| blob | 7404241af9b91dba5585448b95aae49348eb72c2 |
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?
1039 ;; Just gone back over some paren block?
1043 before-sws-pos)))
1046 ;; Just gone back over an ordinary symbol of some sort?
1052 ptok tok
1053 tok boundary-pos
1055 ;; Like a C "continue". Analyze the next sexp.
1058 ;; ObjC method def?
1065 ;; Might we have a bitfield declaration, "<type> <id> : <size>"?
1068 ;; The : <size> and <id> fields?
1078 ;; The <type> field?
1086 bitfield-id-pos nil)))))
1088 ;; Handle labels.
1091 ;; `c-crosses-statement-barrier-p' has found a colon, so we
1092 ;; might be in a label now. Have we got a real label
1093 ;; (including a case label) or something like C++'s "public:"?
1094 ;; A case label might use an expression rather than a token.
1097 ;; Catch C++'s inheritance construct "class foo : bar".
1111 ; been found.
1113 ;; We're in a potential label and it's the first
1114 ;; time we've found something that isn't allowed in
1115 ;; one.
1118 ;; We've moved back by a sexp, so update the token positions.
1120 pptok ptok
1121 ptok tok
1127 ;; If the stack isn't empty there might be errors to report.
1136 ;; Need to investigate closer whether we've crossed
1137 ;; between a substatement and its containing statement.
1141 pptok)
1143 ptok)
1151 after-labels-pos
1158 last-label-pos
1159 pos))
1161 ;; We're in a label. Maybe we should step to the statement
1162 ;; after it.
1167 ;; Might have jumped over several labels. Go to the last one.
1170 ;; Have we got "case <expression>:"?
1180 ;; Skip over the unary operators that can start the statement.
1183 ;; protect AWK post-inc/decrement operators, etc.
1188 ret)))
1191 "Return non-nil if there is a non-comment non-macro punctuation character
1192 between FROM and TO. FROM must not be in a string or comment. The returned
1193 value is the position of the first such character."
1204 "Return non-nil if buffer positions FROM to TO cross one or more
1205 statement or declaration boundaries. The returned value is actually
1206 the position of the earliest boundary char. FROM must not be within
1207 a string or comment.
1209 The variable `c-maybe-labelp' is set to the position of the first `:' that
1210 might start a label (i.e. not part of `::' and not preceded by `?'). If a
1211 single `?' is found, then `c-maybe-labelp' is cleared.
1213 For AWK, a statement which is terminated by an EOL (not a \; or a }) is
1215 the line. If this virtual semicolon is _at_ from, the function recognizes it.
1217 Note that this function might do hidden buffer changes. See the
1218 comment at the start of cc-engine.el for more info."
1220 ;; If the current language has CPP macros, insert # into skip-chars.
1223 c-opt-cpp-symbol ; usually "#"
1225 c-stmt-delim-chars))
1228 lit-range vsemi-pos)
1235 skip-chars
1239 ;; Virtual semicolon?
1249 ;; In a string/comment?
1256 ;; Ignore scope operators.
1260 ;; A question mark. Can't be a label, so stop
1261 ;; looking for more : and ?.
1264 ;; At a CPP construct or a "#" or "##" operator?
1275 ;; In trailing space after an as yet undetected virtual semicolon?
1278 ; unterminated string/regexp.
1283 nil))))))
1286 "Return non-nil if the point is at the first token in a statement
1287 or somewhere in the syntactic whitespace before it.
1290 Any kind of declaration-like construct that occur outside function
1293 Note that this function might do hidden buffer changes. See the
1294 comment at the start of cc-engine.el for more info."
1298 c-maybe-labelp)
1309 "Return non-nil if the point is at the first token in an expression or
1310 statement, or somewhere in the syntactic whitespace before it.
1313 grammar sense: It's any sequence of expression tokens except commas,
1314 unless they are enclosed inside parentheses of some kind. Also, an
1315 expression never continues past an enclosing parenthesis, but it might
1316 contain parenthesis pairs of any sort except braces.
1318 Since expressions never cross statement boundaries, this function also
1319 recognizes statement beginnings, just like `c-at-statement-start-p'.
1321 Note that this function might do hidden buffer changes. See the
1322 comment at the start of cc-engine.el for more info."
1327 c-maybe-labelp)
1335 \f
1336 ;; A set of functions that covers various idiosyncrasies in
1337 ;; implementations of `forward-comment'.
1339 ;; Note: Some emacsen considers incorrectly that any line comment
1340 ;; ending with a backslash continues to the next line. I can't think
1341 ;; of any way to work around that in a reliable way without changing
1342 ;; the buffer, though. Suggestions welcome. ;) (No, temporarily
1343 ;; changing the syntax for backslash doesn't work since we must treat
1344 ;; escapes in string literals correctly.)
1347 "Move forward past whitespace and the closest following comment, if any.
