a
868 ;; statement boundary or passed a "while"-like token.
876 ;; The PDA state handling.
877 ;;
878 ;; Refer to the description of the PDA in the opening
879 ;; comments. In the following OR form, the first leaf
880 ;; attempts to handles one of the specific actions detailed
881 ;; (e.g., finding token "if" whilst in state `else-boundary').
882 ;; We drop through to the second leaf (which handles common
883 ;; state) if no specific handler is found in the first cond.
884 ;; If a parsing error is detected (e.g. an "else" with no
885 ;; preceding "if"), we throw to the enclosing catch.
886 ;;
887 ;; Note that the (eq state 'else) means
888 ;; "we've just passed an else", NOT "we're looking for an
889 ;; else".
906 ;; Since this can cause backtracking we do a
907 ;; little more careful analysis to avoid it:
908 ;; If there's a label in front of the while
909 ;; it can't be part of a do-while.
938 ;; This is state common. We get here when the previous
939 ;; cond statement found no particular state handler.
941 ;; If we have a boundary at the start
942 ;; position we push a frame to go to the
943 ;; previous statement.
952 ;; Is this a real while, or a do-while?
953 ;; The next `when' triggers unless we are SURE that
954 ;; the `while' is not the tail end of a `do-while'.
957 ;; The following kludge is to prevent
958 ;; infinite recursion when called from
959 ;; c-awk-after-if-for-while-condition-p,
960 ;; or the like.
964 ;; Since this can cause backtracking we do a
965 ;; little more careful analysis to avoid it: If
966 ;; the while isn't followed by a (possibly
967 ;; virtual) semicolon it can't be a do-while.
976 ;; We're either past a statement boundary or at the
977 ;; start of a statement, so throw away any label data
978 ;; for the previous one.
980 last-label-pos nil
981 c-maybe-labelp nil))))
983 ;; Step to the previous sexp, but not if we crossed a
984 ;; boundary, since that doesn't consume an sexp.
988 ;; HERE IS THE SINGLE PLACE INSIDE THE PDA LOOP WHERE WE MOVE
989 ;; BACKWARDS THROUGH THE SOURCE.
993 ;; The end position of the area to search for statement
994 ;; barriers in this round.
997 ;; Go back over exactly one logical sexp, taking proper
998 ;; account of macros and escaped EOLs.
1002 ;; Give up if we hit an unbalanced block. Since the
1003 ;; stack won't be empty the code below will report a
1004 ;; suitable error.
1007 ;; Have we moved into a macro?
1010 ;; Have we crossed a statement boundary? If not,
1011 ;; keep going back until we find one or a "real" sexp.
1018 ;; Have we just gone back over an escaped NL? This
1019 ;; doesn't count as a sexp.
1022 ;; Have we crossed a statement boundary?
1025 ;; Are we at a macro beginning?
1027 c-opt-cpp-prefix
1033 ;; 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.
1144 pptok)
1146 ptok)
1154 after-labels-pos
1161 last-label-pos
1162 pos))
1164 ;; We're in a label. Maybe we should step to the statement
1165 ;; after it.
1170 ;; Might have jumped over several labels. Go to the last one.
1173 ;; Have we got "case <expression>:"?
1183 ;; Skip over the unary operators that can start the statement.
1186 ;; protect AWK post-inc/decrement operators, etc.
1191 ret)))
1194 "Return non-nil if there is a non-comment non-macro punctuation character
1195 between FROM and TO. FROM must not be in a string or comment. The returned
1196 value is the position of the first such character."
1207 "Return non-nil if buffer positions FROM to TO cross one or more
1208 statement or declaration boundaries. The returned value is actually
1209 the position of the earliest boundary char. FROM must not be within
1210 a string or comment.
1212 The variable `c-maybe-labelp' is set to the position of the first `:' that
1213 might start a label (i.e. not part of `::' and not preceded by `?'). If a
1214 single `?' is found, then `c-maybe-labelp' is cleared.
1216 For AWK, a statement which is terminated by an EOL (not a \; or a }) is
1218 the line. If this virtual semicolon is _at_ from, the function recognizes it.
1220 Note that this function might do hidden buffer changes. See the
1221 comment at the start of cc-engine.el for more info."
1223 ;; If the current language has CPP macros, insert # into skip-chars.
1226 c-opt-cpp-symbol ; usually "#"
1228 c-stmt-delim-chars))
1231 lit-range vsemi-pos)
1238 skip-chars
1242 ;; Virtual semicolon?
1252 ;; In a string/comment?
1259 ;; Ignore scope operators.
1263 ;; A question mark. Can't be a label, so stop
1264 ;; looking for more : and ?.
1267 ;; At a CPP construct or a "#" or "##" operator?
1278 ;; In trailing space after an as yet undetected virtual semicolon?
1281 ; unterminated string/regexp.
1286 nil))))))
1289 "Return non-nil if the point is at the first token in a statement
1290 or somewhere in the syntactic whitespace before it.
1293 Any kind of declaration-like construct that occur outside function
1296 Note that this function might do hidden buffer changes. See the
1297 comment at the start of cc-engine.el for more info."
1301 c-maybe-labelp)
1312 "Return non-nil if the point is at the first token in an expression or
1313 statement, or somewhere in the syntactic whitespace before it.
1316 grammar sense: It's any sequence of expression tokens except commas,
1317 unless they are enclosed inside parentheses of some kind. Also, an
1318 expression never continues past an enclosing parenthesis, but it might
1319 contain parenthesis pairs of any sort except braces.
1321 Since expressions never cross statement boundaries, this function also
1322 recognizes statement beginnings, just like `c-at-statement-start-p'.
1324 Note that this function might do hidden buffer changes. See the
1325 comment at the start of cc-engine.el for more info."
1330 c-maybe-labelp)
1338 \f
1339 ;; A set of functions that covers various idiosyncrasies in
1340 ;; implementations of `forward-comment'.
1342 ;; Note: Some emacsen considers incorrectly that any line comment
1343 ;; ending with a backslash continues to the next line. I can't think
1344 ;; of any way to work around that in a reliable way without changing
1345 ;; the buffer, though. Suggestions welcome. ;) (No, temporarily
1346 ;; changing the syntax for backslash doesn't work since we must treat
1347 ;; escapes in string literals correctly.)
1350 "Move forward past whitespace and the closest following comment, if any.
1351 Return t if a comment was found, nil otherwise. In either case, the
1352 point is moved past the following whitespace. Line continuations,
1353 i.e. a backslashes followed by line breaks, are treated as whitespace.
1354 The line breaks that end line comments are considered to be the
1355 comment enders, so the point will be put on the beginning of the next
1356 line if it moved past a line comment.
1358 This function does not do any hidden buffer changes."
1366 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1367 ;; forwards at eob.
1368 nil
1370 ;; Emacs includes the ending newline in a b-style (c++)
1371 ;; comment, but XEmacs doesn't. We depend on the Emacs
1372 ;; behavior (which also is symmetric).
1376 t))))
1379 "Move forward past all following whitespace and comments.
1380 Line continuations, i.e. a backslashes followed by line breaks, are
1381 treated as whitespace.
1383 Note that this function might do hidden buffer changes. See the
1384 comment at the start of cc-engine.el for more info."
1387 ;; If forward-comment in at least XEmacs 21 is given a large
1388 ;; positive value, it'll loop all the way through if it hits
1389 ;; eob.
1391 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1392 ;; forwards at eob.
1397 t))))
1400 "Move backward past whitespace and the closest preceding comment, if any.
1401 Return t if a comment was found, nil otherwise. In either case, the
1402 point is moved past the preceding whitespace. Line continuations,
1403 i.e. a backslashes followed by line breaks, are treated as whitespace.
1404 The line breaks that end line comments are considered to be the
1405 comment enders, so the point cannot be at the end of the same line to
1406 move over a line comment.
1408 This function does not do any hidden buffer changes."
1411 ;; When we got newline terminated comments, forward-comment in all
1412 ;; supported emacsen so far will stop at eol of each line not
1413 ;; ending with a comment when moving backwards. This corrects for
1414 ;; that, and at the same time handles line continuations.
1422 ;; Some emacsen (e.g. Emacs 19.34) return t when moving
1423 ;; backwards at bob.
1424 nil
1426 ;; Leave point after the closest following newline if we've
1427 ;; backed up over any above, since forward-comment won't move
1428 ;; backward over a line comment if point is at the end of the
1429 ;; same line.
1434 ;; If forward-comment above succeeded and we're at eol
1435 ;; then the newline we moved over above didn't end a
1436 ;; line comment, so we give it another go.
1439 t))
1441 ;; Emacs <= 20 and XEmacs move back over the closer of a
1442 ;; block comment that lacks an opener.
1445 t)))))
1448 "Move backward past all preceding whitespace and comments.
1449 Line continuations, i.e. a backslashes followed by line breaks, are
1450 treated as whitespace. The line breaks that end line comments are
1451 considered to be the comment enders, so the point cannot be at the end
1452 of the same line to move over a line comment. Unlike
1453 c-backward-syntactic-ws, this function doesn't move back over
1454 preprocessor directives.
1456 Note that this function might do hidden buffer changes. See the
1457 comment at the start of cc-engine.el for more info."
1461 ;; `forward-comment' in some emacsen (e.g. XEmacs 21.4)
1462 ;; return t when moving backwards at bob.
1468 ;; Cope specifically with ^M^J here -
1469 ;; forward-comment sometimes gets stuck after ^Ms,
1470 ;; sometimes after ^M^J.
1474 t)
1478 t))))
1479 moved-comment)
1481 ;; Emacs <= 20 and XEmacs move back over the
1482 ;; closer of a block comment that lacks an opener.
1484 t)
1486 ;; XEmacs treats line continuations as whitespace but
1487 ;; only in the backward direction, which seems a bit
1488 ;; odd. Anyway, this is necessary for Emacs.
1493 t))))))
1495 \f
1496 ;; Tools for skipping over syntactic whitespace.
1498 ;; The following functions use text properties to cache searches over
1499 ;; large regions of syntactic whitespace. It works as follows:
1500 ;;
1501 ;; o If a syntactic whitespace region contains anything but simple
1502 ;; whitespace (i.e. space, tab and line breaks), the text property
1503 ;; `c-in-sws' is put over it. At places where we have stopped
1504 ;; within that region there's also a `c-is-sws' text property.
1505 ;; That since there typically are nested whitespace inside that
1506 ;; must be handled separately, e.g. whitespace inside a comment or
1507 ;; cpp directive. Thus, from one point with `c-is-sws' it's safe
1508 ;; to jump to another point with that property within the same
1509 ;; `c-in-sws' region. It can be likened to a ladder where
1510 ;; `c-in-sws' marks the bars and `c-is-sws' the rungs.
1511 ;;
1512 ;; o The `c-is-sws' property is put on the simple whitespace chars at
1513 ;; a "rung position" and also maybe on the first following char.
1514 ;; As many characters as can be conveniently found in this range
1515 ;; are marked, but no assumption can be made that the whole range
1516 ;; is marked (it could be clobbered by later changes, for
1517 ;; instance).
1518 ;;
1519 ;; Note that some part of the beginning of a sequence of simple
1520 ;; whitespace might be part of the end of a preceding line comment
1521 ;; or cpp directive and must not be considered part of the "rung".
1522 ;; Such whitespace is some amount of horizontal whitespace followed
1523 ;; by a newline. In the case of cpp directives it could also be
1524 ;; two newlines with horizontal whitespace between them.
1525 ;;
1526 ;; The reason to include the first following char is to cope with
1527 ;; "rung positions" that doesn't have any ordinary whitespace. If
1528 ;; `c-is-sws' is put on a token character it does not have
1529 ;; `c-in-sws' set simultaneously. That's the only case when that
1530 ;; can occur, and the reason for not extending the `c-in-sws'
1531 ;; region to cover it is that the `c-in-sws' region could then be
1532 ;; accidentally merged with a following one if the token is only
1533 ;; one character long.
1534 ;;
1535 ;; o On buffer changes the `c-in-sws' and `c-is-sws' properties are
1536 ;; removed in the changed region. If the change was inside
1537 ;; syntactic whitespace that means that the "ladder" is broken, but
1538 ;; a later call to `c-forward-sws' or `c-backward-sws' will use the
1539 ;; parts on either side and use an ordinary search only to "repair"
1540 ;; the gap.
1541 ;;
1542 ;; Special care needs to be taken if a region is removed: If there
1543 ;; are `c-in-sws' on both sides of it which do not connect inside
1544 ;; the region then they can't be joined. If e.g. a marked macro is
1545 ;; broken, syntactic whitespace inside the new text might be
1546 ;; marked. If those marks would become connected with the old
1547 ;; `c-in-sws' range around the macro then we could get a ladder
1548 ;; with one end outside the macro and the other at some whitespace
1549 ;; within it.
1550 ;;
1551 ;; The main motivation for this system is to increase the speed in
1552 ;; skipping over the large whitespace regions that can occur at the
1553 ;; top level in e.g. header files that contain a lot of comments and
1554 ;; cpp directives. For small comments inside code it's probably
1555 ;; slower than using `forward-comment' straightforwardly, but speed is
1556 ;; not a significant factor there anyway.
1558 ; (defface c-debug-is-sws-face
1559 ; '((t (:background "GreenYellow")))
1560 ; "Debug face to mark the `c-is-sws' property.")
1561 ; (defface c-debug-in-sws-face
1562 ; '((t (:underline t)))
1563 ; "Debug face to mark the `c-in-sws' property.")
1565 ; (defun c-debug-put-sws-faces ()
1566 ; ;; Put the sws debug faces on all the `c-is-sws' and `c-in-sws'
1567 ; ;; properties in the buffer.
1568 ; (interactive)
1569 ; (save-excursion
1570 ; (c-save-buffer-state (in-face)
1571 ; (goto-char (point-min))
1572 ; (setq in-face (if (get-text-property (point) 'c-is-sws)
1573 ; (point)))
1574 ; (while (progn
1575 ; (goto-char (next-single-property-change
1576 ; (point) 'c-is-sws nil (point-max)))
1577 ; (if in-face
1578 ; (progn
1579 ; (c-debug-add-face in-face (point) 'c-debug-is-sws-face)
1580 ; (setq in-face nil))
1581 ; (setq in-face (point)))
1582 ; (not (eobp))))
1583 ; (goto-char (point-min))
1584 ; (setq in-face (if (get-text-property (point) 'c-in-sws)
1585 ; (point)))
1586 ; (while (progn
1587 ; (goto-char (next-single-property-change
1588 ; (point) 'c-in-sws nil (point-max)))
1589 ; (if in-face
1590 ; (progn
1591 ; (c-debug-add-face in-face (point) 'c-debug-in-sws-face)
1592 ; (setq in-face nil))
1593 ; (setq in-face (point)))
1594 ; (not (eobp)))))))
1597 ;;`(message ,@args)
1598 )
1601 ;; This macro does a hidden buffer change.
1608 ;; This macro does a hidden buffer change.
1615 ;; This macro does a hidden buffer change.
1622 ;; This macro does a hidden buffer change.
1629 ;; This macro does a hidden buffer change.
1637 ;; Called from `after-change-functions'. Note that if
1638 ;; `c-forward-sws' or `c-backward-sws' are used outside
1639 ;; `c-save-buffer-state' or similar then this will remove the cache
1640 ;; properties right after they're added.
1641 ;;
1642 ;; This function does hidden buffer changes.
1645 ;; Adjust the end to remove the properties in any following simple
1646 ;; ws up to and including the next line break, if there is any
1647 ;; after the changed region. This is necessary e.g. when a rung
1648 ;; marked empty line is converted to a line comment by inserting
1649 ;; "//" before the line break. In that case the line break would
1650 ;; keep the rung mark which could make a later `c-backward-sws'
1651 ;; move into the line comment instead of over it.
1661 ;; Ensure that an `c-in-sws' range gets broken. Note that it isn't
1662 ;; safe to keep a range that was continuous before the change. E.g:
1663 ;;
1664 ;; #define foo
1665 ;; \
1666 ;; bar
1667 ;;
1668 ;; There can be a "ladder" between "#" and "b". Now, if the newline
1669 ;; after "foo" is removed then "bar" will become part of the cpp
1670 ;; directive instead of a syntactically relevant token. In that
1671 ;; case there's no longer syntactic ws from "#" to "b".
1678 ;; Used by `c-forward-syntactic-ws' to implement the unbounded search.
1679 ;;
1680 ;; This function might do hidden buffer changes.
1683 ;; unmarked part of the simple ws region.
1685 rung-is-marked next-rung-is-marked simple-ws-end
1686 ;; `safe-start' is set when it's safe to cache the start position.
1687 ;; It's not set if we've initially skipped over comments and line
1688 ;; continuations since we might have gone out through the end of a
1689 ;; macro then. This provision makes `c-forward-sws' not populate the
1690 ;; cache in the majority of cases, but otoh is `c-backward-sws' by far
1691 ;; more common.
1692 safe-start)
1694 ;; Skip simple ws and do a quick check on the following character to see
1695 ;; if it's anything that can't start syntactic ws, so we can bail out
1696 ;; early in the majority of cases when there just are a few ws chars.
1703 ;; Find the last rung position to avoid setting properties in all
1704 ;; the cases when the marked rung is complete.
1705 ;; (`next-single-property-change' is certain to move at least one
1706 ;; step forward.)
1709 ;; Got no marked rung here. Since the simple ws might have started
1710 ;; inside a line comment or cpp directive we must set `rung-pos' as
1711 ;; high as possible.
1721 ;; The following search is the main reason that `c-in-sws'
1722 ;; and `c-is-sws' aren't combined to one property.
1726 ;; If the `c-in-sws' region extended past the last
1727 ;; `c-is-sws' char we have to go back a bit.
1734 "c-forward-sws cached move %s -> %s (max %s)"
1741 ;; We'll loop here if there is simple ws after the last rung.
1742 ;; That means that there's been some change in it and it's
1743 ;; possible that we've stepped into another ladder, so extend
1744 ;; the previous one to join with it if there is one, and try to
1745 ;; use the cache again.
1747 "c-forward-sws extending rung with [%s..%s] (max %s)"
1750 ;; Remove any `c-in-sws' property from the last char of
1751 ;; the rung before we mark it with `c-is-sws', so that we
1752 ;; won't connect with the remains of a broken "ladder".
1758 last-put-in-sws-pos rung-pos)))
1765 ;; Skipped over comments. Don't cache at eob in case the buffer
1766 ;; is narrowed.
1777 ;; Skip a preprocessor directive.
1784 ;; Don't cache at eob in case the buffer is narrowed.
1787 ;; We've searched over a piece of non-white syntactic ws. See if this
1788 ;; can be cached.
1794 ;; Cache if we haven't skipped comments only, and if we started
1795 ;; either from a marked rung or from a completely uncached
1796 ;; position.
1801 ;; See if there's a marked rung in the encountered simple ws. If
1802 ;; so then we can cache, unless `safe-start' is nil. Even then
1803 ;; we need to do this to check if the cache can be used for the
1804 ;; next step.
1808 safe-start))
1812 "c-forward-sws caching [%s..%s] - [%s..%s] (max %s)"
1816 ;; Remove the properties for any nested ws that might be cached.
1817 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
1818 ;; anyway.
1826 last-put-in-sws-pos rung-pos))
1828 ;; Remove any `c-in-sws' property from the last char of
1829 ;; the rung before we mark it with `c-is-sws', so that we
1830 ;; won't connect with the remains of a broken "ladder".
1833 rung-end-pos))
1836 "c-forward-sws not caching [%s..%s] - [%s..%s] (max %s)"
1840 ;; Set `rung-pos' for the next rung. It's the same thing here as
1841 ;; initially, except that the rung position is set as early as
1842 ;; possible since we can't be in the ending ws of a line comment or
1843 ;; cpp directive now.
1850 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
1851 ;; another one after the point (which might occur when editing inside a
1852 ;; comment or macro).
1856 "c-forward-sws clearing at %s for cache separation"
1857 last-put-in-sws-pos)
1861 ;; If at eob we have to clear the last character before the end
1862 ;; instead since the buffer might be narrowed and there might
1863 ;; be a `c-in-sws' after (point-max). In this case it's
1864 ;; necessary to clear both properties.
1866 "c-forward-sws clearing thoroughly at %s for cache separation"
1869 last-put-in-sws-pos))))
1870 ))))
1873 ;; Used by `c-backward-syntactic-ws' to implement the unbounded search.
1874 ;;
1875 ;; This function might do hidden buffer changes.
1878 ;; part of the simple ws region.
1880 rung-is-marked simple-ws-beg cmt-skip-pos)
1882 ;; Skip simple horizontal ws and do a quick check on the preceding
1883 ;; character to see if it's anything that can't end syntactic ws, so we can
1884 ;; bail out early in the majority of cases when there just are a few ws
1885 ;; chars. Newlines are complicated in the backward direction, so we can't
1886 ;; skip over them.
1893 ;; Try to find a rung position in the simple ws preceding point, so that
1894 ;; we can get a cache hit even if the last bit of the simple ws has
1895 ;; changed recently.
1901 ;; `rung-pos' will be the earliest marked position, which means that
1902 ;; there might be later unmarked parts in the simple ws region.
1903 ;; It's not worth the effort to fix that; the last part of the
1904 ;; simple ws is also typically edited often, so it could be wasted.
1916 ;; The following search is the main reason that `c-in-sws'
1917 ;; and `c-is-sws' aren't combined to one property.
1921 ;; If the `c-in-sws' region extended past the first
1922 ;; `c-is-sws' char we have to go forward a bit.
1927 "c-backward-sws cached move %s <- %s (min %s)"
1939 t
1941 nil))
1943 ;; We'll loop here if there is simple ws before the first rung.
1944 ;; That means that there's been some change in it and it's
1945 ;; possible that we've stepped into another ladder, so extend
1946 ;; the previous one to join with it if there is one, and try to
1947 ;; use the cache again.
1949 "c-backward-sws extending rung with [%s..%s] (min %s)"
1952 ;; Remove any `c-in-sws' property from the last char of
1953 ;; the rung before we mark it with `c-is-sws', so that we
1954 ;; won't connect with the remains of a broken "ladder".
1957 rung-pos)
1961 last-put-in-sws-pos rung-pos))
1970 ;; Inside a cpp directive. See if it should be skipped over.
1973 ;; Move back over all line continuations in the region skipped
1974 ;; over by `c-backward-comments'. If we go past it then we
1975 ;; started inside the cpp directive.
1984 ;; Don't move past the cpp directive if we began inside
1985 ;; it. Note that the position at the end of the last line
1986 ;; of the macro is also considered to be within it.
1988 nil)
1990 ;; It's worthwhile to spend a little bit of effort on finding
1991 ;; the end of the macro, to get a good `simple-ws-beg'
1992 ;; position for the cache. Note that `c-backward-comments'
1993 ;; could have stepped over some comments before going into
1994 ;; the macro, and then `simple-ws-beg' must be kept on the
1995 ;; same side of those comments.
2002 ;; Might happen if comments after the macro were skipped
2003 ;; over.
2007 t)))
2012 simple-ws-beg)
2013 ;; Skipped over comments. Must put point at the end of
2014 ;; the simple ws at point since we might be after a line
2015 ;; comment or cpp directive that's been partially
2016 ;; narrowed out, and we can't risk marking the simple ws
2017 ;; at the end of it.
2019 t)))
2021 ;; We've searched over a piece of non-white syntactic ws. See if this
2022 ;; can be cached.
2027 ;; Cache if we started either from a marked rung or from a
2028 ;; completely uncached position.
2029 rung-is-marked
2032 ;; Cache if there's a marked rung in the encountered simple ws.
2040 "c-backward-sws caching [%s..%s] - [%s..%s] (min %s)"
2045 ;; Remove the properties for any nested ws that might be cached.
2046 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
2047 ;; anyway.