1348 Return t if a comment was found, nil otherwise. In either case, the
1349 point is moved past the following whitespace. Line continuations,
1350 i.e. a backslashes followed by line breaks, are treated as whitespace.
1351 The line breaks that end line comments are considered to be the
1352 comment enders, so the point will be put on the beginning of the next
1353 line if it moved past a line comment.
1355 This function does not do any hidden buffer changes."
1363 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1364 ;; forwards at eob.
1365 nil
1367 ;; Emacs includes the ending newline in a b-style (c++)
1368 ;; comment, but XEmacs doesn't. We depend on the Emacs
1369 ;; behavior (which also is symmetric).
1373 t))))
1376 "Move forward past all following whitespace and comments.
1377 Line continuations, i.e. a backslashes followed by line breaks, are
1378 treated as whitespace.
1380 Note that this function might do hidden buffer changes. See the
1381 comment at the start of cc-engine.el for more info."
1384 ;; If forward-comment in at least XEmacs 21 is given a large
1385 ;; positive value, it'll loop all the way through if it hits
1386 ;; eob.
1388 ;; Some emacsen (e.g. XEmacs 21) return t when moving
1389 ;; forwards at eob.
1394 t))))
1397 "Move backward past whitespace and the closest preceding comment, if any.
1398 Return t if a comment was found, nil otherwise. In either case, the
1399 point is moved past the preceding whitespace. Line continuations,
1400 i.e. a backslashes followed by line breaks, are treated as whitespace.
1401 The line breaks that end line comments are considered to be the
1402 comment enders, so the point cannot be at the end of the same line to
1403 move over a line comment.
1405 This function does not do any hidden buffer changes."
1408 ;; When we got newline terminated comments, forward-comment in all
1409 ;; supported emacsen so far will stop at eol of each line not
1410 ;; ending with a comment when moving backwards. This corrects for
1411 ;; that, and at the same time handles line continuations.
1419 ;; Some emacsen (e.g. Emacs 19.34) return t when moving
1420 ;; backwards at bob.
1421 nil
1423 ;; Leave point after the closest following newline if we've
1424 ;; backed up over any above, since forward-comment won't move
1425 ;; backward over a line comment if point is at the end of the
1426 ;; same line.
1431 ;; If forward-comment above succeeded and we're at eol
1432 ;; then the newline we moved over above didn't end a
1433 ;; line comment, so we give it another go.
1436 t))
1438 ;; Emacs <= 20 and XEmacs move back over the closer of a
1439 ;; block comment that lacks an opener.
1442 t)))))
1445 "Move backward past all preceding whitespace and comments.
1446 Line continuations, i.e. a backslashes followed by line breaks, are
1447 treated as whitespace. The line breaks that end line comments are
1448 considered to be the comment enders, so the point cannot be at the end
1449 of the same line to move over a line comment. Unlike
1450 c-backward-syntactic-ws, this function doesn't move back over
1451 preprocessor directives.
1453 Note that this function might do hidden buffer changes. See the
1454 comment at the start of cc-engine.el for more info."
1458 ;; `forward-comment' in some emacsen (e.g. XEmacs 21.4)
1459 ;; return t when moving backwards at bob.
1465 ;; Cope specifically with ^M^J here -
1466 ;; forward-comment sometimes gets stuck after ^Ms,
1467 ;; sometimes after ^M^J.
1471 t)
1475 t))))
1476 moved-comment)
1478 ;; Emacs <= 20 and XEmacs move back over the
1479 ;; closer of a block comment that lacks an opener.
1481 t)
1483 ;; XEmacs treats line continuations as whitespace but
1484 ;; only in the backward direction, which seems a bit
1485 ;; odd. Anyway, this is necessary for Emacs.
1490 t))))))
1492 \f
1493 ;; Tools for skipping over syntactic whitespace.
1495 ;; The following functions use text properties to cache searches over
1496 ;; large regions of syntactic whitespace. It works as follows:
1497 ;;
1498 ;; o If a syntactic whitespace region contains anything but simple
1499 ;; whitespace (i.e. space, tab and line breaks), the text property
1500 ;; `c-in-sws' is put over it. At places where we have stopped
1501 ;; within that region there's also a `c-is-sws' text property.
1502 ;; That since there typically are nested whitespace inside that
1503 ;; must be handled separately, e.g. whitespace inside a comment or
1504 ;; cpp directive. Thus, from one point with `c-is-sws' it's safe
1505 ;; to jump to another point with that property within the same
1506 ;; `c-in-sws' region. It can be likened to a ladder where
1507 ;; `c-in-sws' marks the bars and `c-is-sws' the rungs.
1508 ;;
1509 ;; o The `c-is-sws' property is put on the simple whitespace chars at
1510 ;; a "rung position" and also maybe on the first following char.
1511 ;; As many characters as can be conveniently found in this range
1512 ;; are marked, but no assumption can be made that the whole range
1513 ;; is marked (it could be clobbered by later changes, for
1514 ;; instance).
1515 ;;
1516 ;; Note that some part of the beginning of a sequence of simple
1517 ;; whitespace might be part of the end of a preceding line comment
1518 ;; or cpp directive and must not be considered part of the "rung".
1519 ;; Such whitespace is some amount of horizontal whitespace followed
1520 ;; by a newline. In the case of cpp directives it could also be
1521 ;; two newlines with horizontal whitespace between them.
1522 ;;
1523 ;; The reason to include the first following char is to cope with
1524 ;; "rung positions" that doesn't have any ordinary whitespace. If
1525 ;; `c-is-sws' is put on a token character it does not have
1526 ;; `c-in-sws' set simultaneously. That's the only case when that
1527 ;; can occur, and the reason for not extending the `c-in-sws'
1528 ;; region to cover it is that the `c-in-sws' region could then be
1529 ;; accidentally merged with a following one if the token is only
1530 ;; one character long.
1531 ;;
1532 ;; o On buffer changes the `c-in-sws' and `c-is-sws' properties are
1533 ;; removed in the changed region. If the change was inside
1534 ;; syntactic whitespace that means that the "ladder" is broken, but
1535 ;; a later call to `c-forward-sws' or `c-backward-sws' will use the
1536 ;; parts on either side and use an ordinary search only to "repair"
1537 ;; the gap.
1538 ;;
1539 ;; Special care needs to be taken if a region is removed: If there
1540 ;; are `c-in-sws' on both sides of it which do not connect inside
1541 ;; the region then they can't be joined. If e.g. a marked macro is
1542 ;; broken, syntactic whitespace inside the new text might be
1543 ;; marked. If those marks would become connected with the old
1544 ;; `c-in-sws' range around the macro then we could get a ladder
1545 ;; with one end outside the macro and the other at some whitespace
1546 ;; within it.
1547 ;;
1548 ;; The main motivation for this system is to increase the speed in
1549 ;; skipping over the large whitespace regions that can occur at the
1550 ;; top level in e.g. header files that contain a lot of comments and
1551 ;; cpp directives. For small comments inside code it's probably
1552 ;; slower than using `forward-comment' straightforwardly, but speed is
1553 ;; not a significant factor there anyway.
1555 ; (defface c-debug-is-sws-face
1556 ; '((t (:background "GreenYellow")))
1557 ; "Debug face to mark the `c-is-sws' property.")
1558 ; (defface c-debug-in-sws-face
1559 ; '((t (:underline t)))
1560 ; "Debug face to mark the `c-in-sws' property.")
1562 ; (defun c-debug-put-sws-faces ()
1563 ; ;; Put the sws debug faces on all the `c-is-sws' and `c-in-sws'
1564 ; ;; properties in the buffer.
1565 ; (interactive)
1566 ; (save-excursion
1567 ; (c-save-buffer-state (in-face)
1568 ; (goto-char (point-min))
1569 ; (setq in-face (if (get-text-property (point) 'c-is-sws)
1570 ; (point)))
1571 ; (while (progn
1572 ; (goto-char (next-single-property-change
1573 ; (point) 'c-is-sws nil (point-max)))
1574 ; (if in-face
1575 ; (progn
1576 ; (c-debug-add-face in-face (point) 'c-debug-is-sws-face)
1577 ; (setq in-face nil))
1578 ; (setq in-face (point)))
1579 ; (not (eobp))))
1580 ; (goto-char (point-min))
1581 ; (setq in-face (if (get-text-property (point) 'c-in-sws)
1582 ; (point)))
1583 ; (while (progn
1584 ; (goto-char (next-single-property-change
1585 ; (point) 'c-in-sws nil (point-max)))
1586 ; (if in-face
1587 ; (progn
1588 ; (c-debug-add-face in-face (point) 'c-debug-in-sws-face)
1589 ; (setq in-face nil))
1590 ; (setq in-face (point)))
1591 ; (not (eobp)))))))
1594 ;;`(message ,@args)
1595 )
1598 ;; This macro does a hidden buffer change.
1605 ;; This macro does a hidden buffer change.
1612 ;; This macro does a hidden buffer change.
1619 ;; This macro does a hidden buffer change.
1626 ;; This macro does a hidden buffer change.
1634 ;; Called from `after-change-functions'. Note that if
1635 ;; `c-forward-sws' or `c-backward-sws' are used outside
1636 ;; `c-save-buffer-state' or similar then this will remove the cache
1637 ;; properties right after they're added.
1638 ;;
1639 ;; This function does hidden buffer changes.
1642 ;; Adjust the end to remove the properties in any following simple
1643 ;; ws up to and including the next line break, if there is any
1644 ;; after the changed region. This is necessary e.g. when a rung
1645 ;; marked empty line is converted to a line comment by inserting
1646 ;; "//" before the line break. In that case the line break would
1647 ;; keep the rung mark which could make a later `c-backward-sws'
1648 ;; move into the line comment instead of over it.