2052 ;; Remove any `c-in-sws' property from the last char of
2053 ;; the rung before we mark it with `c-is-sws', so that we
2054 ;; won't connect with the remains of a broken "ladder".
2057 rung-end-pos)
2060 last-put-in-sws-pos simple-ws-beg)
2061 rung-pos)
2066 "c-backward-sws not caching [%s..%s] - [%s..%s] (min %s)"
2072 ))
2074 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
2075 ;; another one before the point (which might occur when editing inside a
2076 ;; comment or macro).
2080 "c-backward-sws clearing at %s for cache separation"
2083 last-put-in-sws-pos))
2085 ;; If at bob and the buffer is narrowed, we have to clear the
2086 ;; character we're standing on instead since there might be a
2087 ;; `c-in-sws' before (point-min). In this case it's necessary
2088 ;; to clear both properties.
2090 "c-backward-sws clearing thoroughly at %s for cache separation"
2091 last-put-in-sws-pos)
2094 ))))
2096 \f
2097 ;; Other whitespace tools
2099 ;; Is the region (beg end) WS, and is there WS (or BOB/EOB) next to the
2100 ;; region? This is a "heuristic" function. .....
2101 ;;
2102 ;; The motivation for the second bit is to check whether removing this
2103 ;; region would coalesce two symbols.
2104 ;;
2105 ;; FIXME!!! This function doesn't check virtual semicolons in any way. Be
2106 ;; careful about using this function for, e.g. AWK. (2007/3/7)
2115 \f
2116 ;; A system for finding noteworthy parens before the point.
2119 ;; A maximum comfortable scanning distance, e.g. between
2120 ;; `c-state-cache-good-pos' and "HERE" (where we call c-parse-state). When
2121 ;; this distance is exceeded, we take "emergency measures", e.g. by clearing
2122 ;; the cache and starting again from point-min or a beginning of defun. This
2123 ;; value can be tuned for efficiency or set to a lower value for testing.
2127 ;; The state cache used by `c-parse-state' to cut down the amount of
2128 ;; searching. It's the result from some earlier `c-parse-state' call. See
2129 ;; `c-parse-state''s doc string for details of its structure.
2130 ;;
2131 ;; The use of the cached info is more effective if the next
2132 ;; `c-parse-state' call is on a line close by the one the cached state
2133 ;; was made at; the cache can actually slow down a little if the
2134 ;; cached state was made very far back in the buffer. The cache is
2135 ;; most effective if `c-parse-state' is used on each line while moving
2136 ;; forward.
2140 ;; This is a position where `c-state-cache' is known to be correct, or
2141 ;; nil (see below). It's a position inside one of the recorded unclosed
2142 ;; parens or the top level, but not further nested inside any literal or
2143 ;; subparen that is closed before the last recorded position.
2144 ;;
2145 ;; The exact position is chosen to try to be close to yet earlier than
2146 ;; the position where `c-state-cache' will be called next. Right now
2147 ;; the heuristic is to set it to the position after the last found
2148 ;; closing paren (of any type) before the line on which
2149 ;; `c-parse-state' was called. That is chosen primarily to work well
2150 ;; with refontification of the current line.
2151 ;;
2152 ;; 2009-07-28: When `c-state-point-min' and the last position where
2153 ;; `c-parse-state' or for which `c-invalidate-state-cache' was called, are
2154 ;; both in the same literal, there is no such "good position", and
2155 ;; c-state-cache-good-pos is then nil. This is the ONLY circumstance in which
2156 ;; it can be nil. In this case, `c-state-point-min-literal' will be non-nil.
2157 ;;
2158 ;; 2009-06-12: In a brace desert, c-state-cache-good-pos may also be in
2159 ;; the middle of the desert, as long as it is not within a brace pair
2160 ;; recorded in `c-state-cache' or a paren/bracket pair.
2163 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2164 ;; We maintain a simple cache of positions which aren't in a literal, so as to
2165 ;; speed up testing for non-literality.
2167 ;; The approximate interval between entries in `c-state-nonlit-pos-cache'.
2171 ;; A list of buffer positions which are known not to be in a literal or a cpp
2172 ;; construct. This is ordered with higher positions at the front of the list.
2173 ;; Only those which are less than `c-state-nonlit-pos-cache-limit' are valid.
2177 ;; An upper limit on valid entries in `c-state-nonlit-pos-cache'. This is
2178 ;; reduced by buffer changes, and increased by invocations of
2179 ;; `c-state-literal-at'.
2183 ;; A list of buffer positions which are known not to be in a literal. This is
2184 ;; ordered with higher positions at the front of the list. Only those which
2185 ;; are less than `c-state-semi-nonlit-pos-cache-limit' are valid.
2189 ;; An upper limit on valid entries in `c-state-semi-nonlit-pos-cache'. This is
2190 ;; reduced by buffer changes, and increased by invocations of
2191 ;; `c-state-literal-at'. FIXME!!!
2194 ;; Do a parse-partial-sexp from FROM to TO, returning either
2195 ;; (STATE TYPE (BEG . END)) if TO is in a literal; or
2196 ;; (STATE) otherwise,
2197 ;; where STATE is the parsing state at TO, TYPE is the type of the literal
2198 ;; (one of 'c, 'c++, 'string) and (BEG . END) is the boundaries of the literal.
2199 ;;
2200 ;; Unless NOT-IN-DELIMITER is non-nil, when TO is inside a two-character
2201 ;; comment opener, this is recognized as being in a comment literal.
2202 ;;
2203 ;; Only elements 3 (in a string), 4 (in a comment), 5 (following a quote),
2204 ;; 7 (comment type) and 8 (start of comment/string) (and possibly 9) of
2205 ;; STATE are valid.
2208 ty co-st)
2216 nil ; TARGETDEPTH
2217 nil ; STOPBEFORE
2218 s ; OLDSTATE
2236 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2237 ;; string, comment, or macro.
2238 ;;
2239 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2240 ;; MAY NOT contain any positions within macros, since macros are frequently
2241 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2242 ;; We cannot rely on this mechanism whilst determining a cache pos since
2243 ;; this function is also called from outwith `c-parse-state'.
2248 pos npos high-pos lit macro-beg macro-end)
2249 ;; Trim the cache to take account of buffer changes.
2261 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2265 ;; Test for being in a literal. If so, go to after it.
2272 ;; Test for being in a macro. If so, go to after it.
2287 ;; Add one extra element above HERE so as to to avoid the previous
2288 ;; expensive calculation when the next call is close to the current
2289 ;; one. This is especially useful when inside a large macro.
2294 pos))))
2297 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2298 ;; string or comment. It may be in a macro.
2303 pos npos high-pos lit macro-beg macro-end)
2304 ;; Trim the cache to take account of buffer changes.
2316 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2320 ;; Test for being in a literal. If so, go to after it.
2333 pos))))
2336 ;; If position HERE is inside a literal, return (START . END), the
2337 ;; boundaries of the literal (which may be outside the accessible bit of the
2338 ;; buffer). Otherwise, return nil.
2339 ;;
2340 ;; This function is almost the same as `c-literal-limits'. Previously, it
2341 ;; differed in that it was a lower level function, and that it rigorously
2342 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2343 ;; virtually identical to this function.
2351 ;; Return the start of the literal containing POS, or POS itself.
2353 pos))
2356 ;; Return a position outside of a string/comment/macro at or before POS.
2357 ;; STATE is the parse-partial-sexp state at POS.
2361 pos)))
2366 res))))
2368 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2369 ;; Stuff to do with point-min, and coping with any literal there.
2372 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2373 ;; narrowing is likely to affect the parens that are visible before the point.
2379 ;; These two variables define the literal, if any, containing point-min.
2380 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2381 ;; literal. If there's no literal there, they're both nil.
2385 ;; This is the earliest buffer-pos from which scanning can be done. It is
2386 ;; either the end of the literal containing point-min, or point-min itself.
2387 ;; It becomes nil if the buffer is changed earlier than this point.
2389 ;; Return the lowest valid scanning pos. This will be the end of the
2390 ;; literal enclosing point-min, or point-min itself.
2402 ;; Determine the properties of any literal containing POINT-MIN, setting the
2403 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2404 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2406 lit)
2421 c-state-point-min-lit-start nil
2422 c-state-min-scan-pos p-min)))))
2425 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2426 ;; A variable which signals a brace dessert - helpful for reducing the number
2427 ;; of fruitless backward scans.
2430 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2431 ;; that defun has searched backwards for a brace pair and not found one. Its
2432 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2433 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2434 ;; nil when at top level) and FROM is where the backward search started. It
2435 ;; is reset to nil in `c-invalidate-state-cache'.
2438 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2439 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2440 ;; list of like structure.
2442 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2443 ;; (default `c-state-cache') (or nil).
2450 ;; Return the address of the latest brace/bracket/paren (whether left or
2451 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2458 ;; Return the position just after the latest brace/bracket/paren (whether
2459 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2467 ;; From the state-cache, determine the buffer position from which we might
2468 ;; scan forward to HERE to update this cache. This position will be just
2469 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2470 ;; return the earliest position in the accessible region which isn't within
2471 ;; a literal. If the visible portion of the buffer is entirely within a
2472 ;; literal, return NIL.
2474 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2490 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2491 ;; Variables which keep track of preprocessor constructs.
2500 ;; These are the limits of the macro containing point at the previous call of
2501 ;; `c-parse-state', or nil.
2503 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2504 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2506 ;; Return the position of the opening paren/brace/bracket before HERE- which
2507 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2508 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2509 ;;
2510 ;; ............................................
2511 ;; | |
2512 ;; ( [ ( .........#macro.. ) ( ) ] )
2513 ;; ^ ^ ^ ^
2514 ;; | | | |
2515 ;; return HERE- HERE+ TOP
2516 ;;
2517 ;; If there aren't enough opening paren/brace/brackets, return the position
2518 ;; of the outermost one found, or HERE- if there are none. If there are no
2519 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2520 ;; must not be inside literals. Only the accessible portion of the buffer
2521 ;; will be scanned.
2523 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2524 ;; `here'. Go round the next loop each time we pass over such a ")". These
2525 ;; probably match "("s before `here-'.
2537 ;; PART 2: Scan back before `here-' searching for the "("s
2538 ;; matching/mismatching the ")"s found above. We only need to direct the
2539 ;; caller to scan when we've encountered unmatched right parens.
2548 pos))
2551 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2552 ;; to minimize the amount of scanning. HERE is the pertinent position in
2553 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2554 ;; its head trimmed) is known to be good, or nil if there is no such
2555 ;; position.
2556 ;;
2557 ;; The return value is a list, one of the following:
2558 ;;
2559 ;; o - ('forward START-POINT) - scan forward from START-POINT,
2560 ;; which is not less than the highest position in `c-state-cache' below HERE,
2561 ;; which is after GOOD-POS.
2562 ;; o - ('backward nil) - scan backwards (from HERE).
2563 ;; o - ('back-and-forward START-POINT) - like 'forward, but when HERE is earlier
2564 ;; than GOOD-POS.
2565 ;; o - ('IN-LIT nil) - point is inside the literal containing point-min.
2567 strategy ; 'forward, 'backward, or 'IN-LIT.
2568 start-point)
2580 start-point cache-pos)))
2584 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2585 ;; Routines which change `c-state-cache' and associated values.
2587 ;; The region (more precisely, point-min) has changed since we
2588 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2590 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2591 ;; It would be possible to do a better job here and recalculate the top
2592 ;; only.
2596 c-state-cache-good-pos c-state-min-scan-pos
2597 c-state-brace-pair-desert nil))
2599 ;; point-min has MOVED FORWARD.
2601 ;; Is the new point-min inside a (different) literal?
2606 ;; Cut off a bit of the tail from `c-state-cache'.
2608 pa)
2616 c-state-cache-good-pos c-state-min-scan-pos)
2619 )))
2624 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2625 ;; of nesting (not necessarily immediately preceding), push a cons onto
2626 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2627 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2628 ;; UPPER-LIM.
2629 ;;
2630 ;; Return non-nil when this has been done.
2631 ;;
2632 ;; The situation it copes with is this transformation:
2633 ;;
2634 ;; OLD: { (.) {...........}
2635 ;; ^ ^
2636 ;; FROM HERE
2637 ;;
2638 ;; NEW: { {....} (.) {.........
2639 ;; ^ ^ ^
2640 ;; LOWER BRACE PAIR HERE or HERE
2641 ;;
2642 ;; This routine should be fast. Since it can get called a LOT, we maintain
2643 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2644 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2654 ;; Don't start scanning in the middle of a CPP construct unless
2655 ;; it contains HERE - these constructs, in Emacs, are "commented
2656 ;; out" with category properties.
2659 macro-start-or-from
2660 from))
2664 ;; If we're essentially repeating a fruitless search, just give up.
2670 ;; DESERT-LIM. Avoid repeated searching through the cached desert.
2676 ;; CACHE-LIM. This limit will be necessary when an opening
2677 ;; paren at `cache-pos' has just had its matching close paren
2678 ;; inserted into the buffer. `cache-pos' continues to be a
2679 ;; search bound, even though the algorithm below would skip
2680 ;; over the new paren pair.
2689 ;; The top limit is EOB to ensure that `bra' is inside the
2690 ;; accessible part of the buffer at the next scan operation.
2693 ;; In the next pair of nested loops, the inner one moves back past a
2694 ;; pair of (mis-)matching parens or brackets; the outer one moves
2695 ;; back over a sequence of unmatched close brace/paren/bracket each
2696 ;; time round.
2712 ; brace, ignore it.
2715 ;; We've found the desired brace-pair.
2727 ;; We haven't found a brace pair. Record this in the cache.
2730 bra
2735 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2736 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2737 ;; "push" "a" brace pair onto `c-state-cache'.
2738 ;;
2739 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2740 ;; otherwise push it normally.
2741 ;;
2742 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2743 ;; latter is inside a macro, not being a macro containing
2744 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2745 ;; base pair. This latter case is assumed to be rare.
2746 ;;
2747 ;; Note: POINT is not preserved in this routine.
2757 c-state-cache)))
2758 ;; N.B. This defsubst codes one method for the simple, normal case,
2759 ;; and a more sophisticated, slower way for the general case. Don't
2760 ;; eliminate this defsubst - it's a speed optimization.
2764 ;; Scan the buffer from FROM to HERE, adding elements into `c-state-cache'
2765 ;; for braces etc. Return a candidate for `c-state-cache-good-pos'.
2766 ;;
2767 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2768 ;; any. Typically, it is immediately after it. It must not be inside a
2769 ;; literal.
2775 here)))
2778 ; Is actually the pos. to scan for a (/{/[ from,
2779 ; which sometimes is after a silly )/}/].
2782 ; good-pos.
2790 ;; Each time round the following loop, we enter a successively deeper
2791 ;; level of brace/paren nesting. (Except sometimes we "continue at
2792 ;; the existing level".) `pa+1' is a pos inside an opening
2793 ;; brace/paren/bracket, usually just after it.
2796 ;; Each time round the next loop moves forward over an opening then
2797 ;; a closing brace/bracket/paren. This loop is white hot, so it
2798 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2799 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2800 ;; call of `scan-lists' signals an error, which happens when there
2801 ;; are no more b/b/p's to scan.
2812 ;; We've just entered a deeper nesting level.
2814 ;; Insert the brace pair (if present) and the single open
2815 ;; paren/brace/bracket into `c-state-cache' It cannot be
2816 ;; inside a macro, except one around point, because of what
2817 ;; `c-neutralize-syntax-in-CPP' has done.
2819 ;; Insert the opening brace/bracket/paren position.
2821 ;; Clear admin stuff for the next more nested part of the scan.
2825 ;; All open p/b/b's at this nesting level, if any, have probably
2826 ;; been closed by matching/mismatching ones. We're probably
2827 ;; finished - we just need to check for having found an
2828 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2829 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2832 ;; Record the final, innermost, brace-pair if there is one.
2835 ;; Determine a good pos
2843 here-bol))
2852 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2853 ;; not be in it when it is amended for position HERE. This may involve
2854 ;; replacing a CONS element for a brace pair containing HERE with its car.
2855 ;; Additionally, the "outermost" open-brace entry before HERE will be
2856 ;; converted to a cons if the matching close-brace is below HERE.
2857 ;;
2858 ;; START-POINT is a "maximal" "safe position" - there must be no open
2859 ;; parens/braces/brackets between START-POINT and HERE.
2860 ;;
2861 ;; As a second thing, calculate the result of parse-partial-sexp at
2862 ;; PPS-POINT, w.r.t. START-POINT. The motivation here is that
2863 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2864 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2865 ;; needs to be FAST).
2866 ;;
2867 ;; Return a list (GOOD-POS SCAN-BACK-POS CONS-SEPARATED PPS-STATE), where
2868 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2869 ;; to be good (we aim for this to be as high as possible);
2870 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2871 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2872 ;; position to scan backwards from. It is the position of the "{" of the
2873 ;; last element to be removed from `c-state-cache', when that elt is a
2874 ;; cons, otherwise nil.
2875 ;; o - CONS-SEPARATED is t when a cons element in `c-state-cache' has been
2876 ;; replaced by its car because HERE lies inside the brace pair represented
2877 ;; by the cons.
2878 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2888 ; start-point or nil
2895 ; HERE), or nil.
2901 pps-state ; Will be 9 or 10 elements long.
2902 pos
2903 upper-lim ; ,beyond which `c-state-cache' entries are removed
2904 scan-back-pos
2905 cons-separated
2906 pair-beg pps-point-state target-depth)
2908 ;; Remove entries beyond HERE. Also remove any entries inside
2909 ;; a macro, unless HERE is in the same macro.
2914 here
2920 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2921 ;; RBrace and indicate we'll need to search backwards for a previous
2922 ;; brace pair.
2928 cons-separated t))
2930 ;; The next loop jumps forward out of a nested level of parens each
2931 ;; time round; the corresponding elements in `c-state-cache' are
2932 ;; removed. `pos' is just after the brace-pair or the open paren at
2933 ;; (car c-state-cache). There can be no open parens/braces/brackets
2934 ;; between `start-point'/`start-point-actual-macro-start' and HERE,
2935 ;; due to the interface spec to this function.
2938 in-macro-start)))
2940 ; marked by a `category' text property.
2941 start-point))
2952 ;; Do nothing when we've merely reached pps-point.
2953 )
2955 ;; Scan!
2959 target-depth
2960 nil pps-state))
2970 ;; We've just moved out of the paren pair containing the brace-pair
2971 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2972 ;; and is potentially where the open brace of a cons in
2973 ;; c-state-cache will be.
2979 ; containing Lparen
2983 ; together with enclosed brace pair.
2984 ;; (t nil) ; Ignore an unmated Rparen.
2985 )))
2989 nil nil ; TARGETDEPTH, STOPBEFORE
2990 pps-state)))
2992 ;; If the last paren pair we moved out of was actually a brace pair,
2993 ;; insert it into `c-state-cache'.
2998 c-state-cache)))
3003 ;; Strip stale elements of `c-state-cache' by moving backwards through the
3004 ;; buffer, and inform the caller of the scenario detected.
3005 ;;
3006 ;; HERE is the position we're setting `c-state-cache' for.
3007 ;; CACHE-POS (a locally bound variable) is just after the latest recorded
3008 ;; position in `c-state-cache' before HERE, or a position at or near
3009 ;; point-min which isn't in a literal.
3010 ;;
3011 ;; This function must only be called only when (> `c-state-cache-good-pos'
3012 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
3013 ;; optimized to eliminate (or minimize) scanning between these two
3014 ;; positions.
3015 ;;
3016 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
3017 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
3018 ;; could become so after missing elements are inserted into
3019 ;; `c-state-cache'. This is JUST AFTER an opening or closing
3020 ;; brace/paren/bracket which is already in `c-state-cache' or just before
3021 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
3022 ;; before `here''s line, or the start of the literal containing it.
3023 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
3024 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
3025 ;; to scan backwards from.
3026 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
3027 ;; POS and HERE which aren't recorded in `c-state-cache'.
3028 ;;
3029 ;; The comments in this defun use "paren" to mean parenthesis or square
3030 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
3031 ;;
3032 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
3033 ;; | | | | | |
3034 ;; CP E here D C good
3037 pa ren ; positions of "(" and ")"
3038 dropped-cons ; whether the last element dropped from `c-state-cache'
3039 ; was a cons (representing a brace-pair)
3040 good-pos ; see above.
3041 lit ; (START . END) of a literal containing some point.
3042 here-lit-start here-lit-end ; bounds of literal containing `here'
3043 ; or `here' itself.
3048 ;; Remove completely irrelevant entries from `c-state-cache'.
3054 ;; At this stage, (> pos here);
3055 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3060 ;; CASE 1: The top of the cache is a brace pair which now encloses
3061 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3062 ;; knowledge of what's inside these braces, we have no alternative but
3063 ;; to direct the caller to scan the buffer from the opening brace.
3067 ; entry into a { entry, so the caller needs to
3068 ; search for a brace pair before the {.
3070 ;; `here' might be inside a literal. Check for this.
3075 ;; Has `here' just "newly entered" a macro?
3088 ;; `here' might be nested inside any depth of parens (or brackets but
3089 ;; not braces). Scan backwards to find the outermost such opening
3090 ;; paren, if there is one. This will be the scan position to return.
3097 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3098 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3099 ;; a brace pair preceding this, it will already be in `c-state-cache',
3100 ;; unless there was a brace pair after it, i.e. there'll only be one to
3101 ;; scan for if we've just deleted one.
3104 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3105 ;; Further forward scanning isn't needed, but we still need to find a
3106 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3116 ;; CASE 3: After a }/)/] before `here''s BOL.
3120 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3121 ;; literal containing it.
3126 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3127 ;; Externally visible routines.
3131 c-state-cache-good-pos 1
3132 c-state-nonlit-pos-cache nil
3133 c-state-nonlit-pos-cache-limit 1
3134 c-state-semi-nonlit-pos-cache nil
3135 c-state-semi-nonlit-pos-cache-limit 1
3136 c-state-brace-pair-desert nil
3137 c-state-point-min 1
3138 c-state-point-min-lit-type nil
3139 c-state-point-min-lit-start nil
3140 c-state-min-scan-pos 1
3141 c-state-old-cpp-beg nil
3142 c-state-old-cpp-end nil)
3145 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3146 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3147 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3148 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3149 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3150 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3151 ;; (defun c-state-dump ()
3152 ;; ;; For debugging.
3153 ;; ;(message
3154 ;; (concat
3155 ;; (c-sc-qde c-state-cache)
3156 ;; (c-sc-de c-state-cache-good-pos)
3157 ;; (c-sc-qde c-state-nonlit-pos-cache)
3158 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3159 ;; (c-sc-qde c-state-brace-pair-desert)
3160 ;; (c-sc-de c-state-point-min)
3161 ;; (c-sc-de c-state-point-min-lit-type)
3162 ;; (c-sc-de c-state-point-min-lit-start)
3163 ;; (c-sc-de c-state-min-scan-pos)
3164 ;; (c-sc-de c-state-old-cpp-beg)
3165 ;; (c-sc-de c-state-old-cpp-end)))
3166 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3169 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3170 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3171 ;; left in a consistent state.
3172 ;;
3173 ;; This is much like `c-whack-state-after', but it never changes a paren
3174 ;; pair element into an open paren element. Doing that would mean that the
3175 ;; new open paren wouldn't have the required preceding paren pair element.
3176 ;;
3177 ;; This function is called from c-after-change.
3179 ;; The caches of non-literals:
3180 ;; Note that we use "<=" for the possibility of the second char of a two-char
3181 ;; comment opener being typed; this would invalidate any cache position at
3182 ;; HERE.
3188 ;; `c-state-cache':
3189 ;; Case 1: if `here' is in a literal containing point-min, everything
3190 ;; becomes (or is already) nil.