1658 ;; Ensure that an `c-in-sws' range gets broken. Note that it isn't
1659 ;; safe to keep a range that was continuous before the change. E.g:
1660 ;;
1661 ;; #define foo
1662 ;; \
1663 ;; bar
1664 ;;
1665 ;; There can be a "ladder" between "#" and "b". Now, if the newline
1666 ;; after "foo" is removed then "bar" will become part of the cpp
1667 ;; directive instead of a syntactically relevant token. In that
1668 ;; case there's no longer syntactic ws from "#" to "b".
1675 ;; Used by `c-forward-syntactic-ws' to implement the unbounded search.
1676 ;;
1677 ;; This function might do hidden buffer changes.
1680 ;; unmarked part of the simple ws region.
1682 rung-is-marked next-rung-is-marked simple-ws-end
1683 ;; `safe-start' is set when it's safe to cache the start position.
1684 ;; It's not set if we've initially skipped over comments and line
1685 ;; continuations since we might have gone out through the end of a
1686 ;; macro then. This provision makes `c-forward-sws' not populate the
1687 ;; cache in the majority of cases, but otoh is `c-backward-sws' by far
1688 ;; more common.
1689 safe-start)
1691 ;; Skip simple ws and do a quick check on the following character to see
1692 ;; if it's anything that can't start syntactic ws, so we can bail out
1693 ;; early in the majority of cases when there just are a few ws chars.
1700 ;; Find the last rung position to avoid setting properties in all
1701 ;; the cases when the marked rung is complete.
1702 ;; (`next-single-property-change' is certain to move at least one
1703 ;; step forward.)
1706 ;; Got no marked rung here. Since the simple ws might have started
1707 ;; inside a line comment or cpp directive we must set `rung-pos' as
1708 ;; high as possible.
1718 ;; The following search is the main reason that `c-in-sws'
1719 ;; and `c-is-sws' aren't combined to one property.
1723 ;; If the `c-in-sws' region extended past the last
1724 ;; `c-is-sws' char we have to go back a bit.
1731 "c-forward-sws cached move %s -> %s (max %s)"
1738 ;; We'll loop here if there is simple ws after the last rung.
1739 ;; That means that there's been some change in it and it's
1740 ;; possible that we've stepped into another ladder, so extend
1741 ;; the previous one to join with it if there is one, and try to
1742 ;; use the cache again.
1744 "c-forward-sws extending rung with [%s..%s] (max %s)"
1747 ;; Remove any `c-in-sws' property from the last char of
1748 ;; the rung before we mark it with `c-is-sws', so that we
1749 ;; won't connect with the remains of a broken "ladder".
1755 last-put-in-sws-pos rung-pos)))
1762 ;; Skipped over comments. Don't cache at eob in case the buffer
1763 ;; is narrowed.
1774 ;; Skip a preprocessor directive.
1781 ;; Don't cache at eob in case the buffer is narrowed.
1784 ;; We've searched over a piece of non-white syntactic ws. See if this
1785 ;; can be cached.
1791 ;; Cache if we haven't skipped comments only, and if we started
1792 ;; either from a marked rung or from a completely uncached
1793 ;; position.
1798 ;; See if there's a marked rung in the encountered simple ws. If
1799 ;; so then we can cache, unless `safe-start' is nil. Even then
1800 ;; we need to do this to check if the cache can be used for the
1801 ;; next step.
1805 safe-start))
1809 "c-forward-sws caching [%s..%s] - [%s..%s] (max %s)"
1813 ;; Remove the properties for any nested ws that might be cached.
1814 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
1815 ;; anyway.
1823 last-put-in-sws-pos rung-pos))
1825 ;; Remove any `c-in-sws' property from the last char of
1826 ;; the rung before we mark it with `c-is-sws', so that we
1827 ;; won't connect with the remains of a broken "ladder".
1830 rung-end-pos))
1833 "c-forward-sws not caching [%s..%s] - [%s..%s] (max %s)"
1837 ;; Set `rung-pos' for the next rung. It's the same thing here as
1838 ;; initially, except that the rung position is set as early as
1839 ;; possible since we can't be in the ending ws of a line comment or
1840 ;; cpp directive now.
1847 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
1848 ;; another one after the point (which might occur when editing inside a
1849 ;; comment or macro).
1853 "c-forward-sws clearing at %s for cache separation"
1854 last-put-in-sws-pos)
1858 ;; If at eob we have to clear the last character before the end
1859 ;; instead since the buffer might be narrowed and there might
1860 ;; be a `c-in-sws' after (point-max). In this case it's
1861 ;; necessary to clear both properties.
1863 "c-forward-sws clearing thoroughly at %s for cache separation"
1866 last-put-in-sws-pos))))
1867 ))))
1870 ;; Used by `c-backward-syntactic-ws' to implement the unbounded search.
1871 ;;
1872 ;; This function might do hidden buffer changes.
1875 ;; part of the simple ws region.
1877 rung-is-marked simple-ws-beg cmt-skip-pos)
1879 ;; Skip simple horizontal ws and do a quick check on the preceding
1880 ;; character to see if it's anything that can't end syntactic ws, so we can
1881 ;; bail out early in the majority of cases when there just are a few ws
1882 ;; chars. Newlines are complicated in the backward direction, so we can't
1883 ;; skip over them.
1890 ;; Try to find a rung position in the simple ws preceding point, so that
1891 ;; we can get a cache hit even if the last bit of the simple ws has
1892 ;; changed recently.
1898 ;; `rung-pos' will be the earliest marked position, which means that
1899 ;; there might be later unmarked parts in the simple ws region.
1900 ;; It's not worth the effort to fix that; the last part of the
1901 ;; simple ws is also typically edited often, so it could be wasted.
1913 ;; The following search is the main reason that `c-in-sws'
1914 ;; and `c-is-sws' aren't combined to one property.
1918 ;; If the `c-in-sws' region extended past the first
1919 ;; `c-is-sws' char we have to go forward a bit.
1924 "c-backward-sws cached move %s <- %s (min %s)"
1936 t
1938 nil))
1940 ;; We'll loop here if there is simple ws before the first rung.
1941 ;; That means that there's been some change in it and it's
1942 ;; possible that we've stepped into another ladder, so extend
1943 ;; the previous one to join with it if there is one, and try to
1944 ;; use the cache again.
1946 "c-backward-sws extending rung with [%s..%s] (min %s)"
1949 ;; Remove any `c-in-sws' property from the last char of
1950 ;; the rung before we mark it with `c-is-sws', so that we
1951 ;; won't connect with the remains of a broken "ladder".
1954 rung-pos)
1958 last-put-in-sws-pos rung-pos))
1967 ;; Inside a cpp directive. See if it should be skipped over.
1970 ;; Move back over all line continuations in the region skipped
1971 ;; over by `c-backward-comments'. If we go past it then we
1972 ;; started inside the cpp directive.
1981 ;; Don't move past the cpp directive if we began inside
1982 ;; it. Note that the position at the end of the last line
1983 ;; of the macro is also considered to be within it.
1985 nil)
1987 ;; It's worthwhile to spend a little bit of effort on finding
1988 ;; the end of the macro, to get a good `simple-ws-beg'
1989 ;; position for the cache. Note that `c-backward-comments'
1990 ;; could have stepped over some comments before going into
1991 ;; the macro, and then `simple-ws-beg' must be kept on the
1992 ;; same side of those comments.
1999 ;; Might happen if comments after the macro were skipped
2000 ;; over.
2004 t)))
2009 simple-ws-beg)
2010 ;; Skipped over comments. Must put point at the end of
2011 ;; the simple ws at point since we might be after a line
2012 ;; comment or cpp directive that's been partially
2013 ;; narrowed out, and we can't risk marking the simple ws
2014 ;; at the end of it.
2016 t)))
2018 ;; We've searched over a piece of non-white syntactic ws. See if this
2019 ;; can be cached.
2024 ;; Cache if we started either from a marked rung or from a
2025 ;; completely uncached position.
2026 rung-is-marked
2029 ;; Cache if there's a marked rung in the encountered simple ws.
2037 "c-backward-sws caching [%s..%s] - [%s..%s] (min %s)"
2042 ;; Remove the properties for any nested ws that might be cached.
2043 ;; Only necessary for `c-is-sws' since `c-in-sws' will be set
2044 ;; anyway.
2049 ;; Remove any `c-in-sws' property from the last char of
2050 ;; the rung before we mark it with `c-is-sws', so that we
2051 ;; won't connect with the remains of a broken "ladder".
2054 rung-end-pos)
2057 last-put-in-sws-pos simple-ws-beg)
2058 rung-pos)
2063 "c-backward-sws not caching [%s..%s] - [%s..%s] (min %s)"
2069 ))
2071 ;; Make sure that the newly marked `c-in-sws' region doesn't connect to
2072 ;; another one before the point (which might occur when editing inside a
2073 ;; comment or macro).
2077 "c-backward-sws clearing at %s for cache separation"
2080 last-put-in-sws-pos))
2082 ;; If at bob and the buffer is narrowed, we have to clear the
2083 ;; character we're standing on instead since there might be a
2084 ;; `c-in-sws' before (point-min). In this case it's necessary
2085 ;; to clear both properties.
2087 "c-backward-sws clearing thoroughly at %s for cache separation"
2088 last-put-in-sws-pos)
2091 ))))
2093 \f
2094 ;; Other whitespace tools
2096 ;; Is the region (beg end) WS, and is there WS (or BOB/EOB) next to the
2097 ;; region? This is a "heuristic" function. .....
2098 ;;
2099 ;; The motivation for the second bit is to check whether removing this
2100 ;; region would coalesce two symbols.