3194 c-state-cache-good-pos nil
3195 c-state-min-scan-pos nil)
3197 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3198 ;; below `here'. To maintain its consistency, we may need to insert a new
3199 ;; brace pair.
3202 too-high-pa ; recorded {/(/[ next above here, or nil.
3203 dropped-cons ; was the last removed element a brace pair?
3204 pa)
3205 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3212 ;; Do we need to add in an earlier brace pair, having lopped one off?
3219 ;; The brace-pair desert marker:
3227 ;; Find and record all noteworthy parens between some good point earlier in
3228 ;; the file and point. That good point is at least the beginning of the
3229 ;; top-level construct we are in, or the beginning of the preceding
3230 ;; top-level construct if we aren't in one.
3231 ;;
3232 ;; The returned value is a list of the noteworthy parens with the last one
3233 ;; first. If an element in the list is an integer, it's the position of an
3234 ;; open paren (of any type) which has not been closed before the point. If
3235 ;; an element is a cons, it gives the position of a closed BRACE paren
3236 ;; pair[*]; the car is the start brace position and the cdr is the position
3237 ;; following the closing brace. Only the last closed brace paren pair
3238 ;; before each open paren and before the point is recorded, and thus the
3239 ;; state never contains two cons elements in succession. When a close brace
3240 ;; has no matching open brace (e.g., the matching brace is outside the
3241 ;; visible region), it is not represented in the returned value.
3242 ;;
3243 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3244 ;; This defun explicitly treats mismatching parens/braces/brackets as
3245 ;; matching. It is the open brace which makes it a "brace" pair.
3246 ;;
3247 ;; If POINT is within a macro, open parens and brace pairs within
3248 ;; THIS macro MIGHT be recorded. This depends on whether their
3249 ;; syntactic properties have been suppressed by
3250 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3251 ;;
3252 ;; Currently no characters which are given paren syntax with the
3253 ;; syntax-table property are recorded, i.e. angle bracket arglist
3254 ;; parens are never present here. Note that this might change.
3255 ;;
3256 ;; BUG: This function doesn't cope entirely well with unbalanced
3257 ;; parens in macros. (2008-12-11: this has probably been resolved
3258 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3259 ;; following case the brace before the macro isn't balanced with the
3260 ;; one after it:
3261 ;;
3262 ;; {
3263 ;; #define X {
3264 ;; }
3265 ;;
3266 ;; Note to maintainers: this function DOES get called with point
3267 ;; within comments and strings, so don't assume it doesn't!
3268 ;;
3269 ;; This function might do hidden buffer changes.
3272 open-paren-in-column-0-is-defun-start
3273 strategy ; 'forward, 'backward etc..
3274 ;; Candidate positions to start scanning from:
3275 cache-pos ; highest position below HERE already existing in
3276 ; cache (or 1).
3277 good-pos
3278 start-point ; (when scanning forward) a place below HERE where there
3279 ; are no open parens/braces between it and HERE.
3280 bopl-state
3281 res
3282 cons-separated
3284 ;; If POINT-MIN has changed, adjust the cache
3288 ;; Strategy?
3293 ;; SCAN!
3301 ; start-point)
3311 good-pos)))
3323 good-pos)))
3327 c-state-cache-good-pos nil))))
3329 c-state-cache)
3332 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3333 ;;
3334 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3335 ;; of all parens in preprocessor constructs, except for any such construct
3336 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3337 ;; worrying further about macros and template delimiters.
3342 c-state-old-cpp-beg
3349 ;; If PLACE is non-nil, return a marker marking it, otherwise nil.
3350 ;; We (re)use MARKER.
3356 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3357 ;; description of the functionality and return value.
3358 ;;
3359 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3360 ;; of all parens in preprocessor constructs, except for any such construct
3361 ;; containing point. We can then call `c-parse-state-1' without worrying
3362 ;; further about macros and template delimiters.
3369 here-cpp-beg)
3371 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3372 ;; subsystem.
3377 here-cpp-beg here-cpp-end
3383 c-state-old-cpp-end
3386 ;; Debug tool to catch cache inconsistencies. This is called from
3387 ;; 000tests.el.
3404 val))))
3406 c-state-cache-good-pos
3407 c-state-nonlit-pos-cache
3408 c-state-nonlit-pos-cache-limit
3409 c-state-semi-nonlit-pos-cache
3410 c-state-semi-nonlit-pos-cache-limit
3411 c-state-brace-pair-desert
3412 c-state-point-min
3413 c-state-point-min-lit-type
3414 c-state-point-min-lit-start
3415 c-state-min-scan-pos
3416 c-state-old-cpp-beg
3417 c-state-old-cpp-end
3418 c-parse-state-point))))
3436 conses-not-ok))
3453 ;; The cache can actually go further back due to the ad-hoc way
3454 ;; the first paren is found, so try to whack off a bit of its
3455 ;; start before complaining.
3456 ;; (save-excursion
3457 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3458 ;; (c-beginning-of-defun-1)
3459 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3460 ;; (c-beginning-of-defun-1))
3461 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3462 ;; (message (concat "c-parse-state inconsistency at %s: "
3463 ;; "using cache: %s, from scratch: %s")
3464 ;; here res1 res2)))
3467 here res1 res2)
3473 here res1)
3478 res2 ; res1 correct a cascading series of errors ASAP
3479 ))
3485 'c-debug-parse-state
3493 \f
3495 ;; Whack off any state information from PAREN-STATE which lies
3496 ;; before BUFPOS. Not destructive on PAREN-STATE.
3499 car)
3510 ;; Whack off any state information from PAREN-STATE which lies at or
3511 ;; after BUFPOS. Not destructive on PAREN-STATE.
3516 ;; just check the car, because in a balanced brace
3517 ;; expression, it must be impossible for the corresponding
3518 ;; close brace to be before point, but the open brace to
3519 ;; be after.
3521 nil ; whack it off
3523 ;; its possible that the open brace is before
3524 ;; bufpos, but the close brace is after. In that
3525 ;; case, convert this to a non-cons element. The
3526 ;; rest of the state is before bufpos, so we're
3527 ;; done.
3529 ;; we know that both the open and close braces are
3530 ;; before bufpos, so we also know that everything else
3531 ;; on state is before bufpos.
3534 nil ; whack it off
3535 ;; it's before bufpos, so everything else should too.
3538 nil)))
3541 ;; Return the bufpos of the innermost enclosing open paren before
3542 ;; bufpos, or nil if none was found.
3552 enclosingp))
3555 ;; Return the bufpos of the outermost enclosing open paren, or nil
3556 ;; if none was found.
3563 pos))
3566 ;; Return the closest "safe" position recorded on PAREN-STATE that
3567 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3568 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3569 ;; find the closest limit before a given limit that might be nil.
3570 ;;
3571 ;; A "safe" position is a position at or after a recorded open
3572 ;; paren, or after a recorded close paren. The returned position is
3573 ;; thus either the first position after a close brace, or the first
3574 ;; position after an enclosing paren, or at the enclosing paren in
3575 ;; case BUFPOS is immediately after it.
3585 ;; See below.
3588 ;; elem is the position at and not after the opening paren, so
3589 ;; we can go forward one more step unless it's equal to
3590 ;; bufpos. This is useful in some cases avoid an extra paren
3591 ;; level between the safe position and bufpos.
3596 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3597 ;; goes to the closest previous point that is known to be outside
3598 ;; any string literal or comment. `c-state-cache' is used if it has
3599 ;; a position in the vicinity.
3601 elem
3604 ;; Note: Similar code in `c-safe-position'. The
3605 ;; difference is that we accept a safe position at
3606 ;; the point and don't bother to go forward past open
3607 ;; parens.
3622 ;; The position is far back. Try `c-beginning-of-defun-1'
3623 ;; (although we can't be entirely sure it will go to a position
3624 ;; outside a comment or string in current emacsen). FIXME:
3625 ;; Consult `syntax-ppss' here.
3630 \f
3631 ;; Tools for scanning identifiers and other tokens.
3634 "Return non-nil if the point is on or directly after an identifier.
3635 Keywords are recognized and not considered identifiers. If an
3636 identifier is detected, the returned value is its starting position.
3637 If an identifier ends at the point and another begins at it \(can only
3638 happen in Pike) then the point for the preceding one is returned.
3640 Note that this function might do hidden buffer changes. See the
3641 comment at the start of cc-engine.el for more info."
3643 ;; FIXME: Shouldn't this function handle "operator" in C++?
3650 ;; Check for a normal (non-keyword) identifier.
3656 ;; Handle the `<operator> syntax in Pike.
3660 t
3667 ;; Handle the "operator +" syntax in C++.
3684 )))
3687 ;; If the point is at the end of a symbol then skip backward to the
3688 ;; beginning of it. Don't move otherwise. Return non-nil if point
3689 ;; moved.
3690 ;;
3691 ;; This function might do hidden buffer changes.
3694 ;; Handle the `<operator> syntax in Pike.
3700 t
3702 nil)))))
3705 ;; Move to the beginning of the current token. Do not move if not
3706 ;; in the middle of one. BACK-LIMIT may be used to bound the
3707 ;; backward search; if given it's assumed to be at the boundary
3708 ;; between two tokens. Return non-nil if the point is moved, nil
3709 ;; otherwise.
3710 ;;
3711 ;; This function might do hidden buffer changes.
3718 ;; `c-nonsymbol-token-regexp' should always match
3719 ;; since we've skipped backward over punctuation
3720 ;; or paren syntax, but consume one char in case
3721 ;; it doesn't so that we don't leave point before
3722 ;; some earlier incorrect token.
3729 ;; Move to the end of the current token. Do not move if not in the
3730 ;; middle of one. BACK-LIMIT may be used to bound the backward
3731 ;; search; if given it's assumed to be at the boundary between two
3732 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3733 ;;
3734 ;; This function might do hidden buffer changes.
3742 ;; `c-nonsymbol-token-regexp' should always match since
3743 ;; we've skipped backward over punctuation or paren
3744 ;; syntax, but move forward in case it doesn't so that
3745 ;; we don't leave point earlier than we started with.
3761 "Move forward by tokens.
3762 A token is defined as all symbols and identifiers which aren't
3764 treated properly). Point is always either left at the beginning of a
3765 token or not moved at all. COUNT specifies the number of tokens to
3766 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3767 moves to the next token beginning only if not already at one. If
3768 BALANCED is true, move over balanced parens, otherwise move into them.
3769 Also, if BALANCED is true, never move out of an enclosing paren.
3771 LIMIT sets the limit for the movement and defaults to the point limit.
3772 The case when LIMIT is set in the middle of a token, comment or macro
3773 is handled correctly, i.e. the point won't be left there.
3775 Return the number of tokens left to move \(positive or negative). If
3776 BALANCED is true, a move over a balanced paren counts as one. Note
3777 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3778 be returned. Thus, a return value of 0 guarantees that point is at
3779 the requested position and a return value less \(without signs) than
3780 COUNT guarantees that point is at the beginning of some token.
3782 Note that this function might do hidden buffer changes. See the
3783 comment at the start of cc-engine.el for more info."
3790 c-jump-syntax-balanced
3791 c-jump-syntax-unbalanced))
3796 ;; If count is zero we should jump if in the middle of a token.
3803 ;; Skip whitespace. Count this as a move if we did in
3804 ;; fact move.
3808 ;; Moved out of bounds. Make sure the returned count isn't zero.
3813 ;; Use `condition-case' to avoid having the limit tests
3814 ;; inside the loop.
3822 t)
3825 t)
3826 ;; `c-nonsymbol-token-regexp' above should always
3827 ;; match if there are correct tokens. Try to
3828 ;; widen to see if the limit was set in the
3829 ;; middle of one, else fall back to treating
3830 ;; the offending thing as a one character token.
3835 nil)
3838 t))))
3848 count)))
3851 "Move backward by tokens.
3852 See `c-forward-token-2' for details."
3860 c-jump-syntax-balanced
3861 c-jump-syntax-unbalanced))
3865 ;; The count is zero so try to skip to the beginning of the
3866 ;; current token.
3870 ;; The limit is inside the same token, so return 1.
3873 ;; We're not in the middle of a token. If there's
3874 ;; whitespace after the point then we must move backward,
3875 ;; so set count to 1 in that case.
3877 ;; If we're looking at a '#' that might start a cpp
3878 ;; directive then we have to do a more elaborate check.
3891 ;; Use `condition-case' to avoid having to check for buffer
3892 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3901 ;; This can be very inefficient if there's a long
3902 ;; sequence of operator tokens without any separation.
3903 ;; That doesn't happen in practice, anyway.
3913 count)))
3916 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3918 characters are jumped over character by character. This function is
3919 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3924 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3926 characters are jumped over character by character. This function is
3927 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3931 \f
3932 ;; Tools for doing searches restricted to syntactically relevant text.
3935 paren-level not-inside-token
3936 lookbehind-submatch)
3937 "Like `re-search-forward', but only report matches that are found
3938 in syntactically significant text. I.e. matches in comments, macros
3939 or string literals are ignored. The start point is assumed to be
3940 outside any comment, macro or string literal, or else the content of
3941 that region is taken as syntactically significant text.
3943 If PAREN-LEVEL is non-nil, an additional restriction is added to
3944 ignore matches in nested paren sexps. The search will also not go
3945 outside the current list sexp, which has the effect that if the point
3946 should be moved to BOUND when no match is found \(i.e. NOERROR is
3947 neither nil nor t), then it will be at the closing paren if the end of
3948 the current list sexp is encountered first.
3950 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3951 ignored. Things like multicharacter operators and special symbols
3953 constants.
3955 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3956 subexpression in REGEXP. The end of that submatch is used as the
3957 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
3958 isn't used or if that subexpression didn't match then the start
3960 subexpression is never tested before the starting position, so it
3964 subexpression can match the end of nonwhite syntactic whitespace,
3965 i.e. the end of comments or cpp directives. This since the function
3966 skips over such things before resuming the search. It's on the other
3968 matches syntactic whitespace.
3970 Bug: Unbalanced parens inside cpp directives are currently not handled
3971 correctly \(i.e. they don't get ignored as they should) when
3972 PAREN-LEVEL is set.
3974 Note that this function might do hidden buffer changes. See the
3975 comment at the start of cc-engine.el for more info."
3980 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
3983 tmp
3984 ;; Start position for the last search.
3985 search-pos
3986 ;; The `parse-partial-sexp' state between the start position
3987 ;; and the point.
3988 state
3989 ;; The current position after the last state update. The next
3990 ;; `parse-partial-sexp' continues from here.
3992 ;; The position at which to check the state and the state
3993 ;; there. This is separate from `state-pos' since we might
3994 ;; need to back up before doing the next search round.
3995 check-pos check-state
3996 ;; Last position known to end a token.
3998 ;; Set when a valid match is found.
3999 found)
4017 state-pos check-pos paren-level nil state))
4019 check-state state))
4021 ;; NOTE: If we got a look behind subexpression and get
4022 ;; an insignificant match in something that isn't
4023 ;; syntactic whitespace (i.e. strings or in nested
4024 ;; parentheses), then we can never skip more than a
4025 ;; single character from the match start position
4026 ;; (i.e. `state-pos' here) before continuing the
4027 ;; search. That since the look behind subexpression
4028 ;; might match the end of the insignificant region in
4029 ;; the next search.
4033 ;; Match inside a line comment. Skip to eol. Use
4034 ;; `re-search-forward' instead of `skip-chars-forward' to get
4035 ;; the right bound behavior.
4039 ;; Match inside a block comment. Skip to the '*/'.
4046 ;; Match in the middle of the opener of a block or line
4047 ;; comment.
4052 ;; The last `parse-partial-sexp' above might have
4053 ;; stopped short of the real check position if the end
4054 ;; of the current sexp was encountered in paren-level
4055 ;; mode. The checks above are always false in that
4056 ;; case, and since they can do better skipping in
4057 ;; lookbehind-submatch mode, we do them before
4058 ;; checking the paren level.
4062 ;; Check the paren level first since we're short of the
4063 ;; syntactic checking position if the end of the
4064 ;; current sexp was encountered by `parse-partial-sexp'.
4067 ;; Inside a nested paren sexp.
4069 ;; See the NOTE above.
4071 ;; Skip out of the paren quickly.
4075 ;; Have exited the current paren sexp.
4078 ;; The last `parse-partial-sexp' call above
4079 ;; has left us just after the closing paren
4080 ;; in this case, so we can modify the bound
4081 ;; to leave the point at the right position
4082 ;; upon return.
4084 nil)
4088 ;; Match inside a string.
4091 ;; See the NOTE above.
4093 ;; Skip to the end of the string before continuing.
4102 continue)))
4107 ;; Match inside a macro. Skip to the end of it.
4115 (< check-pos
4121 ;; Inside a token.
4123 ;; See the NOTE above.
4128 ;; A real match.
4130 nil)))
4132 ;; Should loop to search again, but take care to avoid
4133 ;; looping on the same spot.
4137 nil
4140 t))))
4146 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4153 ;; Search failed. Set point as appropriate.
4157 nil)))
4162 ;; Return the start of the literal point is in, or nil.
4163 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4164 ;; bound in the caller.
4166 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4167 ;; if it's outside comments and strings.
4170 ;; Pick a safe position as close to the point as possible.
4171 ;;
4172 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4173 ;; position.
4186 ;; Cache positions along the way to use if we have to back up more. We
4187 ;; cache every closing paren on the same level. If the paren cache is
4188 ;; relevant in this region then we're typically already on the same
4189 ;; level as the target position. Note that we might cache positions
4190 ;; after opening parens in case safe-pos is in a nested list. That's
4191 ;; both uncommon and harmless.
4199 ;; If the state contains the start of the containing sexp we cache that
4200 ;; position too, so that parse-partial-sexp in the next run has a bigger
4201 ;; chance of starting at the same level as the target position and thus
4202 ;; will get more good safe positions into the list.
4208 ;; Inside string or comment. Continue search at the
4209 ;; beginning of it.
4213 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4214 i.e. don't stop at positions inside syntactic whitespace or string
4215 literals. Preprocessor directives are also ignored, with the exception
4216 of the one that the point starts within, if any. If LIMIT is given,
4217 it's assumed to be at a syntactically relevant position.
4219 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4220 sexps, and the search will also not go outside the current paren sexp.
4221 However, if LIMIT or the buffer limit is reached inside a nested paren
4222 then the point will be left at the limit.
4224 Non-nil is returned if the point moved, nil otherwise.
4226 Note that this function might do hidden buffer changes. See the
4227 comment at the start of cc-engine.el for more info."
4230 state-2
4231 ;; A list of syntactically relevant positions in descending
4232 ;; order. It's used to avoid scanning repeatedly over
4233 ;; potentially large regions with `parse-partial-sexp' to verify
4234 ;; each position. Used in `c-ssb-lit-begin'
4235 safe-pos-list
4236 ;; The result from `c-beginning-of-macro' at the start position or the
4237 ;; start position itself if it isn't within a macro. Evaluated on
4238 ;; demand.
4239 start-macro-beg
4240 ;; The earliest position after the current one with the same paren
4241 ;; level. Used only when `paren-level' is set.
4242 lit-beg
4247 ;; The next loop "tries" to find the end point each time round,
4248 ;; loops when it hasn't succeeded.
4254 ;; Don't stop inside a literal.
4257 t)))
4269 ;; Not at the right level.
4272 ;; We stop above if we go out of a paren.
4273 ;; Now check whether it precedes or is
4274 ;; nested in the starting sexp.
4278 pps-end-pos paren-level-pos
4279 nil nil state-2))
4282 ;; We've stopped short of the starting position
4283 ;; so the hit was inside a nested list. Go up
4284 ;; until we are at the right level.
4293 nil)
4294 t))
4297 nil))
4299 ;; The hit was outside the list at the start
4300 ;; position. Go to the start of the list and exit.
4302 nil))
4305 ;; Inside a macro.
4313 t
4315 ;; It's inside the same macro we started in so it's
4316 ;; a relevant match.
4318 nil))))))
4322 ;; Skip syntactic ws afterwards so that we don't stop at the
4323 ;; end of a comment if `skip-chars' is something like "^/".
4327 ;; We might want to extend this with more useful return values in
4328 ;; the future.
4331 ;; The following is an alternative implementation of
4332 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4333 ;; track of the syntactic context. It turned out to be generally
4334 ;; slower than the one above which uses forward checks from earlier
4335 ;; safe positions.
4336 ;;
4337 ;;(defconst c-ssb-stop-re
4338 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4339 ;; ;; stop at to avoid going into comments and literals.
4340 ;; (concat
4341 ;; ;; Match comment end syntax and string literal syntax. Also match
4342 ;; ;; '/' for block comment endings (not covered by comment end
4343 ;; ;; syntax).
4344 ;; "\\s>\\|/\\|\\s\""
4345 ;; (if (memq 'gen-string-delim c-emacs-features)
4346 ;; "\\|\\s|"
4347 ;; "")
4348 ;; (if (memq 'gen-comment-delim c-emacs-features)
4349 ;; "\\|\\s!"
4350 ;; "")))
4351 ;;
4352 ;;(defconst c-ssb-stop-paren-re
4353 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4354 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4355 ;;
4356 ;;(defconst c-ssb-sexp-end-re
4357 ;; ;; Regexp matching the ending syntax of a complex sexp.
4358 ;; (concat c-string-limit-regexp "\\|\\s)"))
4359 ;;
4360 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4361 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4362 ;;i.e. don't stop at positions inside syntactic whitespace or string
4363 ;;literals. Preprocessor directives are also ignored. However, if the
4364 ;;point is within a comment, string literal or preprocessor directory to
4365 ;;begin with, its contents is treated as syntactically relevant chars.
4366 ;;If LIMIT is given, it limits the backward search and the point will be
4367 ;;left there if no earlier position is found.
4368 ;;
4369 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4370 ;;sexps, and the search will also not go outside the current paren sexp.
4371 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4372 ;;then the point will be left at the limit.
4373 ;;
4374 ;;Non-nil is returned if the point moved, nil otherwise.
4375 ;;
4376 ;;Note that this function might do hidden buffer changes. See the
4377 ;;comment at the start of cc-engine.el for more info."
4378 ;;
4379 ;; (save-restriction
4380 ;; (when limit
4381 ;; (narrow-to-region limit (point-max)))
4382 ;;
4383 ;; (let ((start (point)))
4384 ;; (catch 'done
4385 ;; (while (let ((last-pos (point))
4386 ;; (stop-pos (progn
4387 ;; (skip-chars-backward skip-chars)
4388 ;; (point))))
4389 ;;
4390 ;; ;; Skip back over the same region as
4391 ;; ;; `skip-chars-backward' above, but keep to
4392 ;; ;; syntactically relevant positions.
4393 ;; (goto-char last-pos)
4394 ;; (while (and
4395 ;; ;; `re-search-backward' with a single char regexp
4396 ;; ;; should be fast.
4397 ;; (re-search-backward
4398 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4399 ;; stop-pos 'move)
4400 ;;
4401 ;; (progn
4402 ;; (cond
4403 ;; ((looking-at "\\s(")
4404 ;; ;; `paren-level' is set and we've found the
4405 ;; ;; start of the containing paren.
4406 ;; (forward-char)
4407 ;; (throw 'done t))
4408 ;;
4409 ;; ((looking-at c-ssb-sexp-end-re)
4410 ;; ;; We're at the end of a string literal or paren
4411 ;; ;; sexp (if `paren-level' is set).
4412 ;; (forward-char)
4413 ;; (condition-case nil
4414 ;; (c-backward-sexp)
4415 ;; (error
4416 ;; (goto-char limit)
4417 ;; (throw 'done t))))
4418 ;;
4419 ;; (t
4420 ;; (forward-char)
4421 ;; ;; At the end of some syntactic ws or possibly
4422 ;; ;; after a plain '/' operator.