2101 ;;
2102 ;; FIXME!!! This function doesn't check virtual semicolons in any way. Be
2103 ;; careful about using this function for, e.g. AWK. (2007/3/7)
2112 \f
2113 ;; A system for finding noteworthy parens before the point.
2116 ;; A maximum comfortable scanning distance, e.g. between
2117 ;; `c-state-cache-good-pos' and "HERE" (where we call c-parse-state). When
2118 ;; this distance is exceeded, we take "emergency measures", e.g. by clearing
2119 ;; the cache and starting again from point-min or a beginning of defun. This
2120 ;; value can be tuned for efficiency or set to a lower value for testing.
2124 ;; The state cache used by `c-parse-state' to cut down the amount of
2125 ;; searching. It's the result from some earlier `c-parse-state' call. See
2126 ;; `c-parse-state''s doc string for details of its structure.
2127 ;;
2128 ;; The use of the cached info is more effective if the next
2129 ;; `c-parse-state' call is on a line close by the one the cached state
2130 ;; was made at; the cache can actually slow down a little if the
2131 ;; cached state was made very far back in the buffer. The cache is
2132 ;; most effective if `c-parse-state' is used on each line while moving
2133 ;; forward.
2137 ;; This is a position where `c-state-cache' is known to be correct, or
2138 ;; nil (see below). It's a position inside one of the recorded unclosed
2139 ;; parens or the top level, but not further nested inside any literal or
2140 ;; subparen that is closed before the last recorded position.
2141 ;;
2142 ;; The exact position is chosen to try to be close to yet earlier than
2143 ;; the position where `c-state-cache' will be called next. Right now
2144 ;; the heuristic is to set it to the position after the last found
2145 ;; closing paren (of any type) before the line on which
2146 ;; `c-parse-state' was called. That is chosen primarily to work well
2147 ;; with refontification of the current line.
2148 ;;
2149 ;; 2009-07-28: When `c-state-point-min' and the last position where
2150 ;; `c-parse-state' or for which `c-invalidate-state-cache' was called, are
2151 ;; both in the same literal, there is no such "good position", and
2152 ;; c-state-cache-good-pos is then nil. This is the ONLY circumstance in which
2153 ;; it can be nil. In this case, `c-state-point-min-literal' will be non-nil.
2154 ;;
2155 ;; 2009-06-12: In a brace desert, c-state-cache-good-pos may also be in
2156 ;; the middle of the desert, as long as it is not within a brace pair
2157 ;; recorded in `c-state-cache' or a paren/bracket pair.
2160 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2161 ;; We maintain a simple cache of positions which aren't in a literal, so as to
2162 ;; speed up testing for non-literality.
2164 ;; The approximate interval between entries in `c-state-nonlit-pos-cache'.
2168 ;; A list of buffer positions which are known not to be in a literal or a cpp
2169 ;; construct. This is ordered with higher positions at the front of the list.
2170 ;; Only those which are less than `c-state-nonlit-pos-cache-limit' are valid.
2174 ;; An upper limit on valid entries in `c-state-nonlit-pos-cache'. This is
2175 ;; reduced by buffer changes, and increased by invocations of
2176 ;; `c-state-literal-at'.
2180 ;; A list of buffer positions which are known not to be in a literal. This is
2181 ;; ordered with higher positions at the front of the list. Only those which
2182 ;; are less than `c-state-semi-nonlit-pos-cache-limit' are valid.
2186 ;; An upper limit on valid entries in `c-state-semi-nonlit-pos-cache'. This is
2187 ;; reduced by buffer changes, and increased by invocations of
2188 ;; `c-state-literal-at'. FIXME!!!
2191 ;; Do a parse-partial-sexp from FROM to TO, returning either
2192 ;; (STATE TYPE (BEG . END)) if TO is in a literal; or
2193 ;; (STATE) otherwise,
2194 ;; where STATE is the parsing state at TO, TYPE is the type of the literal
2195 ;; (one of 'c, 'c++, 'string) and (BEG . END) is the boundaries of the literal.
2196 ;;
2197 ;; Unless NOT-IN-DELIMITER is non-nil, when TO is inside a two-character
2198 ;; comment opener, this is recognized as being in a comment literal.
2199 ;;
2200 ;; Only elements 3 (in a string), 4 (in a comment), 5 (following a quote),
2201 ;; 7 (comment type) and 8 (start of comment/string) (and possibly 9) of
2202 ;; STATE are valid.
2205 ty co-st)
2213 nil ; TARGETDEPTH
2214 nil ; STOPBEFORE
2215 s ; OLDSTATE
2233 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2234 ;; string, comment, or macro.
2235 ;;
2236 ;; NOTE: This function manipulates `c-state-nonlit-pos-cache'. This cache
2237 ;; MAY NOT contain any positions within macros, since macros are frequently
2238 ;; turned into comments by use of the `c-cpp-delimiter' category properties.