4423 ;; (let ((pos (point)))
4424 ;; (c-backward-syntactic-ws)
4425 ;; (if (= pos (point))
4426 ;; ;; Was a plain '/' operator. Go past it.
4427 ;; (backward-char)))))
4428 ;;
4429 ;; (> (point) stop-pos))))
4430 ;;
4431 ;; ;; Now the point is either at `stop-pos' or at some
4432 ;; ;; position further back if `stop-pos' was at a
4433 ;; ;; syntactically irrelevant place.
4434 ;;
4435 ;; ;; Skip additional syntactic ws so that we don't stop
4436 ;; ;; at the end of a comment if `skip-chars' is
4437 ;; ;; something like "^/".
4438 ;; (c-backward-syntactic-ws)
4439 ;;
4440 ;; (< (point) stop-pos))))
4441 ;;
4442 ;; ;; We might want to extend this with more useful return values
4443 ;; ;; in the future.
4444 ;; (/= (point) start))))
4446 \f
4447 ;; Tools for handling comments and string literals.
4450 "Return the type of literal point is in, if any.
4451 The return value is `c' if in a C-style comment, `c++' if in a C++
4452 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4453 is non-nil and in a preprocessor line, or nil if somewhere else.
4454 Optional LIM is used as the backward limit of the search. If omitted,
4455 or nil, `c-beginning-of-defun' is used.
4457 The last point calculated is cached if the cache is enabled, i.e. if
4458 `c-in-literal-cache' is bound to a two element vector.
4460 Note that this function might do hidden buffer changes. See the
4461 comment at the start of cc-engine.el for more info."
4472 "Return a cons of the beginning and end positions of the comment or
4473 string surrounding point (including both delimiters), or nil if point
4475 to start parsing from. If NEAR is non-nil, then the limits of any
4477 spaces and tabs between point and the literal. The search for such a
4478 literal is done first in forward direction. If NOT-IN-DELIMITER is
4479 non-nil, the case when point is inside a starting delimiter won't be
4480 recognized. This only has effect for comments which have starting
4481 delimiters with more than one character.
4483 Note that this function might do hidden buffer changes. See the
4484 comment at the start of cc-engine.el for more info."
4500 ;; Search forward for a literal.
4511 ;; Search backward.
4522 ;; Block comment. Due to the nature of line
4523 ;; comments, they will always be covered by the
4524 ;; normal case above.
4527 ;; If LIM is bogus, beg will be bogus.
4531 ))))
4533 ;; In case external callers use this; it did have a docstring.
4537 "If the argument is a cons of two buffer positions (such as returned by
4538 `c-literal-limits'), and that range contains a C++ style line comment,
4539 then an extended range is returned that contains all adjacent line
4540 comments (i.e. all comments that starts in the same column with no
4541 empty lines or non-whitespace characters between them). Otherwise the
4542 argument is returned.
4544 Note that this function might do hidden buffer changes. See the
4545 comment at the start of cc-engine.el for more info."
4556 ;; Got to take care in the backward direction to handle
4557 ;; comments which are preceded by code.
4571 range)
4575 "Convenience function that given the result of `c-literal-limits',
4576 returns nil or the type of literal that the range surrounds, one
4577 of the symbols 'c, 'c++ or 'string. It's much faster than using
4578 `c-in-literal' and is intended to be used when you need both the
4579 type of a literal and its limits.
4581 Note that this function might do hidden buffer changes. See the
4582 comment at the start of cc-engine.el for more info."
4593 range))
4596 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4597 ;; This doesn't preserve point.
4606 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4607 ;; (default point). This is done by going back further in the buffer then
4608 ;; searching forward for literals. The position found won't be in a
4609 ;; literal. We start searching for the sought position TRY-SIZE (default
4610 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4611 ;; :-)
4619 stack elt size
4622 ;; Move forward one literal each time round this loop.
4623 ;; Move forward to the start of a comment or string.
4625 pos
4626 start
4627 nil ; target-depth
4628 nil ; stop-before
4629 s ; state
4632 ;; Gather details of the non-literal-bit - starting pos and size.
4636 pos))
4640 ;; Move forward to the end of the comment/string.
4644 start
4645 nil ; target-depth
4646 nil ; stop-before
4647 s ; state
4651 ;; Now try and find enough non-literal characters recorded on the stack.
4652 ;; Go back one recorded literal each time round this loop.
4654 stack)
4659 ;; Have we found enough yet?
4669 ;; Return a buffer position about HOW-FAR non-literal characters forward
4670 ;; from START-POS (default point), which must not be inside a literal.
4677 ;; Scan over counted characters.
4679 pos
4681 nil ; target-depth
4682 nil ; stop-before
4683 s ; state
4687 ;; Scan over literal characters.
4690 pos
4692 nil ; target-depth
4693 nil ; stop-before
4694 s ; state
4699 \f
4700 ;; `c-find-decl-spots' and accompanying stuff.
4702 ;; Variables used in `c-find-decl-spots' to cache the search done for
4703 ;; the first declaration in the last call. When that function starts,
4704 ;; it needs to back up over syntactic whitespace to look at the last
4705 ;; token before the region being searched. That can sometimes cause
4706 ;; moves back and forth over a quite large region of comments and
4707 ;; macros, which would be repeated for each changed character when
4708 ;; we're called during fontification, since font-lock refontifies the
4709 ;; current line for each change. Thus it's worthwhile to cache the
4710 ;; first match.
4711 ;;
4712 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4713 ;; the syntactic whitespace less or equal to some start position.
4714 ;; There's no cached value if it's nil.
4715 ;;
4716 ;; `c-find-decl-match-pos' is the match position if
4717 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4718 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4729 ; (defface c-debug-decl-spot-face
4730 ; '((t (:background "Turquoise")))
4731 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4732 ; (defface c-debug-decl-sws-face
4733 ; '((t (:background "Khaki")))
4734 ; "Debug face to mark the syntactic whitespace between the declaration
4735 ; spots and the preceding token end.")
4751 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4752 ;; but it contains lots of free variables that refer to things
4753 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4754 ;; if there is a match, otherwise at `cfd-limit'.
4755 ;;
4756 ;; The macro moves point forward to the next putative start of a declaration
4757 ;; or cfd-limit. This decl start is the next token after a "declaration
4758 ;; prefix". The declaration prefix is the earlier of `cfd-prop-match' and
4759 ;; `cfd-re-match'. `cfd-match-pos' is set to the decl prefix.
4760 ;;
4761 ;; This macro might do hidden buffer changes.
4764 ;; Find the next property match position if we haven't got one already.
4775 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4776 ;; got one already.
4782 ;; Each time round, the next `while' moves forward over a pseudo match
4783 ;; of `c-decl-prefix-or-start-re' which is either inside a literal, or
4784 ;; is a ":" not preceded by "public", etc.. `cfd-re-match' and
4785 ;; `cfd-re-match-end' get set.
4792 ;; No match. Finish up and exit the loop.
4794 nil)
4797 ;; Matched the end of a token preceding a decl spot.
4801 ;; Matched a token that start a decl spot.
4804 c-literal-faces)
4805 ;; Pseudo match inside a comment or string literal. Skip out
4806 ;; of comments and string literals.
4818 ;; Found a ":" which isn't part of "public:", etc.
4819 t)
4822 ;; If our match was at the decl start, we have to back up over the
4823 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4824 ;; any decl spots in the syntactic ws.
4829 ;; Choose whichever match is closer to the start.
4832 cfd-re-match nil)
4834 cfd-prop-match nil))
4839 ;; Skip forward past comments only so we don't skip macros.
4841 ;; Set the position to continue at. We can avoid going over
4842 ;; the comments skipped above a second time, but it's possible
4843 ;; that the comment skipping has taken us past `cfd-prop-match'
4844 ;; since the property might be used inside comments.
4850 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4851 ;; label from the point to CFD-LIMIT.
4852 ;;
4853 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4854 ;; arguments: The first is the end position of the token preceding the spot,
4855 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4856 ;; the match is inside a macro. Point should be moved forward by at least
4857 ;; one token.
4858 ;;
4859 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4860 ;; it should return non-nil to ensure that the next search will find them.
4861 ;;
4862 ;; Such a spot is:
4863 ;; o The first token after bob.
4864 ;; o The first token after the end of submatch 1 in
4865 ;; `c-decl-prefix-or-start-re' when that submatch matches.
4866 ;; o The start of each `c-decl-prefix-or-start-re' match when
4867 ;; submatch 1 doesn't match.
4868 ;; o The first token after the end of each occurrence of the
4869 ;; `c-type' text property with the value `c-decl-end', provided
4870 ;; `c-type-decl-end-used' is set.
4871 ;;
4872 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4873 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4874 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4875 ;;
4876 ;; If the match is inside a macro then the buffer is narrowed to the
4877 ;; end of it, so that CFD-FUN can investigate the following tokens
4878 ;; without matching something that begins inside a macro and ends
4879 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4880 ;; CFD-FACE-CHECKLIST checks exist.
4881 ;;
4882 ;; The spots are visited approximately in order from top to bottom.
4883 ;; It's however the positions where `c-decl-prefix-or-start-re'
4884 ;; matches and where `c-decl-end' properties are found that are in
4885 ;; order. Since the spots often are at the following token, they
4886 ;; might be visited out of order insofar as more spots are reported
4887 ;; later on within the syntactic whitespace between the match
4888 ;; positions and their spots.
4889 ;;
4890 ;; It's assumed that comments and strings are fontified in the
4891 ;; searched range.
4892 ;;
4893 ;; This is mainly used in fontification, and so has an elaborate
4894 ;; cache to handle repeated calls from the same start position; see
4895 ;; the variables above.
4896 ;;
4897 ;; All variables in this function begin with `cfd-' to avoid name
4898 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4899 ;;
4900 ;; This function might do hidden buffer changes.
4904 ;; The end of the token preceding the decl spot last found
4905 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4906 ;; no match.
4907 cfd-re-match
4908 ;; The end position of the last `c-decl-prefix-or-start-re'
4909 ;; match. If this is greater than `cfd-continue-pos', the
4910 ;; next regexp search is started here instead.
4912 ;; The end of the last `c-decl-end' found by
4913 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4914 ;; match. If searching for the property isn't needed then we
4915 ;; disable it by setting it to `cfd-limit' directly.
4917 ;; The end of the token preceding the decl spot last found by
4918 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4919 ;; bob. `cfd-limit' if there's no match. In other words,
4920 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4922 ;; The position to continue searching at.
4923 cfd-continue-pos
4924 ;; The position of the last "real" token we've stopped at.
4925 ;; This can be greater than `cfd-continue-pos' when we get
4926 ;; hits inside macros or at `c-decl-end' positions inside
4927 ;; comments.
4929 ;; The end position of the last entered macro.
4932 ;; Initialize by finding a syntactically relevant start position
4933 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4934 ;; search unless we're at bob.
4937 ;; Must back up a bit since we look for the end of the previous
4938 ;; statement or declaration, which is earlier than the first
4939 ;; returned match.
4942 ;; First we need to move to a syntactically relevant position.
4943 ;; Begin by backing out of comment or string literals.
4946 ;; Try to use the faces to back up to the start of the
4947 ;; literal. FIXME: What if the point is on a declaration
4948 ;; inside a comment?
4954 ;; XEmacs doesn't fontify the quotes surrounding string
4955 ;; literals.
4964 ;; Don't trust the literal to contain only literal faces
4965 ;; (the font lock package might not have fontified the
4966 ;; start of it at all, for instance) so check that we have
4967 ;; arrived at something that looks like a start or else
4968 ;; resort to `c-literal-limits'.
4975 ;; The start is in a literal. If the limit is in the same
4976 ;; one we don't have to find a syntactic position etc. We
4977 ;; only check that if the limit is at or before bonl to save
4978 ;; time; it covers the by far most common case when font-lock
4979 ;; refontifies the current line only.
4990 ;; Completely inside a literal. Set up variables to trig the
4991 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
4992 ;; find a suitable start position.
4995 ;; Check if the region might be completely inside a macro, to
4996 ;; optimize that like the completely-inside-literal above.
5003 ;; (Maybe) completely inside a macro. Only need to trig the
5004 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
5005 ;; set things up.
5007 start-in-macro t))
5010 ;; Back out of any macro so we don't miss any declaration
5011 ;; that could follow after it.
5015 ;; Now we're at a proper syntactically relevant position so we
5016 ;; can use the cache. But first clear it if it applied
5017 ;; further down.
5022 ;; Don't have to do this if the cache is relevant here,
5023 ;; typically if the same line is refontified again. If
5024 ;; we're just some syntactic whitespace further down we can
5025 ;; still use the cache to limit the skipping.
5028 ;; If we hit `c-find-decl-syntactic-pos' and
5029 ;; `c-find-decl-match-pos' is set then we install the cached
5030 ;; values. If we hit `c-find-decl-syntactic-pos' and
5031 ;; `c-find-decl-match-pos' is nil then we know there's no decl
5032 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
5033 ;; and so we can continue the search from this point. If we
5034 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
5035 ;; the right spot to begin searching anyway.
5037 c-find-decl-match-pos)
5039 cfd-continue-pos syntactic-pos)
5044 ;; Always consider bob a match to get the first
5045 ;; declaration in the file. Do this separately instead of
5046 ;; letting `c-decl-prefix-or-start-re' match bob, so that
5047 ;; regexp always can consume at least one character to
5048 ;; ensure that we won't get stuck in an infinite loop.
5053 ;; Do an initial search now. In the bob case above it's
5054 ;; only done to search for a `c-decl-end' spot.
5058 cfd-match-pos)))))
5060 ;; Advance `cfd-continue-pos' if it's before the start position.
5061 ;; The closest continue position that might have effect at or
5062 ;; after the start depends on what we started in. This also
5063 ;; finds a suitable start position in the special cases when the
5064 ;; region is completely within a literal or macro.
5069 ;; If we're in a macro then it's the closest preceding token
5070 ;; in the macro. Check this before `start-in-literal',
5071 ;; since if we're inside a literal in a macro, the preceding
5072 ;; token is earlier than any `c-decl-end' spot inside the
5073 ;; literal (comment).
5075 ;; The only syntactic ws in macros are comments.
5081 ;; If we're in a comment it can only be the closest
5082 ;; preceding `c-decl-end' position within that comment, if
5083 ;; any. Go back to the beginning of such a property so that
5084 ;; `c-find-decl-prefix-search' will find the end of it.
5085 ;; (Can't stop at the end and install it directly on
5086 ;; `cfd-prop-match' since that variable might be cleared
5087 ;; after `cfd-fun' below.)
5088 ;;
5089 ;; Note that if the literal is a string then the property
5090 ;; search will simply skip to the beginning of it right
5091 ;; away.
5103 ;; Didn't find any property inside the comment, so we can
5104 ;; skip it entirely. (This won't skip past a string, but
5105 ;; that'll be handled quickly by the next
5106 ;; `c-find-decl-prefix-search' anyway.)
5112 ;; If we started in normal code, the only match that might
5113 ;; apply before the start is what we already got in
5114 ;; `cfd-match-pos' so we can continue at the start position.
5115 ;; (Note that we don't get here if the first match is below
5116 ;; it.)
5119 ;; Delete found matches if they are before our new continue
5120 ;; position, so that `c-find-decl-prefix-search' won't back up
5121 ;; to them later on.
5129 ;; This is the normal case and we got a proper syntactic
5130 ;; position. If there's a match then it's always outside
5131 ;; macros and comments, so advance to the next token and set
5132 ;; `cfd-token-pos'. The loop below will later go back using
5133 ;; `cfd-continue-pos' to fix declarations inside the
5134 ;; syntactic ws.
5142 ;; Have one of the special cases when the region is completely
5143 ;; within a literal or macro. `cfd-continue-pos' is set to a
5144 ;; good start position for the search, so do it.
5147 ;; Now loop. Round what? (ACM, 2006/7/5). We already got the first match.
5154 ;; Kludge to filter out matches on the "<" that
5155 ;; aren't open parens, for the sake of languages
5156 ;; that got `c-recognize-<>-arglists' set.
5161 ;; If `cfd-continue-pos' is less or equal to
5162 ;; `cfd-token-pos', we've got a hit inside a macro
5163 ;; that's in the syntactic whitespace before the last
5164 ;; "real" declaration we've checked. If they're equal
5165 ;; we've arrived at the declaration a second time, so
5166 ;; there's nothing to do.
5170 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5171 ;; we're still searching for declarations embedded in
5172 ;; the syntactic whitespace. In that case we need
5173 ;; only to skip comments and not macros, since they
5174 ;; can't be nested, and that's already been done in
5175 ;; `c-find-decl-prefix-search'.
5180 ;; Continue if the following token fails the
5181 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5188 t)))
5199 ;; Narrow to the end of the macro if we got a hit inside
5200 ;; one, to avoid recognizing things that start inside the
5201 ;; macro and end outside it.
5203 ;; Not in the same macro as in the previous round.
5214 t
5217 t)
5218 ;; The matched token was the last thing in the macro,
5219 ;; so the whole match is bogus.
5221 nil))))
5228 ;; Restore limits if we did macro narrowing above.
5235 ; cfd-match-pos, etc.
5237 \f
5238 ;; A cache for found types.
5240 ;; Buffer local variable that contains an obarray with the types we've
5241 ;; found. If a declaration is recognized somewhere we record the
5242 ;; fully qualified identifier in it to recognize it as a type
5243 ;; elsewhere in the file too. This is not accurate since we do not
5244 ;; bother with the scoping rules of the languages, but in practice the
5245 ;; same name is seldom used as both a type and something else in a
5246 ;; file, and we only use this as a last resort in ambiguous cases (see
5247 ;; `c-forward-decl-or-cast-1').
5248 ;;
5249 ;; Not every type need be in this cache. However, things which have
5250 ;; ceased to be types must be removed from it.
5251 ;;
5252 ;; Template types in C++ are added here too but with the template
5253 ;; arglist replaced with "<>" in references or "<" for the one in the
5254 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5255 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5256 ;; template specs can be fairly sized programs in themselves) and
5257 ;; improves the hit ratio (it's a type regardless of the template
5258 ;; args; it's just not the same type, but we're only interested in
5259 ;; recognizing types, not telling distinct types apart). Note that
5260 ;; template types in references are added here too; from the example
5261 ;; above there will also be an entry "Foo<".
5266 ;; Clears `c-found-types'.
5270 ;; Add the given region as a type in `c-found-types'. If the region
5271 ;; doesn't match an existing type but there is a type which is equal
5272 ;; to the given one except that the last character is missing, then
5273 ;; the shorter type is removed. That's done to avoid adding all
5274 ;; prefixes of a type as it's being entered and font locked. This
5275 ;; doesn't cover cases like when characters are removed from a type
5276 ;; or added in the middle. We'd need the position of point when the
5277 ;; font locking is invoked to solve this well.
5278 ;;
5279 ;; This function might do hidden buffer changes.
5286 ;; Remove the "NAME" from c-found-types, if present.
5290 ;; Return non-nil if the given region contains a type in
5291 ;; `c-found-types'.
5292 ;;
5293 ;; This function might do hidden buffer changes.
5295 c-found-types))
5298 ;; Return all the types in `c-found-types' as a sorted list of
5299 ;; strings.
5303 type-list)))
5304 c-found-types)
5307 ;; Shut up the byte compiler.
5311 ;; An after change function which, in conjunction with the info in
5312 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5313 ;; from `c-found-types', should this type have become stale. For
5314 ;; example, this happens to "foo" when "foo \n bar();" becomes
5315 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5316 ;; the fontification.
5317 ;;
5318 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5319 ;; type?
5339 ;; Changing the amount of (already existing) whitespace - don't do anything.
5344 ;; The syntactic relationship which defined a "found type" has been
5345 ;; destroyed.
5348 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5349 )))
5351 \f
5352 ;; Setting and removing syntax properties on < and > in languages (C++
5353 ;; and Java) where they can be template/generic delimiters as well as
5354 ;; their normal meaning of "less/greater than".
5356 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5357 ;; be delimiters, they are marked as such with the category properties
5358 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5360 ;; STRATEGY:
5361 ;;
5362 ;; It is impossible to determine with certainty whether a <..> pair in
5363 ;; C++ is two comparison operators or is template delimiters, unless
5364 ;; one duplicates a lot of a C++ compiler. For example, the following
5365 ;; code fragment:
5366 ;;
5367 ;; foo (a < b, c > d) ;
5368 ;;
5369 ;; could be a function call with two integer parameters (each a
5370 ;; relational expression), or it could be a constructor for class foo
5371 ;; taking one parameter d of templated type "a < b, c >". They are
5372 ;; somewhat easier to distinguish in Java.
5373 ;;
5374 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5375 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5376 ;; individually when their context so indicated. This gave rise to
5377 ;; intractable problems when one of a matching pair was deleted, or
5378 ;; pulled into a literal.]
5379 ;;
5380 ;; At each buffer change, the syntax-table properties are removed in a
5381 ;; before-change function and reapplied, when needed, in an
5382 ;; after-change function. It is far more important that the
5383 ;; properties get removed when they they are spurious than that they
5384 ;; be present when wanted.
5385 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5387 ;; POS (default point) is at a < character. If it is marked with
5388 ;; open paren syntax-table text property, remove the property,
5389 ;; together with the close paren property on the matching > (if
5390 ;; any).
5405 ;; POS (default point) is at a > character. If it is marked with
5406 ;; close paren syntax-table property, remove the property, together
5407 ;; with the open paren property on the matching < (if any).
5422 ;; POS (default point) is at a < or > character. If it has an
5423 ;; open/close paren syntax-table property, remove this property both
5424 ;; from the current character and its partner (which will also be
5425 ;; thusly marked).
5435 ;; POS (default point) is at a < character. If it is both marked
5436 ;; with open/close paren syntax-table property, and has a matching >
5437 ;; (also marked) which is after LIM, remove the property both from
5438 ;; the current > and its partner. Return t when this happens, nil
5439 ;; when it doesn't.
5453 t)))
5456 ;; POS (default point) is at a > character. If it is both marked
5457 ;; with open/close paren syntax-table property, and has a matching <
5458 ;; (also marked) which is before LIM, remove the property both from
5459 ;; the current < and its partner. Return t when this happens, nil
5460 ;; when it doesn't.
5474 t)))
5476 ;; Set by c-common-init in cc-mode.el.
5481 ;; Unmark certain pairs of "< .... >" which are currently marked as
5482 ;; template/generic delimiters. (This marking is via syntax-table
5483 ;; text properties).
5484 ;;
5485 ;; These pairs are those which are in the current "statement" (i.e.,
5486 ;; the region between the {, }, or ; before BEG and the one after
5487 ;; END), and which enclose any part of the interval (BEG END).
5488 ;;
5489 ;; Note that in C++ (?and Java), template/generic parens cannot
5490 ;; enclose a brace or semicolon, so we use these as bounds on the
5491 ;; region we must work on.
5492 ;;
5493 ;; This function is called from before-change-functions (via
5494 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5495 ;; and point is undefined, both at entry and exit.
5496 ;;
5497 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5498 ;; 2010-01-29.
5502 new-beg new-end need-new-beg need-new-end)
5503 ;; Locate the barrier before the changed region
5508 ;; Remove the syntax-table properties from each pertinent <...> pair.
5509 ;; Firsly, the ones with the < before beg and > after beg.
5514 ;; Locate the barrier after END.
5519 ;; Remove syntax-table properties from the remaining pertinent <...>
5520 ;; pairs, those with a > after end and < before end.
5525 ;; Extend the fontification region, if needed.
5537 ;; This is called from `after-change-functions' when
5538 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5539 ;; chars with paren syntax become part of another operator like "<<"
5540 ;; or ">=".
5541 ;;
5542 ;; This function might do hidden buffer changes.
5575 \f
5576 ;; Handling of small scale constructs like types and names.
5578 ;; Dynamically bound variable that instructs `c-forward-type' to also
5579 ;; treat possible types (i.e. those that it normally returns 'maybe or
5580 ;; 'found for) as actual types (and always return 'found for them).
5581 ;; This means that it records them in `c-record-type-identifiers' if
5582 ;; that is set, and that it adds them to `c-found-types'.
5585 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5586 ;; mark up successfully parsed arglists with paren syntax properties on
5587 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5588 ;; `c-type' property of each argument separating comma.
5589 ;;
5590 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5591 ;; all arglists for side effects (i.e. recording types), otherwise it
5592 ;; exploits any existing paren syntax properties to quickly jump to the
5593 ;; end of already parsed arglists.
5594 ;;
5595 ;; Marking up the arglists is not the default since doing that correctly
5596 ;; depends on a proper value for `c-restricted-<>-arglists'.
5599 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5600 ;; not accept arglists that contain binary operators.
5601 ;;
5602 ;; This is primarily used to handle C++ template arglists. C++
5603 ;; disambiguates them by checking whether the preceding name is a
5604 ;; template or not. We can't do that, so we assume it is a template
5605 ;; if it can be parsed as one. That usually works well since
5606 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5607 ;; in almost all cases would be pointless.
5608 ;;
5609 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5610 ;; should let the comma separate the function arguments instead. And
5611 ;; in a context where the value of the expression is taken, e.g. in
5612 ;; "if (a < b || c > d)", it's probably not a template.
5615 ;; Dynamically bound variables that instructs
5616 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5617 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5618 ;; `c-forward-label' to record the ranges of all the type and
5619 ;; reference identifiers they encounter. They will build lists on
5620 ;; these variables where each element is a cons of the buffer
5621 ;; positions surrounding each identifier. This recording is only
5622 ;; activated when `c-record-type-identifiers' is non-nil.
5623 ;;
5624 ;; All known types that can't be identifiers are recorded, and also
5625 ;; other possible types if `c-promote-possible-types' is set.
5626 ;; Recording is however disabled inside angle bracket arglists that
5627 ;; are encountered inside names and other angle bracket arglists.
5628 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5629 ;; instead.
5630 ;;
5631 ;; Only the names in C++ template style references (e.g. "tmpl" in
5632 ;; "tmpl<a,b>::foo") are recorded as references, other references
5633 ;; aren't handled here.
5634 ;;
5635 ;; `c-forward-label' records the label identifier(s) on
5636 ;; `c-record-ref-identifiers'.
5640 ;; This variable will receive a cons cell of the range of the last
5641 ;; single identifier symbol stepped over by `c-forward-name' if it's
5642 ;; successful. This is the range that should be put on one of the
5643 ;; record lists above by the caller. It's assigned nil if there's no
5644 ;; such symbol in the name.
5649 ;; Always true.
5659 ;; Always true.
5667 ;; Dynamically bound variable that instructs `c-forward-type' to
5668 ;; record the ranges of types that only are found. Behaves otherwise
5669 ;; like `c-record-type-identifiers'.
5673 ;; Used internally in `c-forward-keyword-clause' to move forward
5674 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5675 ;; possibly is prefixed by keywords and their associated clauses.
5676 ;; Try with a type/name first to not trip up on those that begin
5677 ;; with a keyword. Return t if a known or found type is moved
5678 ;; over. The point is clobbered if nil is returned. If range
5679 ;; recording is enabled, the identifier is recorded on as a type
5680 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5681 ;;
5682 ;; This macro might do hidden buffer changes.
5687 nil
5694 t)))
5697 ;; Used internally in `c-forward-keyword-clause' to move forward
5698 ;; over a comma separated list of types or names using
5699 ;; `c-forward-keyword-prefixed-id'.
5700 ;;
5701 ;; This macro might do hidden buffer changes.
5712 ;; Submatch MATCH in the current match data is assumed to surround a
5713 ;; token. If it's a keyword, move over it and any immediately
5714 ;; following clauses associated with it, stopping at the start of
5715 ;; the next token. t is returned in that case, otherwise the point
5716 ;; stays and nil is returned. The kind of clauses that are
5717 ;; recognized are those specified by `c-type-list-kwds',
5718 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5719 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5720 ;; and `c-<>-arglist-kwds'.
5721 ;;
5722 ;; This function records identifier ranges on
5723 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5724 ;; `c-record-type-identifiers' is non-nil.
5725 ;;
5726 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5727 ;; apply directly after the keyword, the type list is moved over
5728 ;; only when there is no unaccounted token before it (i.e. a token
5729 ;; that isn't moved over due to some other keyword list). The
5730 ;; identifier ranges in the list are still recorded if that should
5731 ;; be done, though.
5732 ;;
5733 ;; This function might do hidden buffer changes.
5736 ;; The call to `c-forward-<>-arglist' below is made after
5737 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5738 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5739 ;; should therefore be nil.
5751 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5756 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5761 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5768 ;; Use `c-forward-type' on every identifier we can find
5769 ;; inside the paren, to record the types.
5794 ;; If we are at the colon already, we move over the type
5795 ;; list after it.
5801 ;; Not at the colon, so stop here. But the identifier
5802 ;; ranges in the type list later on should still be
5803 ;; recorded.
5806 ;; If a keyword matched both one of the types above and
5807 ;; this one, we match `c-colon-type-list-re' after the
5808 ;; clause matched above.
5815 ;; There's a type after the `c-colon-type-list-re' match
5816 ;; after a keyword in `c-colon-type-list-kwds'.
5820 t)))
5822 ;; cc-mode requires cc-fonts.
5826 ;; The point is assumed to be at a "<". Try to treat it as the open
5827 ;; paren of an angle bracket arglist and move forward to the
5828 ;; corresponding ">". If successful, the point is left after the
5829 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5830 ;; returned. If ALL-TYPES is t then all encountered arguments in
5831 ;; the arglist that might be types are treated as found types.
5832 ;;
5833 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5834 ;; function handles text properties on the angle brackets and argument
5835 ;; separating commas.
5836 ;;
5837 ;; `c-restricted-<>-arglists' controls how lenient the template
5838 ;; arglist recognition should be.
5839 ;;
5840 ;; This function records identifier ranges on
5841 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5842 ;; `c-record-type-identifiers' is non-nil.
5843 ;;
5844 ;; This function might do hidden buffer changes.
5847 ;; If `c-record-type-identifiers' is set then activate
5848 ;; recording of any found types that constitute an argument in
5849 ;; the arglist.
5857 ;; `nconc' doesn't mind that the tail of
5858 ;; `c-record-found-types' is t.
5861 t)
5864 nil)))
5867 ;; Recursive part of `c-forward-<>-arglist'.
5868 ;;
5869 ;; This function might do hidden buffer changes.
5872 ;; Cover this so that any recorded found type ranges are
5873 ;; automatically lost if it turns out to not be an angle
5874 ;; bracket arglist. It's propagated through the return value
5875 ;; on successful completion.
5877 ;; List that collects the positions after the argument
5878 ;; separating ',' in the arglist.
5879 arg-start-pos)
5880 ;; If the '<' has paren open syntax then we've marked it as an angle
5881 ;; bracket arglist before, so skip to the end.
5889 t
5890 ;; Got unmatched paren angle brackets. We don't clear the paren
5891 ;; syntax properties and retry, on the basis that it's very
5892 ;; unlikely that paren angle brackets become operators by code
5893 ;; manipulation. It's far more likely that it doesn't match due
5894 ;; to narrowing or some temporary change.
5896 nil))
5901 ;; go forward one non-alphanumeric character (group) per iteration of
5902 ;; this loop.
5909 ;; All encountered identifiers are types, so set the
5910 ;; promote flag and parse the type.
5933 ;; Note: These regexps exploit the match order in \| so
5934 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5936 ;; Stop on ',', '|', '&', '+' and '-' to catch
5937 ;; common binary operators that could be between
5938 ;; two comparison expressions "a<b" and "c>d".
5940 nil t t))
5944 ;; Either an operator starting with '>' or the end of
5945 ;; the angle bracket arglist.
5952 ;; The angle bracket arglist is finished.
5964 ;; Either an operator starting with '<' or a nested arglist.
5968 ;; The '<' begins a multi-char operator.
5972 ;; We're at a nested <.....>
5978 ;; There's always an identifier before an angle
5979 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
5980 ;; or `c-<>-arglist-kwds'.
5997 ;; It was an angle bracket arglist.
6000 ;; Record the identifier before the template as a type
6001 ;; or reference depending on whether the arglist is last
6002 ;; in a qualified identifier.
6012 ;; At a "less than" operator.
6015 )))
6022 ;; Just another argument. Record the position. The
6023 ;; type check stuff that made us stop at it is at
6024 ;; the top of the loop.
6028 ;; Got a character that can't be in an angle bracket
6029 ;; arglist argument. Abort using `throw', since
6030 ;; it's useless to try to find a surrounding arglist
6031 ;; if we're nested.
6037 ;; The point is assumed to be directly after a ">". Try to treat it
6038 ;; as the close paren of an angle bracket arglist and move back to
6039 ;; the corresponding "<". If successful, the point is left at
6040 ;; the "<" and t is returned, otherwise the point isn't moved and
6041 ;; nil is returned. ALL-TYPES is passed on to
6042 ;; `c-forward-<>-arglist'.
6043 ;;
6044 ;; If the optional LIMIT is given, it bounds the backward search.
6045 ;; It's then assumed to be at a syntactically relevant position.
6046 ;;
6047 ;; This is a wrapper around `c-forward-<>-arglist'. See that
6048 ;; function for more details.
6057 t
6058 ;; See corresponding note in `c-forward-<>-arglist'.
6060 nil)
6067 t
6068 ;; Stopped at bob or a char that isn't allowed in an
6069 ;; arglist, so we've failed.
6071 nil)
6076 ;; If we moved then the "<" was part of some
6077 ;; multicharacter token.
6078 t
6084 ;; Matched the arglist. Break the while.
6086 nil)
6088 ;; We started from a non-paren ">" inside an
6089 ;; arglist.
6091 nil)
6093 ;; Matched a shorter arglist. Can be a nested
6094 ;; one so continue looking.
6096 t))
6097 t))))))
6102 ;; Move forward over a complete name if at the beginning of one,
6103 ;; stopping at the next following token. A keyword, as such,
6104 ;; doesn't count as a name. If the point is not at something that
6105 ;; is recognized as a name then it stays put.
6106 ;;
6107 ;; A name could be something as simple as "foo" in C or something as
6108 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6109 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6110 ;; int>::*volatile const" in C++ (this function is actually little
6111 ;; more than a `looking-at' call in all modes except those that,
6112 ;; like C++, have `c-recognize-<>-arglists' set).
6113 ;;
6114 ;; Return
6115 ;; o - nil if no name is found;
6116 ;; o - 'template if it's an identifier ending with an angle bracket
6117 ;; arglist;
6118 ;; o - 'operator of it's an operator identifier;
6119 ;; o - t if it's some other kind of name.
6120 ;;
6121 ;; This function records identifier ranges on
6122 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6123 ;; `c-record-type-identifiers' is non-nil.
6124 ;;
6125 ;; This function might do hidden buffer changes.
6128 ;; Turn off `c-promote-possible-types' here since we might
6129 ;; call `c-forward-<>-arglist' and we don't want it to promote
6130 ;; every suspect thing in the arglist to a type. We're
6131 ;; typically called from `c-forward-type' in this case, and
6132 ;; the caller only wants the top level type that it finds to
6133 ;; be promoted.
6134 c-promote-possible-types)
6140 ;; Check for keyword. We go to the last symbol in
6141 ;; `c-identifier-key' first.
6154 ;; "template" is only valid inside an
6155 ;; identifier if preceded by "::".
6160 t))
6162 ;; Handle a C++ operator or template identifier.
6166 ;; Got "... ::template".
6170 res subres))))
6173 ;; Got a cast operator.
6177 ;; Now we should match a sequence of either
6178 ;; '*', '&' or a name followed by ":: *",
6179 ;; where each can be followed by a sequence
6180 ;; of `c-opt-type-modifier-key'.
6183 t)
6193 t))))
6201 ;; Got some other operator.
6209 nil)
6211 ;; `id-start' is equal to `id-end' if we've jumped over
6212 ;; an identifier that doesn't end with a symbol token.
6213 ;; That can occur e.g. for Java import directives on the
6214 ;; form "foo.bar.*".
6221 res t)))
6231 ;; Got a concatenated identifier. This handles the
6232 ;; cases with tricky syntactic whitespace that aren't
6233 ;; covered in `c-identifier-key'.
6236 t)
6240 ;; Maybe an angle bracket arglist.
6248 c-last-identifier-range nil)
6253 ;; Continue if there's an identifier concatenation
6254 ;; operator after the template argument.
6260 t)
6265 nil)))
6266 )))))
6269 res))
6272 ;; Move forward over a type spec if at the beginning of one,
6273 ;; stopping at the next following token. The keyword "typedef"
6274 ;; isn't part of a type spec here.
6275 ;;
6276 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6277 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6278 ;; The current (2009-03-10) intention is to convert all uses of
6279 ;; `c-forward-type' to call with this parameter set, then to
6280 ;; eliminate it.
6281 ;;
6282 ;; Return
6283 ;; o - t if it's a known type that can't be a name or other
6284 ;; expression;
6285 ;; o - 'known if it's an otherwise known type (according to
6286 ;; `*-font-lock-extra-types');
6287 ;; o - 'prefix if it's a known prefix of a type;
6288 ;; o - 'found if it's a type that matches one in `c-found-types';
6289 ;; o - 'maybe if it's an identifier that might be a type; or
6290 ;; o - nil if it can't be a type (the point isn't moved then).
6291 ;;
6292 ;; The point is assumed to be at the beginning of a token.
6293 ;;
6294 ;; Note that this function doesn't skip past the brace definition
6295 ;; that might be considered part of the type, e.g.
6296 ;; "enum {a, b, c} foo".
6297 ;;
6298 ;; This function records identifier ranges on
6299 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6300 ;; `c-record-type-identifiers' is non-nil.
6301 ;;
6302 ;; This function might do hidden buffer changes.
6310 ;; Skip leading type modifiers. If any are found we know it's a
6311 ;; prefix of a type.
6320 ; "typedef".
6328 ;; In many languages the name can be used without the
6329 ;; prefix, so we add it to `c-found-types'.
6332 c-last-identifier-range)
6354 id-range c-last-identifier-range))))
6366 id-end)
6368 ;; Looking at a primitive or known type identifier. We've
6369 ;; checked for a name first so that we don't go here if the
6370 ;; known type match only is a prefix of another name.
6381 ;; There might be more keywords for the type.
6409 ;; A normal identifier.
6419 ;; It's an identifier that has been used as
6420 ;; a type somewhere else.
6421 'found
6422 ;; It's an identifier that might be a type.
6425 ;; A template is a type.
6429 ;; Otherwise it's an operator identifier, which is not a type.
6434 ;; Skip trailing type modifiers. If any are found we know it's
6435 ;; a type.
6441 ;; Step over any type suffix operator. Do not let the existence
6442 ;; of these alter the classification of the found type, since
6443 ;; these operators typically are allowed in normal expressions
6444 ;; too.
6451 ;; Look for a trailing operator that concatenates the type
6452 ;; with a following one, and if so step past that one through
6453 ;; a recursive call. Note that we don't record concatenated
6454 ;; types in `c-found-types' - it's the component types that
6455 ;; are recorded when appropriate.
6458 c-promote-possible-types))
6459 ;; If we can't promote then set `c-record-found-types' so that
6460 ;; we can merge in the types from the second part afterwards if
6461 ;; it turns out to be a known type there.
6464 subres)
6473 ;; If either operand certainly is a type then both are, but we
6474 ;; don't let the existence of the operator itself promote two
6475 ;; uncertain types to a certain one.
6494 ;; Merge in the ranges of any types found by the second
6495 ;; `c-forward-type'.
6497 ;; `nconc' doesn't mind that the tail of
6498 ;; `c-record-found-types' is t.
6500 c-record-type-identifiers))))
6508 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6510 res))
6513 ;; Used for Java code only at the moment. Assumes point is on the
6514 ;; @, moves forward an annotation. returns nil if there is no
6515 ;; annotation at point.
6522 t)))
6525 ;; Pull the next open brace from PS (which has the form of paren-state),
6526 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
6534 ;; Test whether we are in a C++ member initializer list, and if so, go back
6535 ;; to the introducing ":", returning the position of the opening paren of
6536 ;; the function's arglist. Otherwise return nil, leaving point unchanged.
6539 res)
6577 res))
6579 \f
6580 ;; Handling of large scale constructs like statements and declarations.
6582 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6583 ;; defsubst or perhaps even a defun, but it contains lots of free
6584 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6586 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6587 ;; of types when parsing a declaration, which means that it
6588 ;; sometimes consumes the identifier in the declaration as a type.
6589 ;; This is used to "backtrack" and make the last type be treated as
6590 ;; an identifier instead.
6593 ;; These identifiers are bound only in the inner let.
6595 identifier-start type-start
6596 got-parens nil
6597 got-identifier t
6598 got-suffix t
6599 got-suffix-after-parens id-start
6603 t
6604 backup-at-type))
6606 id-start backup-id-start)
6608 id-start start-pos))
6610 ;; When these flags already are set we've found specifiers that
6611 ;; unconditionally signal these attributes - backtracking doesn't
6612 ;; change that. So keep them set in that case.
6619 ;; This identifier is bound only in the inner let.
6623 ;; Move forward over a declaration or a cast if at the start of one.
6624 ;; The point is assumed to be at the start of some token. Nil is
6625 ;; returned if no declaration or cast is recognized, and the point
6626 ;; is clobbered in that case.
6627 ;;
6628 ;; If a declaration is parsed:
6629 ;;
6630 ;; The point is left at the first token after the first complete
6631 ;; declarator, if there is one. The return value is a cons where
6632 ;; the car is the position of the first token in the declarator. (See
6633 ;; below for the cdr.)
6634 ;; Some examples:
6635 ;;
6636 ;; void foo (int a, char *b) stuff ...
6637 ;; car ^ ^ point
6638 ;; float (*a)[], b;
6639 ;; car ^ ^ point
6640 ;; unsigned int a = c_style_initializer, b;
6641 ;; car ^ ^ point
6642 ;; unsigned int a (cplusplus_style_initializer), b;
6643 ;; car ^ ^ point (might change)
6644 ;; class Foo : public Bar {}
6645 ;; car ^ ^ point
6646 ;; class PikeClass (int a, string b) stuff ...
6647 ;; car ^ ^ point
6648 ;; enum bool;
6649 ;; car ^ ^ point
6650 ;; enum bool flag;
6651 ;; car ^ ^ point
6652 ;; void cplusplus_function (int x) throw (Bad);
6653 ;; car ^ ^ point
6654 ;; Foo::Foo (int b) : Base (b) {}
6655 ;; car ^ ^ point
6656 ;;
6657 ;; The cdr of the return value is non-nil when a
6658 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6659 ;; Specifically it is a dotted pair (A . B) where B is t when a
6660 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6661 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6662 ;; specifier is present. I.e., (some of) the declared
6663 ;; identifier(s) are types.
6664 ;;
6665 ;; If a cast is parsed:
6666 ;;
6667 ;; The point is left at the first token after the closing paren of
6668 ;; the cast. The return value is `cast'. Note that the start
6669 ;; position must be at the first token inside the cast parenthesis
6670 ;; to recognize it.
6671 ;;
6672 ;; PRECEDING-TOKEN-END is the first position after the preceding
6673 ;; token, i.e. on the other side of the syntactic ws from the point.
6674 ;; Use a value less than or equal to (point-min) if the point is at
6675 ;; the first token in (the visible part of) the buffer.
6676 ;;
6677 ;; CONTEXT is a symbol that describes the context at the point:
6678 ;; 'decl In a comma-separated declaration context (typically
6679 ;; inside a function declaration arglist).
6680 ;; '<> In an angle bracket arglist.
6681 ;; 'arglist Some other type of arglist.
6682 ;; nil Some other context or unknown context. Includes
6683 ;; within the parens of an if, for, ... construct.
6684 ;;
6685 ;; LAST-CAST-END is the first token after the closing paren of a
6686 ;; preceding cast, or nil if none is known. If
6687 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6688 ;; the position after the closest preceding call where a cast was
6689 ;; matched. In that case it's used to discover chains of casts like
6690 ;; "(a) (b) c".
6691 ;;
6692 ;; This function records identifier ranges on
6693 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6694 ;; `c-record-type-identifiers' is non-nil.
6695 ;;
6696 ;; This function might do hidden buffer changes.
6699 ;; first type, i.e. after any leading specifiers. It might
6700 ;; also point at the beginning of the preceding syntactic
6701 ;; whitespace.
6703 ;; Set to the result of `c-forward-type'.
6704 at-type
6705 ;; The position of the first token in what we currently
6706 ;; believe is the type in the declaration or cast, after any
6707 ;; specifiers and their associated clauses.
6708 type-start
6709 ;; The position of the first token in what we currently
6710 ;; believe is the declarator for the first identifier. Set
6711 ;; when the type is found, and moved forward over any
6712 ;; `c-decl-hangon-kwds' and their associated clauses that
6713 ;; occurs after the type.
6714 id-start
6715 ;; These store `at-type', `type-start' and `id-start' of the
6716 ;; identifier before the one in those variables. The previous
6717 ;; identifier might turn out to be the real type in a
6718 ;; declaration if the last one has to be the declarator in it.
6719 ;; If `backup-at-type' is nil then the other variables have
6720 ;; undefined values.
6721 backup-at-type backup-type-start backup-id-start
6722 ;; Set if we've found a specifier (apart from "typedef") that makes
6723 ;; the defined identifier(s) types.
6724 at-type-decl
6725 ;; Set if we've a "typedef" keyword.
6726 at-typedef
6727 ;; Set if we've found a specifier that can start a declaration
6728 ;; where there's no type.
6729 maybe-typeless
6730 ;; If a specifier is found that also can be a type prefix,
6731 ;; these flags are set instead of those above. If we need to
6732 ;; back up an identifier, they are copied to the real flag
6733 ;; variables. Thus they only take effect if we fail to
6734 ;; interpret it as a type.
6735 backup-at-type-decl backup-maybe-typeless
6736 ;; Whether we've found a declaration or a cast. We might know
6737 ;; this before we've found the type in it. It's 'ids if we've
6738 ;; found two consecutive identifiers (usually a sure sign, but
6739 ;; we should allow that in labels too), and t if we've found a
6740 ;; specifier keyword (a 100% sure sign).
6741 at-decl-or-cast
6742 ;; Set when we need to back up to parse this as a declaration
6743 ;; but not as a cast.
6744 backup-if-not-cast
6745 ;; For casts, the return position.
6746 cast-end
6747 ;; Save `c-record-type-identifiers' and
6748 ;; `c-record-ref-identifiers' since ranges are recorded
6749 ;; speculatively and should be thrown away if it turns out
6750 ;; that it isn't a declaration or cast.
6757 ;; Check for a type. Unknown symbols are treated as possible
6758 ;; types, but they could also be specifiers disguised through
6759 ;; macros like __INLINE__, so we recognize both types and known
6760 ;; specifiers after them too.
6764 ;; Look for a specifier keyword clause.
6776 ;; Found a known or possible type or a prefix of a known type.
6779 ;; Got two identifiers with nothing but whitespace
6780 ;; between them. That can only happen in declarations.
6784 ;; If the previous identifier is a found type we
6785 ;; record it as a real one; it might be some sort of
6786 ;; alias for a prefix like "unsigned".
6793 backup-type-start type-start
6794 backup-id-start id-start
6795 at-type found-type
6796 type-start start
6798 ;; The previous ambiguous specifier/type turned out
6799 ;; to be a type since we've parsed another one after
6800 ;; it, so clear these backup flags.