2239 ;; We cannot rely on this mechanism whilst determining a cache pos since
2240 ;; this function is also called from outwith `c-parse-state'.
2245 pos npos high-pos lit macro-beg macro-end)
2246 ;; Trim the cache to take account of buffer changes.
2258 ;; Add an element to `c-state-nonlit-pos-cache' each iteration.
2262 ;; Test for being in a literal. If so, go to after it.
2269 ;; Test for being in a macro. If so, go to after it.
2284 ;; Add one extra element above HERE so as to to avoid the previous
2285 ;; expensive calculation when the next call is close to the current
2286 ;; one. This is especially useful when inside a large macro.
2291 pos))))
2294 ;; Return a buffer position before HERE which is "safe", i.e. outside any
2295 ;; string or comment. It may be in a macro.
2300 pos npos high-pos lit macro-beg macro-end)
2301 ;; Trim the cache to take account of buffer changes.
2313 ;; Add an element to `c-state-semi-nonlit-pos-cache' each iteration.
2317 ;; Test for being in a literal. If so, go to after it.
2330 pos))))
2333 ;; If position HERE is inside a literal, return (START . END), the
2334 ;; boundaries of the literal (which may be outside the accessible bit of the
2335 ;; buffer). Otherwise, return nil.
2336 ;;
2337 ;; This function is almost the same as `c-literal-limits'. Previously, it
2338 ;; differed in that it was a lower level function, and that it rigorously
2339 ;; followed the syntax from BOB. `c-literal-limits' is now (2011-12)
2340 ;; virtually identical to this function.
2348 ;; Return the start of the literal containing POS, or POS itself.
2350 pos))
2353 ;; Return a position outside of a string/comment/macro at or before POS.
2354 ;; STATE is the parse-partial-sexp state at POS.
2358 pos)))
2363 res))))
2365 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2366 ;; Stuff to do with point-min, and coping with any literal there.
2369 ;; This is (point-min) when `c-state-cache' was last calculated. A change of
2370 ;; narrowing is likely to affect the parens that are visible before the point.
2376 ;; These two variables define the literal, if any, containing point-min.
2377 ;; Their values are, respectively, 'string, c, or c++, and the start of the
2378 ;; literal. If there's no literal there, they're both nil.
2382 ;; This is the earliest buffer-pos from which scanning can be done. It is
2383 ;; either the end of the literal containing point-min, or point-min itself.
2384 ;; It becomes nil if the buffer is changed earlier than this point.
2386 ;; Return the lowest valid scanning pos. This will be the end of the
2387 ;; literal enclosing point-min, or point-min itself.
2399 ;; Determine the properties of any literal containing POINT-MIN, setting the
2400 ;; variables `c-state-point-min-lit-type', `c-state-point-min-lit-start',
2401 ;; and `c-state-min-scan-pos' accordingly. The return value is meaningless.
2403 lit)
2418 c-state-point-min-lit-start nil
2419 c-state-min-scan-pos p-min)))))
2422 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2423 ;; A variable which signals a brace dessert - helpful for reducing the number
2424 ;; of fruitless backward scans.
2427 ;; Used only in `c-append-lower-brace-pair-to-state-cache'. It is set when
2428 ;; that defun has searched backwards for a brace pair and not found one. Its
2429 ;; value is either nil or a cons (PA . FROM), where PA is the position of the
2430 ;; enclosing opening paren/brace/bracket which bounds the backwards search (or
2431 ;; nil when at top level) and FROM is where the backward search started. It
2432 ;; is reset to nil in `c-invalidate-state-cache'.
2435 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2436 ;; Lowish level functions/macros which work directly on `c-state-cache', or a
2437 ;; list of like structure.
2439 ;; Return the address of the top left brace/bracket/paren recorded in CACHE
2440 ;; (default `c-state-cache') (or nil).
2447 ;; Return the address of the latest brace/bracket/paren (whether left or
2448 ;; right) recorded in CACHE (default `c-state-cache') or nil.
2455 ;; Return the position just after the latest brace/bracket/paren (whether
2456 ;; left or right) recorded in CACHE (default `c-state-cache') or nil.
2464 ;; From the state-cache, determine the buffer position from which we might
2465 ;; scan forward to HERE to update this cache. This position will be just
2466 ;; after a paren/brace/bracket recorded in the cache, if possible, otherwise
2467 ;; return the earliest position in the accessible region which isn't within
2468 ;; a literal. If the visible portion of the buffer is entirely within a
2469 ;; literal, return NIL.
2471 ;(while (>= (or (c-state-cache-top-lparen c) 1) here)
2487 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2488 ;; Variables which keep track of preprocessor constructs.
2497 ;; These are the limits of the macro containing point at the previous call of
2498 ;; `c-parse-state', or nil.
2500 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2501 ;; Defuns which analyze the buffer, yet don't change `c-state-cache'.