6801 backup-at-type-decl nil
6802 backup-maybe-typeless nil))
6806 ;; Handle known specifier keywords and
6807 ;; `c-decl-hangon-kwds' which can occur after known
6808 ;; types.
6811 ;; It's a hang-on keyword that can occur anywhere.
6815 ;; Move the identifier start position if
6816 ;; we've passed a type.
6818 ;; Otherwise treat this as a specifier and
6819 ;; move the fallback position.
6823 ;; It's an ordinary specifier so we know that
6824 ;; anything before this can't be the type.
6826 start-pos kwd-clause-end)
6829 ;; It's ambiguous whether this keyword is a
6830 ;; specifier or a type prefix, so set the backup
6831 ;; flags. (It's assumed that `c-forward-type'
6832 ;; moved further than `c-forward-keyword-clause'.)
6840 ;; This test only happens after we've scanned a type.
6841 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6846 ;; Haven't matched a type so it's an unambiguous
6847 ;; specifier keyword and we know we're in a
6848 ;; declaration.
6853 ;; If the type isn't known we continue so that we'll jump
6854 ;; over all specifiers and type identifiers. The reason
6855 ;; to do this for a known type prefix is to make things
6856 ;; like "unsigned INT16" work.
6861 ;; If a known type was found, we still need to skip over any
6862 ;; hangon keyword clauses after it. Otherwise it has already
6863 ;; been done in the loop above.
6869 ;; A prefix type is itself a primitive type when it's not
6870 ;; followed by another type.
6874 ;; Got no type but set things up to continue anyway to handle
6875 ;; the various cases when a declaration doesn't start with a
6876 ;; type.
6881 ;; If it's C++ then check if the last "type" ends on the form
6882 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
6883 ;; (con|de)structor.
6894 ;; Cheating in the handling of syntactic ws below.
6901 name))
6902 ;; It is a (con|de)structor name. In that case the
6903 ;; declaration is typeless so zap out any preceding
6904 ;; identifier(s) that we might have taken as types.
6907 backup-at-type nil
6908 id-start type-start))))))
6910 ;; Check for and step over a type decl expression after the thing
6911 ;; that is or might be a type. This can't be skipped since we
6912 ;; need the correct end position of the declarator for
6913 ;; `max-type-decl-end-*'.
6915 ;; True if there's a non-open-paren match of
6916 ;; `c-type-decl-prefix-key'.
6917 got-prefix
6918 ;; True if the declarator is surrounded by a parenthesis pair.
6919 got-parens
6920 ;; True if there is an identifier in the declarator.
6921 got-identifier
6922 ;; True if there's a non-close-paren match of
6923 ;; `c-type-decl-suffix-key'.
6924 got-suffix
6925 ;; True if there's a prefix match outside the outermost
6926 ;; paren pair that surrounds the declarator.
6927 got-prefix-before-parens
6928 ;; True if there's a suffix match outside the outermost
6929 ;; paren pair that surrounds the declarator. The value is
6930 ;; the position of the first suffix match.
6931 got-suffix-after-parens
6932 ;; True if we've parsed the type decl to a token that is
6933 ;; known to end declarations in this context.
6934 at-decl-end
6935 ;; The earlier values of `at-type' and `type-start' if we've
6936 ;; shifted the type backwards.
6937 identifier-type identifier-start
6938 ;; If `c-parse-and-markup-<>-arglists' is set we need to
6939 ;; turn it off during the name skipping below to avoid
6940 ;; getting `c-type' properties that might be bogus. That
6941 ;; can happen since we don't know if
6942 ;; `c-restricted-<>-arglists' will be correct inside the
6943 ;; arglist paren that gets entered.
6944 c-parse-and-markup-<>-arglists
6945 ;; Start of the identifier for which `got-identifier' was set.
6946 name-start)
6950 ;; Skip over type decl prefix operators. (Note similar code in
6951 ;; `c-font-lock-declarators'.)
6961 ;; If the third submatch matches in C++ then
6962 ;; we're looking at an identifier that's a
6963 ;; prefix only if it specifies a member pointer.
6968 ;; We only check for a trailing "::" and
6969 ;; let the "*" that should follow be
6970 ;; matched in the next round.
6972 ;; It turned out to be the real identifier,
6973 ;; so stop.
6974 nil))
6975 t))
6989 ;; Skip over an identifier.
6996 ;; Skip over type decl suffix operators.
7003 t)
7007 t)
7013 ;; No suffix matched. We might have matched the
7014 ;; identifier as a type and the open paren of a
7015 ;; function arglist as a type decl prefix. In that
7016 ;; case we should "backtrack": Reinterpret the last
7017 ;; type as the identifier, move out of the arglist and
7018 ;; continue searching for suffix operators.
7019 ;;
7020 ;; Do this even if there's no preceding type, to cope
7021 ;; with old style function declarations in K&R C,
7022 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
7023 ;; style declarations. That isn't applicable in an
7024 ;; arglist context, though.
7029 maybe-typeless
7030 backup-maybe-typeless
7037 t))
7044 at-type)
7045 ;; Have found no identifier but `c-typeless-decl-kwds' has
7046 ;; matched so we know we're inside a declaration. The
7047 ;; preceding type must be the identifier instead.
7051 at-decl-or-cast
7054 ;; CASE 1
7056 ;; Encountered something inside parens that isn't matched by
7057 ;; the `c-type-decl-*' regexps, so it's not a type decl
7058 ;; expression. Try to skip out to the same paren depth to
7059 ;; not confuse the cast check below.
7061 ;; If we've found a specifier keyword then it's a
7062 ;; declaration regardless.
7070 ;; Now we've collected info about various characteristics of
7071 ;; the construct we're looking at. Below follows a decision
7072 ;; tree based on that. It's ordered to check more certain
7073 ;; signs before less certain ones.
7078 ;; CASE 2
7081 ;; Got another identifier directly after the type, so it's a
7082 ;; declaration.
7088 ;; (not got-suffix-after-parens)
7090 maybe-typeless
7091 backup-maybe-typeless
7096 ;; Got a declaration of the form "foo bar (gnu);" or "bar
7097 ;; (gnu);" where we've recognized "bar" as the type and "gnu"
7098 ;; as the declarator. In this case it's however more likely
7099 ;; that "bar" is the declarator and "gnu" a function argument
7100 ;; or initializer (if `c-recognize-paren-inits' is set),
7101 ;; since the parens around "gnu" would be superfluous if it's
7102 ;; a declarator. Shift the type one step backward.
7105 ;; Found no identifier.
7110 ;; CASE 3
7112 ;; Got a plain list of identifiers. If a colon follows it's
7113 ;; a valid label, or maybe a bitfield. Otherwise the last
7114 ;; one probably is the declared identifier and we should
7115 ;; back up to the previous type, providing it isn't a cast.
7119 ;; If we've found a specifier keyword then it's a
7120 ;; declaration regardless.
7131 ;; CASE 4
7135 ;; Got a plain list of identifiers followed by some suffix.
7136 ;; If this isn't a cast then the last identifier probably is
7137 ;; the declared one and we should back up to the previous
7138 ;; type.
7142 ;; CASE 5
7144 ;; If the type is known we know that there can't be any
7145 ;; identifier somewhere else, and it's only in declarations in
7146 ;; e.g. function prototypes and in casts that the identifier may
7147 ;; be left out.
7151 ;; Only got a single identifier (parsed as a type so far).
7152 ;; CASE 6
7154 ;; Check that the identifier isn't at the start of an
7155 ;; expression.
7156 at-decl-end
7159 ;; Inside an arglist that contains declarations. If K&R
7160 ;; style declarations and parenthesis style initializers
7161 ;; aren't allowed then the single identifier must be a
7162 ;; type, else we require that it's known or found
7163 ;; (primitive types are handled above).
7168 ;; Inside a template arglist. Accept known and found
7169 ;; types; other identifiers could just as well be
7170 ;; constants in C++.
7173 ;; CASE 7
7174 ;; Can't be a valid declaration or cast, but if we've found a
7175 ;; specifier it can't be anything else either, so treat it as
7176 ;; an invalid/unfinished declaration or cast.
7184 maybe-typeless
7185 backup-maybe-typeless
7189 c-after-suffixed-type-maybe-decl-key))))))
7190 ;; Got an empty paren pair and a preceding type that probably
7191 ;; really is the identifier. Shift the type backwards to make
7192 ;; the last one the identifier. This is analogous to the
7193 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7194 ;; above.
7195 ;;
7196 ;; Exception: In addition to the conditions in that
7197 ;; "backtracking" code, do not shift backward if we're not
7198 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7199 ;; Since there's no preceding type, the shift would mean that
7200 ;; the declaration is typeless. But if the regexp doesn't match
7201 ;; then we will simply fall through in the tests below and not
7202 ;; recognize it at all, so it's better to try it as an abstract
7203 ;; declarator instead.
7206 ;; Still no identifier.
7207 ;; CASE 8
7209 ;; Require `got-prefix' together with either `got-parens' or
7210 ;; `got-suffix' to recognize it as an abstract declarator:
7211 ;; `got-parens' only is probably an empty function call.
7212 ;; `got-suffix' only can build an ordinary expression together
7213 ;; with the preceding identifier which we've taken as a type.
7214 ;; We could actually accept on `got-prefix' only, but that can
7215 ;; easily occur temporarily while writing an expression so we
7216 ;; avoid that case anyway. We could do a better job if we knew
7217 ;; the point when the fontification was invoked.
7220 ;; CASE 9
7224 got-suffix-after-parens
7226 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7227 ;; normal function call after all (or perhaps a C++ style object
7228 ;; instantiation expression).
7231 ;; CASE 10
7233 ;; By now we've located the type in the declaration that we know
7234 ;; we're in.
7237 ;; CASE 11
7245 ;; Shift the type backward in the case that there's a
7246 ;; single identifier inside parens. That can only
7247 ;; occur in K&R style function declarations so it's
7248 ;; more likely that it really is a function call.
7249 ;; Therefore we only do this after
7250 ;; `c-after-suffixed-type-decl-key' has matched.
7252 got-suffix-after-parens))
7253 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7256 ;; CASE 12
7259 ;; It's a declaration if a known type precedes it and it can't be a
7260 ;; function call.
7263 ;; If we get here we can't tell if this is a type decl or a normal
7264 ;; expression by looking at it alone. (That's under the assumption
7265 ;; that normal expressions always can look like type decl expressions,
7266 ;; which isn't really true but the cases where it doesn't hold are so
7267 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7268 ;; the effort to look for them.)
7270 ;;; 2008-04-16: commented out the next form, to allow the function to recognize
7271 ;;; "foo (int bar)" in CC (an implicit type (in class foo) without a semicolon)
7272 ;;; as a(n almost complete) declaration, enabling it to be fontified.
7273 ;; CASE 13
7274 ;; (unless (or at-decl-end (looking-at "=[^=]"))
7275 ;; If this is a declaration it should end here or its initializer(*)
7276 ;; should start here, so check for allowed separation tokens. Note
7277 ;; that this rule doesn't work e.g. with a K&R arglist after a
7278 ;; function header.
7279 ;;
7280 ;; *) Don't check for C++ style initializers using parens
7281 ;; since those already have been matched as suffixes.
7282 ;;
7283 ;; If `at-decl-or-cast' is then we've found some other sign that
7284 ;; it's a declaration or cast, so then it's probably an
7285 ;; invalid/unfinished one.
7286 ;; (throw 'at-decl-or-cast at-decl-or-cast))
7288 ;; Below are tests that only should be applied when we're certain to
7289 ;; not have parsed halfway through an expression.
7291 ;; CASE 14
7293 ;; The expression starts with a known type so treat it as a
7294 ;; declaration.
7297 ;; CASE 15
7299 ;; In C++ we check if the identifier is a known type, since
7300 ;; (con|de)structors use the class name as identifier.
7301 ;; We've always shifted over the identifier as a type and
7302 ;; then backed up again in this case.
7303 identifier-type
7306 ;; `at-type' probably won't be 'found for
7307 ;; destructors since the "~" is then part of the
7308 ;; type name being checked against the list of
7309 ;; known types, so do a check without that
7310 ;; operator.
7315 c-identifier-syntax-table
7319 ;; We have already parsed the type earlier,
7320 ;; so it'd be possible to cache the end
7321 ;; position instead of redoing it here, but
7322 ;; then we'd need to keep track of another
7323 ;; position everywhere.
7331 ;; CASE 16
7333 at-type
7337 ;; Got something like "foo * bar;". Since we're not inside an
7338 ;; arglist it would be a meaningless expression because the
7339 ;; result isn't used. We therefore choose to recognize it as
7340 ;; a declaration. Do not allow a suffix since it could then
7341 ;; be a function call.
7344 ;; CASE 17
7349 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7350 ;; be an odd expression or it could be a declaration. Treat
7351 ;; it as a declaration if "a" has been used as a type
7352 ;; somewhere else (if it's a known type we won't get here).
7355 ;; CASE 18
7361 ;; Got a type followed by an abstract declarator. If `got-prefix'
7362 ;; is set it's something like "a *" without anything after it. If
7363 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7364 ;; or similar, which we accept only if the context rules out
7365 ;; expressions.
7368 ;; If we had a complete symbol table here (which rules out
7369 ;; `c-found-types') we should return t due to the disambiguation rule
7370 ;; (in at least C++) that anything that can be parsed as a declaration
7371 ;; is a declaration. Now we're being more defensive and prefer to
7372 ;; highlight things like "foo (bar);" as a declaration only if we're
7373 ;; inside an arglist that contains declarations.
7376 ;; The point is now after the type decl expression.
7379 ;; Check for a cast.
7382 c-cast-parens
7384 ;; Should be the first type/identifier in a cast paren.
7388 ;; The closing paren should follow.
7393 ;; There should be a primary expression after it.
7402 ;; Check if the expression begins with a prefix keyword.
7405 ;; Expression begins with an ambiguous operator. Treat
7406 ;; it as a cast if it's a type decl or if we've
7407 ;; recognized the type somewhere else.
7410 ;; Unless it's a keyword, it's the beginning of a primary
7411 ;; expression.
7413 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7414 ;; that it matched a whole one so that we don't e.g. confuse
7415 ;; the operator '-' with '->'. It's ok if it matches further,
7416 ;; though, since it e.g. can match the float '.5' while the
7417 ;; operator regexp only matches '.'.
7421 ;; There should either be a cast before it or something that isn't an
7422 ;; identifier or close paren.
7430 ;; It's a symbol. Accept it only if it's one of the
7431 ;; keywords that can precede an expression (without
7432 ;; surrounding parens).
7435 ;; Check that it isn't a close paren (block close is ok,
7436 ;; though).
7438 ;; Check that it isn't a nonsymbol identifier.
7441 ;; Handle the cast.
7451 ;; We're at a declaration. Highlight the type and the following
7452 ;; declarators.
7458 ;; Make sure to propagate the `c-decl-arg-start' property to
7459 ;; the next argument if it's set in this one, to cope with
7460 ;; interactive refontification.
7464 ;; There seems no reason to exclude a token from
7465 ;; fontification just because it's "a known type that can't
7466 ;; be a name or other expression". 2013-09-18.
7467 )
7478 ;; False alarm. Restore the recorded ranges.
7480 c-record-ref-identifiers save-rec-ref-ids)
7481 nil))))
7484 ;; Assuming that point is at the beginning of a token, check if it starts a
7485 ;; label and if so move over it and return non-nil (t in default situations,
7486 ;; specific symbols (see below) for interesting situations), otherwise don't
7487 ;; move and return nil. "Label" here means "most things with a colon".
7488 ;;
7489 ;; More precisely, a "label" is regarded as one of:
7490 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7491 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7492 ;; bare "case", should the colon be missing. We return t;
7493 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7494 ;; return t;
7495 ;; (iv) One of QT's "extended" C++ variants of
7496 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7497 ;; Returns the symbol `qt-2kwds-colon'.
7498 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7499 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7500 ;; colon). Currently (2006-03), this applies only to Objective C's
7501 ;; keywords "@private", "@protected", and "@public". Returns t.
7502 ;;
7503 ;; One of the things which will NOT be recognized as a label is a bit-field
7504 ;; element of a struct, something like "int foo:5".
7505 ;;
7506 ;; The end of the label is taken to be just after the colon, or the end of
7507 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7508 ;; after the end on return. The terminating char gets marked with
7509 ;; `c-decl-end' to improve recognition of the following declaration or
7510 ;; statement.
7511 ;;
7512 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7513 ;; label, if any, has already been marked up like that.
7514 ;;
7515 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7516 ;; after the preceding token, i.e. on the other side of the
7517 ;; syntactic ws from the point. Use a value less than or equal to
7518 ;; (point-min) if the point is at the first token in (the visible
7519 ;; part of) the buffer.
7520 ;;
7521 ;; The optional LIMIT limits the forward scan for the colon.
7522 ;;
7523 ;; This function records the ranges of the label symbols on
7524 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7525 ;; non-nil.
7526 ;;
7527 ;; This function might do hidden buffer changes.
7530 label-end
7531 qt-symbol-idx
7532 macro-start ; if we're in one.
7533 label-type
7534 kwd)
7536 ;; "case" or "default" (Doesn't apply to AWK).
7539 ;; Record only the keyword itself for fontification, since in
7540 ;; case labels the following is a constant expression and not
7541 ;; a label.
7545 ;; Find the label end.
7549 ;; Stop on chars that aren't allowed in expressions,
7550 ;; and on operator chars that would be meaningless
7551 ;; there. FIXME: This doesn't cope with ?: operators.
7559 t)
7561 ;; It's an unfinished label. We consider the keyword enough
7562 ;; to recognize it as a label, so that it gets fontified.
7563 ;; Leave the point at the end of it, but don't put any
7564 ;; `c-decl-end' marker.
7566 t))))
7568 ;; @private, @protected, @public, in Objective C, or similar.
7571 ;; For a `c-opt-extra-label-key' match, we record the whole
7572 ;; thing for fontification. That's to get the leading '@' in
7573 ;; Objective-C protection labels fontified.
7580 ;; All other cases of labels.
7583 ;; A colon label must have something before the colon.
7586 ;; Check that we're not after a token that can't precede a label.
7588 ;; Trivially succeeds when there's no preceding token.
7589 ;; Succeeds when we're at a virtual semicolon.
7598 ;; Check if we're after a label, if we're after a closing
7599 ;; paren that belong to statement, and with
7600 ;; `c-label-prefix-re'. It's done in different order
7601 ;; depending on `assume-markup' since the checks have
7602 ;; different expensiveness.
7629 ;; Might be after another label, so check it recursively.
7633 ;; Essentially the same as the
7634 ;; `c-syntactic-re-search-forward' regexp below.
7640 ;; Note: the following should work instead of the
7641 ;; narrow-to-region above. Investigate why not,
7642 ;; sometime. ACM, 2006-03-31.
7643 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7644 ;; macro-start t)
7646 ;; If the caller turned on recording for us,
7647 ;; it shouldn't apply when we check the
7648 ;; preceding label.
7649 c-record-type-identifiers)
7650 ;; A label can't start at a cpp directive. Check for
7651 ;; this, since c-forward-syntactic-ws would foul up on it.
7656 ;; Point is still at the beginning of the possible label construct.
7657 ;;
7658 ;; Check that the next nonsymbol token is ":", or that we're in one
7659 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7660 ;; arguments. FIXME: Should build this regexp from the language
7661 ;; constants.
7663 ;; public: protected: private:
7673 ;; QT double keyword like "protected slots:" or goto target.
7692 'qt-1kwd-colon
7704 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7714 t)
7717 ;; Record the identifiers in the label for fontification, unless
7718 ;; it begins with `c-label-kwds' in which case the following
7719 ;; identifiers are part of a (constant) expression that
7720 ;; shouldn't be fontified.
7732 ;; Not a label.
7734 label-type))
7737 ;; Assuming the point is at the beginning of a token, try to move
7738 ;; forward to the end of the Objective-C directive that starts
7739 ;; there. Return t if a directive was fully recognized, otherwise
7740 ;; the point is moved as far as one could be successfully parsed and
7741 ;; nil is returned.
7742 ;;
7743 ;; This function records identifier ranges on
7744 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7745 ;; `c-record-type-identifiers' is non-nil.
7746 ;;
7747 ;; This function might do hidden buffer changes.
7750 start-char
7752 lim
7753 ;; Turn off recognition of angle bracket arglists while parsing
7754 ;; types here since the protocol reference list might then be
7755 ;; considered part of the preceding name or superclass-name.
7766 t)
7775 ;; Handle the name of the class itself.
7777 ; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7778 ; at EOB.
7785 ;; Look for ": superclass-name" or "( category-name )".
7796 ;; Look for a protocol reference list.
7802 t))))
7808 t)
7811 nil)))
7814 ;; Go to the first non-whitespace after the colon that starts a
7815 ;; multiple inheritance introduction. Optional LIM is the farthest
7816 ;; back we should search.
7817 ;;
7818 ;; This function might do hidden buffer changes.
7826 ;; Return nil if we aren't in a method definition, otherwise the
7827 ;; position of the initial [+-].
7828 ;;
7829 ;; This function might do hidden buffer changes.
7835 ))
7837 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7839 ;; Return non-nil if point is within a gcc \"asm\" block.
7840 ;;
7841 ;; This should be called with point inside an argument list.
7842 ;;
7843 ;; Only one level of enclosing parentheses is considered, so for
7844 ;; instance `nil' is returned when in a function call within an asm
7845 ;; operand.
7846 ;;
7847 ;; This function might do hidden buffer changes.
7859 a function.
7861 More precisely, being at the top-level means that point is either
7862 outside any enclosing block (such as a function definition), or
7863 directly inside a class, namespace or other block that contains
7864 another declaration level.
7866 If point is not at the top-level (e.g. it is inside a method
7867 definition), then nil is returned. Otherwise, if point is at a
7868 top-level not enclosed within a class definition, t is returned.
7869 Otherwise, a 2-vector is returned where the zeroth element is the
7870 buffer position of the start of the class declaration, and the first
7871 element is the buffer position of the enclosing class's opening
7872 brace.
7874 Note that this function might do hidden buffer changes. See the
7875 comment at the start of cc-engine.el for more info."
7881 ;; Return non-nil if the point is in the region after the argument
7882 ;; list of a function and its opening brace (or semicolon in case it
7883 ;; got no body). If there are K&R style argument declarations in
7884 ;; that region, the point has to be inside the first one for this
7885 ;; function to recognize it.
7886 ;;
7887 ;; If successful, the point is moved to the first token after the
7888 ;; function header (see `c-forward-decl-or-cast-1' for details) and
7889 ;; the position of the opening paren of the function arglist is
7890 ;; returned.
7891 ;;
7892 ;; The point is clobbered if not successful.
7893 ;;
7894 ;; LIM is used as bound for backward buffer searches.
7895 ;;
7896 ;; This function might do hidden buffer changes.
7909 ;; There should not be a '=' or ',' between beg and the
7910 ;; start of the declaration since that means we were in the
7911 ;; "expression part" of the declaration.
7916 ;; Check that there's an arglist paren in the
7917 ;; declaration.
7920 ;; The declarator is a paren expression, so skip past it
7921 ;; so that we don't get stuck on that instead of the
7922 ;; function arglist.
7926 ;; Don't trip up on "operator ()".
7933 ;; Return the position of the first argument declaration if point is
7934 ;; inside a K&R style argument declaration list, nil otherwise.
7935 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
7936 ;; position that bounds the backward search for the argument list.
7937 ;;
7938 ;; Point must be within a possible K&R region, e.g. just before a top-level
7939 ;; "{". It must be outside of parens and brackets. The test can return
7940 ;; false positives otherwise.
7941 ;;
7942 ;; This function might do hidden buffer changes.
7946 ;; If we're in a macro, our search range is restricted to it. Narrow to
7947 ;; the searchable range.
7951 before-lparen after-rparen
7953 ; we give up
7956 ;; Search backwards for the defun's argument list. We give up if we
7957 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
7958 ;; a knr region) or BOB.
7959 ;;
7960 ;; The criterion for a paren structure being the arg list is:
7961 ;; o - there is non-WS stuff after it but before any "{"; AND
7962 ;; o - the token after it isn't a ";" AND
7963 ;; o - it is preceded by either an identifier (the function name) or
7964 ;; a macro expansion like "DEFUN (...)"; AND
7965 ;; o - its content is a non-empty comma-separated list of identifiers
7966 ;; (an empty arg list won't have a knr region).
7967 ;;
7968 ;; The following snippet illustrates these rules:
7969 ;; int foo (bar, baz, yuk)
7970 ;; int bar [] ;
7971 ;; int (*baz) (my_type) ;
7972 ;; int (*) (void) (*yuk) (void) ;
7973 ;; {
7984 ; parens/brackets.
7988 ;; We're inside a paren. Could it be our argument list....?
7994 ;; FIXME!!! What about macros between the parens? 2007/01/20
7997 ;; It can't be the arg list if next token is ; or {
8002 ;; Is the thing preceding the list an identifier (the
8003 ;; function name), or a macro expansion?
8013 ;; Have we got a non-empty list of comma-separated
8014 ;; identifiers?
8028 ;; ...Yes. We've identified the function's argument list.
8034 ;; ...No. The current parens aren't the function's arg list.
8041 ;; skip forward over conditional at point, including any predicate
8042 ;; statements in parentheses. No error checking is performed.
8043 ;;
8044 ;; This function might do hidden buffer changes.
8046 ;; else if()
8051 ;; do, else, try, finally
8056 ;; for, if, while, switch, catch, synchronized, foreach
8060 ;; If looking at the token after a conditional then return the
8061 ;; position of its start, otherwise return nil.
8062 ;;
8063 ;; This function might do hidden buffer changes.
8074 ;; If the point is after an operator identifier that isn't handled
8075 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
8076 ;; position of the start of that identifier is returned. nil is
8077 ;; returned otherwise. The point may be anywhere in the syntactic
8078 ;; whitespace after the last token of the operator identifier.
8079 ;;
8080 ;; This function might do hidden buffer changes.
8092 ;; Assuming point is at a brace that opens a statement block of some
8093 ;; kind, move to the proper anchor point for that block. It might
8094 ;; need to be adjusted further by c-add-stmt-syntax, but the
8095 ;; position at return is suitable as start position for that
8096 ;; function.
8097 ;;
8098 ;; This function might do hidden buffer changes.
8105 ;; Assuming point is at a brace that opens the block of a top level
8106 ;; declaration of some kind, move to the proper anchor point for
8107 ;; that block.
8108 ;;
8109 ;; This function might do hidden buffer changes.
8114 ;; Search forward for the end of the "header" of the current
8115 ;; declaration. That's the position where the definition body
8116 ;; starts, or the first variable initializer, or the ending
8117 ;; semicolon. I.e. search forward for the closest following
8118 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
8119 ;; _after_ the first found token, or at point-max if none is found.
8120 ;;
8121 ;; This function might do hidden buffer changes.
8126 ;; In C++ we need to take special care to handle operator
8127 ;; tokens and those pesky template brackets.
8132 ;; Handle operator identifiers, i.e. ignore any
8133 ;; operator token preceded by "operator".
8140 t
8142 nil)))))
8151 ;; Go to the beginning of the current declaration, or the beginning
8152 ;; of the previous one if already at the start of it. Point won't
8153 ;; be moved out of any surrounding paren. Return a cons cell of the
8154 ;; form (MOVE . KNR-POS). MOVE is like the return value from
8155 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
8156 ;; style argument declarations (and they are to be recognized) then
8157 ;; KNR-POS is set to the start of the first such argument
8158 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
8159 ;; position that bounds the backward search.
8160 ;;
8161 ;; NB: Cases where the declaration continues after the block, as in
8162 ;; "struct foo { ... } bar;", are currently recognized as two
8163 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
8164 ;;
8165 ;; This function might do hidden buffer changes.
8171 ;; `c-beginning-of-statement-1' stops at a block start, but we
8172 ;; want to continue if the block doesn't begin a top level
8173 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8174 ;; or an open paren.
8176 ;; Go back one "statement" each time round the loop until we're just
8177 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8178 ;; an ObjC method. This will move over a multiple declaration whose
8179 ;; components are comma separated.
8181 ;; Must check with c-opt-method-key in ObjC mode.
8191 ;; Check that we don't move from the first thing in a
8192 ;; macro to its header.
8198 move tentative-move))
8203 ;; Handle K&R argdecls. Back up after the "statement" jumped
8204 ;; over by `c-beginning-of-statement-1', unless it was the
8205 ;; function body, in which case we're sitting on the opening
8206 ;; brace now. Then test if we're in a K&R argdecl region and
8207 ;; that we started at the other side of the first argdecl in
8208 ;; it.
8218 'previous
8220 knr-argdecl-start))
8223 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8224 ;; statement, so the result will be 'previous if we've moved over any.
8225 ;; So change our result back to 'same if necessary.
8226 ;;
8227 ;; If they were brace list initializers we might not have moved over a
8228 ;; declaration boundary though, so change it to 'same if we've moved
8229 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8230 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8231 ;; potentially can search over a large amount of text.). Take special
8232 ;; pains not to get mislead by C++'s "operator=", and the like.
8235 c++-template-syntax-table
8241 ; no more, or get to one which isn't an "operator ="
8244 c-overloadable-operators-regexp
8245 c-opt-op-identifier-prefix
8260 ;; Assuming point is at the start of a declaration (as detected by
8261 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8262 ;; `c-beginning-of-decl-1', this function handles the case when a
8263 ;; block is followed by identifiers in e.g. struct declarations in C
8264 ;; or C++. If a proper end was found then t is returned, otherwise
8265 ;; point is moved as far as possible within the current sexp and nil
8266 ;; is returned. This function doesn't handle macros; use
8267 ;; `c-end-of-macro' instead in those cases.
8268 ;;
8269 ;; This function might do hidden buffer changes.
8272 c++-template-syntax-table
8280 ;; Stopped at the ';' in a K&R argdecl section which is
8281 ;; detected using the same criteria as in
8282 ;; `c-beginning-of-decl-1'. Move to the following block
8283 ;; start.
8287 ;; Encountered a block in the declaration. Jump over it.
8298 ;; Check for `c-opt-block-decls-with-vars-key'
8299 ;; before the first paren.
8302 c-opt-block-decls-with-vars-key
8304 lim t t t)
8307 ;; Check that the first following paren is
8308 ;; the block.
8310 lim t t t)
8312 ;; The declaration doesn't have any of the
8313 ;; `c-opt-block-decls-with-vars' keywords in the
8314 ;; beginning, so it ends here at the end of the block.
8322 nil)))
8325 ;; Assuming the point is at an open brace, check if it starts a
8326 ;; block that contains another declaration level, i.e. that isn't a
8327 ;; statement block or a brace list, and if so return non-nil.
8328 ;;
8329 ;; If the check is successful, the return value is the start of the
8330 ;; keyword that tells what kind of construct it is, i.e. typically
8331 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8332 ;; the point will be at the start of the construct, before any
8333 ;; leading specifiers, otherwise it's at the returned position.
8334 ;;
8335 ;; The point is clobbered if the check is unsuccessful.
8336 ;;
8337 ;; CONTAINING-SEXP is the position of the open of the surrounding
8338 ;; paren, or nil if none.
8339 ;;
8340 ;; The optional LIMIT limits the backward search for the start of
8341 ;; the construct. It's assumed to be at a syntactically relevant
8342 ;; position.
8343 ;;
8344 ;; If any template arglists are found in the searched region before
8345 ;; the open brace, they get marked with paren syntax.
8346 ;;
8347 ;; This function might do hidden buffer changes.
8354 ;; Could be at the end of a template arglist.
8365 ;; Note: Can't get bogus hits inside template arglists below since they
8366 ;; have gotten paren syntax above.
8368 ;; If `goto-start' is set we begin by searching for the
8369 ;; first possible position of a leading specifier list.
8370 ;; The `c-decl-block-key' search continues from there since
8371 ;; we know it can't match earlier.
8374 open-brace t t)
8376 t)
8377 t)
8384 ;; Found a keyword that can't be a type?
8387 ;; Can be a type too, in which case it's the return type of a
8388 ;; function (under the assumption that no declaration level
8389 ;; block construct starts with a type).
8392 ;; Jumped over a type, but it could be a declaration keyword
8393 ;; followed by the declared identifier that we've jumped over
8394 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8395 ;; then we should be at the declarator now, so check for a
8396 ;; valid declarator start.
8397 ;;
8398 ;; Note: This doesn't cope with the case when a declared
8399 ;; identifier is followed by e.g. '(' in a language where '('
8400 ;; also might be part of a declarator expression. Currently
8401 ;; there's no such language.
8405 ;; In Pike a list of modifiers may be followed by a brace
8406 ;; to make them apply to many identifiers. Note that the
8407 ;; match data will be empty on return in this case.
8413 ;; Use this variant to avoid yet another special regexp.
8417 t)))
8419 ;; Got a match.
8422 ;; Back up over any preceding specifiers and their clauses
8423 ;; by going forward from `first-specifier-pos', which is the
8424 ;; earliest possible position where the specifier list can
8425 ;; start.
8431 ;; Accept any plain symbol token on the ground that
8432 ;; it's a specifier masked through a macro (just
8433 ;; like `c-forward-decl-or-cast-1' skip forward over
8434 ;; such tokens).
8435 ;;
8436 ;; Could be more restrictive wrt invalid keywords,
8437 ;; but that'd only occur in invalid code so there's
8438 ;; no use spending effort on it.
8444 ;; Can't parse a declaration preamble and is still
8445 ;; before `kwd-start'. That means `first-specifier-pos'
8446 ;; was in some earlier construct. Search again.
8450 ;; Got no preamble before the block declaration keyword.
8456 kwd-start)))
8459 ;; Check if the closest containing paren sexp is a declaration
8460 ;; block, returning a 2 element vector in that case. Aref 0
8461 ;; contains the bufpos at boi of the class key line, and aref 1
8462 ;; contains the bufpos of the open brace. This function is an
8463 ;; obsolete wrapper for `c-looking-at-decl-block'.
8464 ;;
8465 ;; This function might do hidden buffer changes.
8473 nil))
8478 ;; Return the buffer position of the most enclosing decl-block brace (in the
8479 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8480 ;; none was found.
8489 open-brace))
8492 ;; Return the position of the L-brace if point is inside a brace list
8493 ;; initialization of an array, etc. This is an approximate function,
8494 ;; designed for speed over accuracy. It will not find every bracelist, but
8495 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8496 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8497 ;; is everywhere else.
8506 b-pos)))
8509 ;; We're at a "{". Move back to the enum-like keyword that starts this
8510 ;; declaration and return t, otherwise don't move and return nil.
8512 up-sexp-pos before-identifier)
8526 t)
8531 t)
8537 ;; return the buffer position of the beginning of the brace list
8538 ;; statement if we're inside a brace list, otherwise return nil.
8539 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8540 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8541 ;; braces
8542 ;;
8543 ;; N.B.: This algorithm can potentially get confused by cpp macros
8544 ;; placed in inconvenient locations. It's a trade-off we make for
8545 ;; speed.
8546 ;;
8547 ;; This function might do hidden buffer changes.
8549 ;; This will pick up brace list declarations.
8553 ;; this will pick up array/aggregate init lists, even if they are nested.
8556 ;; Pike can have class definitions anywhere, so we must
8557 ;; check for the class key here.
8559 c-decl-block-key))
8560 bufpos braceassignp lim next-containing macro-start)
8562 containing-sexp)
8573 ;; We're in an in-expression block of some kind. Do not
8574 ;; check nesting. We deliberately set the limit to the
8575 ;; containing sexp, so that c-looking-at-inexpr-block
8576 ;; doesn't check for an identifier before it.
8578 ;; see if the open brace is preceded by = or [...] in
8579 ;; this statement, but watch out for operator=
8582 ;; Checks to do only on the first sexp before the brace.
8585 ;; In Java, an initialization brace list may follow
8586 ;; directly after "new Foo[]", so check for a "new"
8587 ;; earlier.
8593 ;; Carry on looking if this is an
8594 ;; identifier (may contain "." in Java)
8595 ;; or another "[]" sexp.
8598 ;; Checks to do on all sexps before the brace, up to the
8599 ;; beginning of the statement.
8607 ;; We've seen a =, but must check earlier tokens so
8608 ;; that it isn't something that should be ignored.
8614 ;; Check for operator =
8617 nil)
8618 ;; Check for `<opchar>= in Pike.
8621 ;; Special case for Pikes
8622 ;; `[]=, since '[' is not in
8623 ;; the punctuation class.
8626 nil)
8628 ;; make sure we're not in a C++ template
8629 ;; argument assignment
8641 ))))
8642 nil)
8649 ;; We've hit the beginning of the aggregate list.
8654 ;; Brace lists can't contain a semicolon, so we're done.
8659 ;; We've a macro whose expansion starts with the '{'.
8660 ;; Heuristically, if we have a ';' in it we've not got a
8661 ;; brace list, otherwise we have.
8668 containing-sexp nil)
8671 ;; Go up one level
8673 lim nil
8674 next-containing nil)))))
8676 bufpos))
8677 ))
8680 ;; If we're looking at the start of a pike-style list, i.e., `({Â })',
8681 ;; `([Â ])', `(<Â >)', etc., a cons of a cons of its starting and ending
8682 ;; positions and its entry in c-special-brace-lists is returned, nil
8683 ;; otherwise. The ending position is nil if the list is still open.
8684 ;; LIM is the limit for forward search. The point may either be at
8685 ;; the `(' or at the following paren character. Tries to check the
8686 ;; matching closer, but assumes it's correct if no balanced paren is
8687 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8688 ;; a special brace list).
8689 ;;
8690 ;; This function might do hidden buffer changes.
8695 inner-beg end type)
8710 nil)))))
8720 ;; Check balancing of the inner paren
8721 ;; below.
8724 t)
8725 ;; If the inner char isn't a paren then
8726 ;; we can't check balancing, so just
8727 ;; check the char before the outer
8728 ;; closing paren.
8743 ;; Return non-nil if between two statements or declarations, assuming
8744 ;; point is not inside a literal or comment.
8745 ;;
8746 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8747 ;; are recommended instead.
8748 ;;
8749 ;; This function might do hidden buffer changes.
8754 ;; Return non-nil if we're looking at the beginning of a block
8755 ;; inside an expression. The value returned is actually a cons of
8756 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8757 ;; position of the beginning of the construct.
8758 ;;
8759 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8760 ;; position of the closest containing list. If it's nil, the
8761 ;; containing paren isn't used to decide whether we're inside an
8762 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8763 ;; needs to be farther back.
8764 ;;
8765 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8766 ;; brace block might be done. It should only be used when the
8767 ;; construct can be assumed to be complete, i.e. when the original
8768 ;; starting position was further down than that.
8769 ;;
8770 ;; This function might do hidden buffer changes.
8775 ;; Look at the character after point only as a last resort
8776 ;; when we can't disambiguate.
8796 ;; If the class definition is at the start of
8797 ;; a statement, we don't consider it an
8798 ;; in-expression class.
8806 ;; Also, in Pike we treat it as an
8807 ;; in-expression class if it's used in an
8808 ;; object clone expression.
8824 nil)
8832 ;; Accept several square bracket sexps for
8833 ;; Java array initializations.
8841 block-follows
8842 containing-sexp
8851 nil
8854 res))))
8857 ;; Returns non-nil if we're looking at the end of an in-expression
8858 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
8859 ;; PAREN-STATE is the paren state relevant at the current position.
8860 ;;
8861 ;; This function might do hidden buffer changes.
8863 ;; We currently only recognize a block.
8866 containing-sexp)
8875 paren-state)
8876 containing-sexp)))))
8879 ;; Is there a "virtual semicolon" at POS or point?
8880 ;; (See cc-defs.el for full details of "virtual semicolons".)
8881 ;;
8882 ;; This is true when point is at the last non syntactic WS position on the
8883 ;; line, there is a macro call last on the line, and this particular macro's
8884 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
8885 ;; semicolon.
8893 c-macro-with-semi-re
8896 ;; Check we've got nothing after this except comments and empty lines
8897 ;; joined by escaped EOLs.
8910 nil)
8918 ;; Check for one of the listed macros being before point.
8922 t))
8930 \f
8931 ;; `c-guess-basic-syntax' and the functions that precedes it below
8932 ;; implements the main decision tree for determining the syntactic
8933 ;; analysis of the current line of code.
8935 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
8936 ;; auto newline analysis.
8940 ;; BRACEPOS is the position of a brace in a construct like "namespace
8941 ;; Bar {". Return the anchor point in this construct; this is the
8942 ;; earliest symbol on the brace's line which isn't earlier than
8943 ;; "namespace".
8944 ;;
8945 ;; Currently (2007-08-17), "like namespace" means "matches
8946 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
8947 ;; or anything like that.
8957 ;; A simple function to prepend a new syntax element to
8958 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
8959 ;; should always be dynamically bound but since we read it first
8960 ;; we'll fail properly anyway if this function is misused.
8962 c-syntactic-context)))
8965 ;; Like `c-add-syntax' but appends to the end of the syntax list.
8966 ;; (Normally not necessary.)
8971 syntax-extra-args
8972 stop-at-boi-only
8973 containing-sexp
8974 paren-state)
8975 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
8976 ;; needed with further syntax elements of the types `substatement',
8977 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
8978 ;; `defun-block-intro'.
8979 ;;
8980 ;; Do the generic processing to anchor the given syntax symbol on
8981 ;; the preceding statement: Skip over any labels and containing
8982 ;; statements on the same line, and then search backward until we
8983 ;; find a statement or block start that begins at boi without a
8984 ;; label or comment.
8985 ;;
8986 ;; Point is assumed to be at the prospective anchor point for the
8987 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
8988 ;; skip past open parens and containing statements. Most of the added
8989 ;; syntax elements will get the same anchor point - the exception is
8990 ;; for an anchor in a construct like "namespace"[*] - this is as early
8991 ;; as possible in the construct but on the same line as the {.
8992 ;;
8993 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
8994 ;;
8995 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
8996 ;; syntax symbol. They are appended after the anchor point.
8997 ;;
8998 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
8999 ;; if the current statement starts there.
9000 ;;
9001 ;; Note: It's not a problem if PAREN-STATE "overshoots"
9002 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
9003 ;;
9004 ;; This function might do hidden buffer changes.
9007 ;; This is by far the most common case, so let's give it special
9008 ;; treatment.
9013 ;; Set when we're on a label, so that we don't stop there.
9014 ;; FIXME: To be complete we should check if we're on a label
9015 ;; now at the start.
9016 on-label)
9018 ;; Use point as the anchor point for "namespace", "extern", etc.
9022 syntax-extra-args)
9024 ;; Loop while we have to back out of containing blocks.
9029 ;; Loop while we have to back up statements.
9031 on-label
9034 ;; Skip past any comments that stands between the
9035 ;; statement start and boi.
9043 ;; Skip to the beginning of this statement or backward
9044 ;; another one.
9052 ;; If we didn't move we're at the start of a block and
9053 ;; have to continue outside it.
9062 ;; Special case to avoid deeper and deeper indentation
9063 ;; of "else if" clauses.
9064 )
9069 ;; If stop-at-boi-only is nil, we shouldn't back up
9070 ;; over previous or containing statements to try to
9071 ;; reach boi, so go back to the last position and
9072 ;; exit.
9079 ;; If we've moved into another statement then we
9080 ;; should no longer try to stop in the middle of a
9081 ;; line.
9084 ;; Record this as a substatement if we skipped up one
9085 ;; level.
9088 )))
9090 containing-sexp)
9092 ;; Now we have to go out of this block.
9095 ;; Don't stop in the middle of a special brace list opener
9096 ;; like "({".
9108 ;; Analyze the construct in front of the block we've stepped out
9109 ;; from and add the right syntactic element for it.
9112 step-type)
9115 ;; Stepped out of a parenthesis block, so we're in an
9116 ;; expression now.
9128 ;; Stepped out of an in-expression statement. This
9129 ;; syntactic element won't get an anchor pos.
9132 ;; A parenthesis normally belongs to an arglist.
9140 on-label nil))
9142 ;; Stepped out of a brace block.
9154 containing-sexp)))
9156 'defun-block-intro
9158 nil))
9162 c-other-decl-block-key-in-symbols-alist))
9169 ;; Always done if the open brace was at boi. The
9170 ;; c-beginning-of-statement-1 call above is necessary
9171 ;; anyway, to decide the type of block-intro to add.
9174 ))
9176 ;; Fill in the current point as the anchor for all the symbols
9177 ;; added above.
9186 )))
9189 containing-decl-open
9190 containing-decl-start
9191 containing-decl-kwd
9192 paren-state)
9193 ;; The inclass and class-close syntactic symbols are added in
9194 ;; several places and some work is needed to fix everything.
9195 ;; Therefore it's collected here.
9196 ;;
9197 ;; This function might do hidden buffer changes.
9202 containing-decl-open)
9204 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9205 ;; here, but we have to do like this for compatibility.
9215 char-after-ip
9216 beg-of-same-or-containing-stmt
9217 containing-sexp
9218 paren-state)
9219 ;; This function contains the decision tree reached through both
9220 ;; cases 18 and 10. It's a continued statement or top level
9221 ;; construct of some kind.
9222 ;;
9223 ;; This function might do hidden buffer changes.
9230 ;; (CASE A removed.)
9231 ;; CASE B: open braces for class or brace-lists
9238 ;; CASE B.1: class-open
9245 ;; CASE B.2: brace-list-open
9250 indent-point t t t)))
9251 ;; The most semantically accurate symbol here is
9252 ;; brace-list-open, but we normally report it simply as a
9253 ;; statement-cont. The reason is that one normally adjusts
9254 ;; brace-list-open for brace lists as top-level constructs,
9255 ;; and brace lists inside statements is a completely different
9256 ;; context. C.f. case 5A.3.
9259 ;; Turn off the dwim above when we're
9260 ;; analyzing the nature of the brace
9261 ;; for the auto newline feature.
9262 'brace-list-open
9264 nil nil
9265 containing-sexp paren-state))
9267 ;; CASE B.3: The body of a function declared inside a normal
9268 ;; block. Can occur e.g. in Pike and when using gcc
9269 ;; extensions, but watch out for macros followed by blocks.
9270 ;; C.f. cases E, 16F and 17G.
9276 ;; Turn off recognition of constructs that lacks a
9277 ;; type in this case, since that's more likely to be
9278 ;; a macro followed by a block.
9281 containing-sexp paren-state))
9283 ;; CASE B.4: Continued statement with block open. The most
9284 ;; accurate analysis is perhaps `statement-cont' together with
9285 ;; `block-open' but we play DWIM and use `substatement-open'
9286 ;; instead. The rationale is that this typically is a macro
9287 ;; followed by a block which makes it very similar to a
9288 ;; statement with a substatement block.
9291 containing-sexp paren-state))
9292 ))
9294 ;; CASE C: iostream insertion or extraction operator
9298 ;; If there is no preceding streamop in the statement
9299 ;; then indent this line as a normal statement-cont.
9303 t))))
9305 ;; CASE E: In the "K&R region" of a function declared inside a
9306 ;; normal block. C.f. case B.3.
9308 ;; Check that the next token is a '{'. This works as
9309 ;; long as no language that allows nested function
9310 ;; definitions allows stuff like member init lists, K&R
9311 ;; declarations or throws clauses there.
9312 ;;
9313 ;; Note that we do a forward search for something ahead
9314 ;; of the indentation line here. That's not good since
9315 ;; the user might not have typed it yet. Unfortunately
9316 ;; it's exceedingly tricky to recognize a function
9317 ;; prototype in a code block without resorting to this.
9325 ;; Turn off recognition of constructs that lacks a
9326 ;; type in this case, since that's more likely to be
9327 ;; a macro followed by a block.
9330 containing-sexp paren-state))
9332 ;;CASE F: continued statement and the only preceding items are
9333 ;;annotations.
9341 t)
9348 ;; CASE G: a template list continuation?
9349 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9367 ;; In a nested template arglist.
9373 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9374 ;; template aware.
9377 ;; CASE D: continued statement.
9381 containing-sexp paren-state))
9382 )))
9384 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9385 ;; 2005/11/29).
9386 ;;;###autoload
9388 "Return the syntactic context of the current line."
9394 open-paren-in-column-0-is-defun-start
9395 ;; A whole ugly bunch of various temporary variables. Have
9396 ;; to declare them here since it's not possible to declare
9397 ;; a variable with only the scope of a cond test and the
9398 ;; following result clauses, and most of this function is a
9399 ;; single gigantic cond. :P
9400 literal char-before-ip before-ws-ip char-after-ip macro-start
9401 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9402 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9403 containing-<
9404 ;; The following record some positions for the containing
9405 ;; declaration block if we're directly within one:
9406 ;; `containing-decl-open' is the position of the open
9407 ;; brace. `containing-decl-start' is the start of the
9408 ;; declaration. `containing-decl-kwd' is the keyword
9409 ;; symbol of the keyword that tells what kind of block it
9410 ;; is.
9411 containing-decl-open
9412 containing-decl-start
9413 containing-decl-kwd
9414 ;; The open paren of the closest surrounding sexp or nil if
9415 ;; there is none.
9416 containing-sexp
9417 ;; The position after the closest preceding brace sexp
9418 ;; (nested sexps are ignored), or the position after
9419 ;; `containing-sexp' if there is none, or (point-min) if
9420 ;; `containing-sexp' is nil.
9421 lim
9422 ;; The paren state outside `containing-sexp', or at
9423 ;; `indent-point' if `containing-sexp' is nil.
9425 ;; There's always at most one syntactic element which got
9426 ;; an anchor pos. It's stored in syntactic-relpos.
9427 syntactic-relpos
9430 ;; Check if we're directly inside an enclosing declaration
9431 ;; level block.
9440 containing-sexp)
9441 t)))
9444 containing-sexp nil)
9449 ;; Init some position variables.
9458 ;; Ignore balanced paren. The next entry
9459 ;; can't be another one.
9462 ;; If there is no surrounding open paren then
9463 ;; put the last balanced pair back on paren-state.
9465 containing-sexp nil)))
9469 ;; If we're in a parenthesis list then ',' delimits the
9470 ;; "statements" rather than being an operator (with the
9471 ;; exception of the "for" clause). This difference is
9472 ;; typically only noticeable when statements are used in macro
9473 ;; arglists.
9477 ;; cache char before and after indent point, and move point to
9478 ;; the most likely position to perform the majority of tests
9487 ;; are we in a literal?
9490 ;; now figure out syntactic qualities of the current line
9493 ;; CASE 1: in a string.
9497 ;; CASE 2: in a C or C++ style comment.
9499 ;; This is a kludge for XEmacs where we use
9500 ;; `buffer-syntactic-context', which doesn't correctly
9501 ;; recognize "\*/" to end a block comment.
9502 ;; `parse-partial-sexp' which is used by
9503 ;; `c-literal-limits' will however do that in most
9504 ;; versions, which results in that we get nil from
9505 ;; `c-literal-limits' even when `c-in-literal' claims
9506 ;; we're inside a comment.
9510 ;; CASE 3: in a cpp preprocessor macro continuation.
9520 ;; If at the beginning of the body of a #define
9521 ;; directive then analyze as cpp-define-intro
9522 ;; only. Go on with the syntactic analysis
9523 ;; otherwise. in-macro-expr is set if we're in a
9524 ;; cpp expression, i.e. before the #define body
9525 ;; or anywhere in a non-#define directive.
9532 nil)))))
9536 ;; CASE 11: an else clause?
9540 containing-sexp paren-state))
9542 ;; CASE 12: while closure of a do/while construct?
9550 containing-sexp paren-state))
9552 ;; CASE 13: A catch or finally clause? This case is simpler
9553 ;; than if-else and do-while, because a block is required
9554 ;; after every try, catch and finally.
9562 t)
9566 t)
9569 t))
9574 containing-sexp paren-state))
9576 ;; CASE 18: A substatement we can recognize by keyword.
9586 ;; Ought to cache the result from the
9587 ;; c-beginning-of-statement-1 calls here.
9597 ;; Step up to the containing statement if we
9598 ;; stayed in the same one.
9606 ;; There was no containing statement after all.
9608 placeholder))
9610 ;; Require a parenthesis after these keywords.
9611 ;; Necessary to catch e.g. synchronized in Java,
9612 ;; which can be used both as statement and
9613 ;; modifier.
9619 ;; CASE 18A: Simple substatement.
9625 containing-sexp paren-state))
9631 containing-sexp paren-state))
9634 containing-sexp paren-state))))
9636 ;; CASE 18B: Some other substatement. This is shared
9637 ;; with case 10.
9639 char-after-ip
9640 placeholder
9641 lim
9642 paren-state)))
9644 ;; CASE 14: A case or default label
9650 containing-sexp))
9653 ;; Got a bogus label at the top level. In lack of better
9654 ;; alternatives, anchor it on (point-min).
9657 ;; CASE 15: any other label
9664 containing-decl-open
9665 containing-decl-start
9666 containing-decl-kwd
9667 paren-state))
9668 ;; Append access-label with the same anchor point as
9669 ;; inclass gets.
9675 containing-sexp))
9680 ;; If the surrounding statement is a switch then
9681 ;; let's analyze all labels as switch labels, so
9682 ;; that they get lined up consistently.
9683 'case-label
9689 ;; A label on the top level. Treat it as a class
9690 ;; context. (point-min) is the closest we get to the
9691 ;; class open brace.
9694 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9695 ;; 17E.
9698 containing-sexp
9699 ;; Have to turn on the heuristics after
9700 ;; the point even though it doesn't work
9701 ;; very well. C.f. test case class-16.pike.
9702 t))
9707 ;; It's a statement block or an anonymous class.
9709 ;; It's a Pike lambda. Check whether we are between the
9710 ;; lambda keyword and the argument list or at the defun
9711 ;; opener.
9713 'inline-open
9719 paren-state)
9723 ;; CASE 5: Line is inside a declaration level block or at top level.
9727 ;; CASE 5A: we are looking at a defun, brace list, class,
9728 ;; or inline-inclass method opening brace
9735 ;; CASE 5A.1: Non-class declaration block open.
9750 ;; Special case for extern-lang-open.
9751 'extern-lang-open
9753 nil t containing-sexp paren-state))
9755 ;; CASE 5A.2: we are looking at a class opening brace
9764 ;; CASE 5A.3: brace list open
9789 ))))
9793 ;; We're in Java and have found that the open brace
9794 ;; belongs to a "new Foo[]" initialization list,
9795 ;; which means the brace list is part of an
9796 ;; expression and not a top level definition. We
9797 ;; therefore treat it as any topmost continuation
9798 ;; even though the semantically correct symbol still
9799 ;; is brace-list-open, on the same grounds as in
9800 ;; case B.2.
9806 ;; CASE 5A.4: inline defun open
9812 containing-decl-open
9813 containing-decl-start
9814 containing-decl-kwd
9815 paren-state))
9817 ;; CASE 5A.5: ordinary defun open
9825 ;; Bogus to use bol here, but it's the legacy. (Resolved,
9826 ;; 2007-11-09)
9827 ))))
9829 ;; CASE 5R: Member init list. (Used to be part of CASE 5B.1)
9830 ;; Note there is no limit on the backward search here, since member
9831 ;; init lists can, in practice, be very large.
9837 ;; There is no preceding member init clause.
9838 ;; Indent relative to the beginning of indentation
9839 ;; for the topmost-intro line that contains the
9840 ;; prototype's open paren.
9843 ;; Indent relative to the first member init clause.
9848 ;; CASE 5B: After a function header but before the body (or
9849 ;; the ending semicolon if there's no body).
9856 ;; CASE 5B.1: Member init list.
9858 ;; There is no preceding member init clause.
9859 ;; Indent relative to the beginning of indentation
9860 ;; for the topmost-intro line that contains the
9861 ;; prototype's open paren.
9865 ;; CASE 5B.2: K&R arg decl intro
9872 containing-decl-open
9873 containing-decl-start
9874 containing-decl-kwd
9875 paren-state)))
9877 ;; CASE 5B.4: Nether region after a C++ or Java func
9878 ;; decl, which could include a `throws' declaration.
9882 )))
9884 ;; CASE 5C: inheritance line. could be first inheritance
9885 ;; line, or continuation of a multiple inheritance
9893 ;; watch out for scope operator
9898 )))
9907 ;; for Java
9912 cont done)
9917 done t))
9921 )
9923 injava-inher)
9926 ))
9929 ;; CASE 5C.1: non-hanging colon on an inher intro
9933 ;; don't add inclass symbol since relative point already
9934 ;; contains any class offset
9935 )
9937 ;; CASE 5C.2: hanging colon on an inher intro
9943 containing-decl-open
9944 containing-decl-start
9945 containing-decl-kwd
9946 paren-state)))
9948 ;; CASE 5C.3: in a Java implements/extends
9962 )))
9964 ;; CASE 5C.4: a continued inheritance line
9968 ;; don't add inclass symbol since relative point already
9969 ;; contains any class offset
9970 )))
9972 ;; CASE 5P: AWK pattern or function or continuation
9973 ;; thereof.
9979 'topmost-intro-cont
9981 nil nil
9982 containing-sexp paren-state))
9984 ;; CASE 5D: this could be a top-level initialization, a
9985 ;; member init list continuation, or a template argument
9986 ;; list continuation.
9988 ;; Note: We use the fact that lim is always after any
9989 ;; preceding brace sexp.
9999 t))
10004 t)
10010 nil)
10011 t))
10013 ;; NB: No c-after-special-operator-id stuff in this
10014 ;; clause - we assume only C++ needs it.
10019 ;; CASE 5D.3: perhaps a template list continuation?
10034 ;; In a nested template arglist.
10040 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
10041 ;; template aware.
10044 ;; CASE 5D.4: perhaps a multiple inheritance line?
10056 t)
10061 ;; CASE 5D.5: Continuation of the "expression part" of a
10062 ;; top level construct. Or, perhaps, an unrecognized construct.
10074 ;; A preceding comma at the top level means that a
10075 ;; new variable declaration starts here. Use
10076 ;; topmost-intro-cont for it, for consistency with
10077 ;; the first variable declaration. C.f. case 5N.
10080 nil nil containing-sexp paren-state))
10081 ))
10083 ;; CASE 5F: Close of a non-class declaration level block.
10087 ;; This is inconsistent: Should use `containing-decl-open'
10088 ;; here if it's at boi, like in case 5J.
10092 ;; Special case for compatibility with the
10093 ;; extern-lang syntactic symbols.
10094 'extern-lang-close
10097 nil t
10099 paren-state))
10101 ;; CASE 5G: we are looking at the brace which closes the
10102 ;; enclosing nested class decl
10107 containing-decl-open
10108 containing-decl-start
10109 containing-decl-kwd
10110 paren-state))
10112 ;; CASE 5H: we could be looking at subsequent knr-argdecls
10119 ;; Do an extra check to avoid tripping up on
10120 ;; statements that occur in invalid contexts
10121 ;; (e.g. in macro bodies where we don't really
10122 ;; know the context of what we're looking at).
10129 ;; CASE 5I: ObjC method definition.
10134 ;; Handle the case when it's the first (non-comment)
10135 ;; thing in the buffer. Can't look for a 'same return
10136 ;; value from cbos1 since ObjC directives currently
10137 ;; aren't recognized fully, so that we get 'same
10138 ;; instead of 'previous if it moved over a preceding
10139 ;; directive.
10143 ;; CASE 5N: At a variable declaration that follows a class
10144 ;; definition or some other block declaration that doesn't
10145 ;; end at the closing '}'. C.f. case 5D.5.
10154 ;; Speed up the backward search a bit.
10159 ;; The '}' is unbalanced.
10160 nil
10165 containing-sexp paren-state))
10167 ;; NOTE: The point is at the end of the previous token here.
10169 ;; CASE 5J: we are at the topmost level, make
10170 ;; sure we skip back past any access specifiers
10172 ;; A macro continuation line is never at top level.
10194 nil)))))
10195 ;; For historic reasons we anchor at bol of the last
10196 ;; line of the previous declaration. That's clearly
10197 ;; highly bogus and useless, and it makes our lives hard
10198 ;; to remain compatible. :P
10209 ;; Special case for compatibility with the
10210 ;; extern-lang syntactic symbols.
10211 'inextern-lang
10214 nil t
10216 paren-state))
10218 containing-decl-open
10219 containing-decl-start
10220 containing-decl-kwd
10221 paren-state)))
10223 macro-start
10228 ;; CASE 5K: we are at an ObjC method definition
10229 ;; continuation line.
10238 ;; CASE 5L: we are at the first argument of a template
10239 ;; arglist that begins on the previous line.
10247 ;; CASE 5Q: we are at a statement within a macro.
10252 ;;CASE 5N: We are at a topmost continuation line and the only
10253 ;;preceding items are annotations.
10260 t)
10266 ;; CASE 5M: we are at a topmost continuation line
10277 ))
10280 ;; (CASE 6 has been removed.)
10282 ;; CASE 7: line is an expression, not a statement. Most
10283 ;; likely we are either in a function prototype or a function
10284 ;; call argument list
10292 ;; CASE 7A: we are looking at the arglist closing paren.
10293 ;; C.f. case 7F.
10305 paren-state))
10307 ;; CASE 7B: Looking at the opening brace of an
10308 ;; in-expression block or brace list. C.f. cases 4, 16A
10309 ;; and 17E.
10313 paren-state))
10317 placeholder
10320 paren-state)
10321 containing-sexp)))
10322 ;; placeholder is nil if it's a block directly in
10323 ;; a function arglist. That makes us skip out of
10324 ;; this case.
10325 )))
10330 paren-state)
10334 ;; CASE 7C: we are looking at the first argument in an empty
10335 ;; argument list. Use arglist-close if we're actually
10336 ;; looking at a close paren or bracket.
10348 paren-state))
10350 ;; CASE 7D: we are inside a conditional test clause. treat
10351 ;; these things as statements
10361 ))
10363 ;; CASE 7E: maybe a continued ObjC method call. This is the
10364 ;; case when we are inside a [] bracketed exp, and what
10365 ;; precede the opening bracket is not an identifier.
10373 )))
10375 ;; CASE 7F: we are looking at an arglist continuation line,
10376 ;; but the preceding argument is on the same line as the
10377 ;; opening paren. This case includes multi-line
10378 ;; mathematical paren groupings, but we could be on a
10379 ;; for-list continuation line. C.f. case 7A.
10396 paren-state))
10398 ;; CASE 7G: we are looking at just a normal arglist
10399 ;; continuation line
10402 ))
10404 ;; CASE 8: func-local multi-inheritance line
10414 ;; CASE 8A: non-hanging colon on an inher intro
10419 ;; CASE 8B: hanging colon on an inher intro
10423 ;; CASE 8C: a continued inheritance line
10427 )))
10429 ;; CASE 9: we are inside a brace-list
10439 ;; CASE 9A: In the middle of a special brace list opener.
10459 ;; CASE 9B: brace-list-close brace
10461 ;; Check special brace list closer.
10468 ;; We were between the special close char and the `)'.
10471 ;; We were before the special close char.
10475 ;; Normal brace list check.
10486 ;; Prepare for the rest of the cases below by going to the
10487 ;; token following the opening brace
10500 ;; CASE 9C: we're looking at the first line in a brace-list
10511 ;; CASE 9D: this is just a later brace-list-entry or
10512 ;; brace-entry-open
10521 ))
10522 ))))
10524 ;; CASE 10: A continued statement or top level construct.
10534 ;; This is shared with case 18.
10536 char-after-ip
10537 placeholder
10538 containing-sexp
10539 paren-state))
10541 ;; CASE 16: block close brace, possibly closing the defun or
10542 ;; the class
10544 ;; From here on we have the next containing sexp in lim.
10549 ;; CASE 16E: Closing a statement block? This catches
10550 ;; cases where it's preceded by a statement keyword,
10551 ;; which works even when used in an "invalid" context,
10552 ;; e.g. a macro argument.
10557 ;; CASE 16A: closing a lambda defun or an in-expression
10558 ;; block? C.f. cases 4, 7B and 17E.
10561 nil))
10563 'inline-close
10573 paren-state)
10577 ;; CASE 16B: does this close an inline or a function in
10578 ;; a non-class declaration level block?
10585 nil))
10595 ;; CASE 16F: Can be a defun-close of a function declared
10596 ;; in a statement block, e.g. in Pike or when using gcc
10597 ;; extensions, but watch out for macros followed by
10598 ;; blocks. Let it through to be handled below.
10599 ;; C.f. cases B.3 and 17G.
10605 ;; Turn off recognition of constructs that
10606 ;; lacks a type in this case, since that's more
10607 ;; likely to be a macro followed by a block.
10614 ;; CASE 16C: If there is an enclosing brace then this is
10615 ;; a block close since defun closes inside declaration
10616 ;; level blocks have been handled above.
10618 ;; If the block is preceded by a case/switch label on
10619 ;; the same line, we anchor at the first preceding label
10620 ;; at boi. The default handling in c-add-stmt-syntax
10621 ;; really fixes it better, but we do like this to keep
10622 ;; the indentation compatible with version 5.28 and
10623 ;; earlier. C.f. case 17H.
10630 ;; c-backward-to-block-anchor not necessary here; those
10631 ;; situations are handled in case 16E above.
10634 ;; CASE 16D: Only top level defun close left.
10640 paren-state))
10641 ))
10643 ;; CASE 19: line is an expression, not a statement, and is directly
10644 ;; contained by a template delimiter. Most likely, we are in a
10645 ;; template arglist within a statement. This case is based on CASE
10646 ;; 7. At some point in the future, we may wish to create more
10647 ;; syntactic symbols such as `template-intro',
10648 ;; `template-cont-nonempty', etc., and distinguish between them as we
10649 ;; do for `arglist-intro' etc. (2009-12-07).
10655 ; '<') before indent-point.
10663 paren-state))
10665 ;; CASE 17: Statement or defun catchall.
10668 ;; Back up statements until we find one that starts at boi.
10671 containing-sexp)))
10675 nil)
10680 ;; CASE 17B: continued statement
10684 containing-sexp paren-state))
10686 ;; CASE 17A: After a case/default label?
10694 'statement-case-open
10696 nil t containing-sexp paren-state))
10698 ;; CASE 17D: any old statement
10705 containing-sexp paren-state)
10709 ;; CASE 17I: Inside a substatement block.
10711 ;; The following tests are all based on containing-sexp.
10713 ;; From here on we have the next containing sexp in lim.
10718 lim paren-state)
10722 ;; CASE 17E: first statement in an in-expression block.
10723 ;; C.f. cases 4, 7B and 16A.
10726 nil))
10728 'defun-block-intro
10737 paren-state)
10743 ;; CASE 17F: first statement in an inline, or first
10744 ;; statement in a top-level defun. we can tell this is it
10745 ;; if there are no enclosing braces that haven't been
10746 ;; narrowed out by a class (i.e. don't use bod here).
10749 paren-state)))
10755 nil))))
10760 ;; CASE 17G: First statement in a function declared inside
10761 ;; a normal block. This can occur in Pike and with
10762 ;; e.g. the gcc extensions, but watch out for macros
10763 ;; followed by blocks. C.f. cases B.3 and 16F.
10769 ;; Turn off recognition of constructs that lacks
10770 ;; a type in this case, since that's more likely
10771 ;; to be a macro followed by a block.
10777 lim paren-state))
10779 ;; CASE 17H: First statement in a block.
10781 ;; If the block is preceded by a case/switch label on the
10782 ;; same line, we anchor at the first preceding label at
10783 ;; boi. The default handling in c-add-stmt-syntax is
10784 ;; really fixes it better, but we do like this to keep the
10785 ;; indentation compatible with version 5.28 and earlier.
10786 ;; C.f. case 16C.
10793 ;; c-backward-to-block-anchor not necessary here; those
10794 ;; situations are handled in case 17I above.
10796 lim paren-state))
10799 ))
10800 )
10802 ;; now we need to look at any modifiers
10806 ;; are we looking at a comment only line?
10811 ;; we might want to give additional offset to friends (in C++).
10816 ;; Set syntactic-relpos.
10822 nil)
10823 t))
10826 ;; Start of or a continuation of a preprocessor directive?
10841 ;; If inside a cpp expression, i.e. anywhere in a
10842 ;; cpp directive except a #define body, we only let
10843 ;; through the syntactic analysis that is internal
10844 ;; in the expression. That means the arglist
10845 ;; elements, if they are anchored inside the cpp
10846 ;; expression.
10855 ;; Inside a #define body and the syntactic analysis is
10856 ;; anchored on the start of the #define. In this case
10857 ;; we add cpp-define-intro to get the extra
10858 ;; indentation of the #define body.
10861 ;; return the syntax
10862 c-syntactic-context)))
10864 \f
10865 ;; Indentation calculation.
10868 ;; offset can be a number, a function, a variable, a list, or one of
10869 ;; the symbols + or -
10870 ;;
10871 ;; This function might do hidden buffer changes.
10890 langelem symbol))
10896 offset symbol)
10897 nil)
10911 Error evaluating offset %S for %s: \
10912 Cannot combine absolute offset %S with relative %S in `%s' method"
10918 Error evaluating offset %S for %s: \
10919 Cannot combine relative offset %S with absolute %S in `%s' method"
10924 res))
10942 Error evaluating offset %S for %s: \
10943 Cannot combine absolute offsets %S and %S in `add' method"
10947 res))
10956 res))))
10963 nil))))
10967 res
10969 offset symbol res)
10970 nil)))
10973 ;; Get offset from LANGELEM which is a list beginning with the
10974 ;; syntactic symbol and followed by any analysis data it provides.
10975 ;; That data may be zero or more elements, but if at least one is
10976 ;; given then the first is the anchor position (or nil). The symbol
10977 ;; is matched against `c-offsets-alist' and the offset calculated
10978 ;; from that is returned.
10979 ;;
10980 ;; This function might do hidden buffer changes.
10990 offset
10994 ))
10997 ;; This is a compatibility wrapper for `c-calc-offset' in case
10998 ;; someone is calling it directly. It takes an old style syntactic
10999 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
11000 ;; new list form.
11001 ;;
11002 ;; This function might do hidden buffer changes.
11009 ;; Calculate the syntactic indentation from a syntactic description
11010 ;; as returned by `c-guess-syntax'.
11011 ;;
11012 ;; Note that topmost-intro always has an anchor position at bol, for
11013 ;; historical reasons. It's often used together with other symbols
11014 ;; that has more sane positions. Since we always use the first
11015 ;; found anchor position, we rely on that these other symbols always
11016 ;; precede topmost-intro in the LANGELEMS list.
11017 ;;
11018 ;; This function might do hidden buffer changes.
11026 ;; Got an absolute column that overrides any indentation
11027 ;; we've collected so far, but not the relative
11028 ;; indentation we might get for the nested structures
11029 ;; further down the langelems list.
11033 ;; Got a relative change of the current calculated
11034 ;; indentation.
11037 ;; Use the anchor position from the first syntactic
11038 ;; element with one.
11048 indent)))
11050 \f
11053 ;;; cc-engine.el ends here