2503 ;; Return the position of the opening paren/brace/bracket before HERE- which
2504 ;; matches the outermost close p/b/b between HERE+ and TOP. Except when
2505 ;; there's a macro, HERE- and HERE+ are the same. Like this:
2506 ;;
2507 ;; ............................................
2508 ;; | |
2509 ;; ( [ ( .........#macro.. ) ( ) ] )
2510 ;; ^ ^ ^ ^
2511 ;; | | | |
2512 ;; return HERE- HERE+ TOP
2513 ;;
2514 ;; If there aren't enough opening paren/brace/brackets, return the position
2515 ;; of the outermost one found, or HERE- if there are none. If there are no
2516 ;; closing p/b/bs between HERE+ and TOP, return HERE-. HERE-/+ and TOP
2517 ;; must not be inside literals. Only the accessible portion of the buffer
2518 ;; will be scanned.
2520 ;; PART 1: scan from `here+' up to `top', accumulating ")"s which enclose
2521 ;; `here'. Go round the next loop each time we pass over such a ")". These
2522 ;; probably match "("s before `here-'.
2534 ;; PART 2: Scan back before `here-' searching for the "("s
2535 ;; matching/mismatching the ")"s found above. We only need to direct the
2536 ;; caller to scan when we've encountered unmatched right parens.
2545 pos))
2548 ;; Determine the scanning strategy for adjusting `c-parse-state', attempting
2549 ;; to minimize the amount of scanning. HERE is the pertinent position in
2550 ;; the buffer, GOOD-POS is a position where `c-state-cache' (possibly with
2551 ;; its head trimmed) is known to be good, or nil if there is no such
2552 ;; position.
2553 ;;
2554 ;; The return value is a list, one of the following:
2555 ;;
2556 ;; o - ('forward START-POINT) - scan forward from START-POINT,
2557 ;; which is not less than the highest position in `c-state-cache' below HERE,
2558 ;; which is after GOOD-POS.
2559 ;; o - ('backward nil) - scan backwards (from HERE).
2560 ;; o - ('back-and-forward START-POINT) - like 'forward, but when HERE is earlier
2561 ;; than GOOD-POS.
2562 ;; o - ('IN-LIT nil) - point is inside the literal containing point-min.
2564 strategy ; 'forward, 'backward, or 'IN-LIT.
2565 start-point)
2577 start-point cache-pos)))
2581 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
2582 ;; Routines which change `c-state-cache' and associated values.
2584 ;; The region (more precisely, point-min) has changed since we
2585 ;; calculated `c-state-cache'. Amend `c-state-cache' accordingly.
2587 ;; If point-min has MOVED BACKWARDS then we drop the state completely.
2588 ;; It would be possible to do a better job here and recalculate the top
2589 ;; only.
2593 c-state-cache-good-pos c-state-min-scan-pos
2594 c-state-brace-pair-desert nil))
2596 ;; point-min has MOVED FORWARD.
2598 ;; Is the new point-min inside a (different) literal?
2603 ;; Cut off a bit of the tail from `c-state-cache'.
2605 pa)
2613 c-state-cache-good-pos c-state-min-scan-pos)
2616 )))
2621 ;; If there is a brace pair preceding FROM in the buffer, at the same level
2622 ;; of nesting (not necessarily immediately preceding), push a cons onto
2623 ;; `c-state-cache' to represent it. FROM must not be inside a literal. If
2624 ;; UPPER-LIM is non-nil, we append the highest brace pair whose "}" is below
2625 ;; UPPER-LIM.
2626 ;;
2627 ;; Return non-nil when this has been done.
2628 ;;
2629 ;; The situation it copes with is this transformation:
2630 ;;
2631 ;; OLD: { (.) {...........}
2632 ;; ^ ^
2633 ;; FROM HERE
2634 ;;
2635 ;; NEW: { {....} (.) {.........
2636 ;; ^ ^ ^
2637 ;; LOWER BRACE PAIR HERE or HERE
2638 ;;
2639 ;; This routine should be fast. Since it can get called a LOT, we maintain
2640 ;; `c-state-brace-pair-desert', a small cache of "failures", such that we
2641 ;; reduce the time wasted in repeated fruitless searches in brace deserts.
2651 ;; Don't start scanning in the middle of a CPP construct unless
2652 ;; it contains HERE - these constructs, in Emacs, are "commented
2653 ;; out" with category properties.
2656 macro-start-or-from
2657 from))
2661 ;; If we're essentially repeating a fruitless search, just give up.
2667 ;; DESERT-LIM. Avoid repeated searching through the cached desert.
2673 ;; CACHE-LIM. This limit will be necessary when an opening
2674 ;; paren at `cache-pos' has just had its matching close paren
2675 ;; inserted into the buffer. `cache-pos' continues to be a
2676 ;; search bound, even though the algorithm below would skip
2677 ;; over the new paren pair.
2686 ;; The top limit is EOB to ensure that `bra' is inside the
2687 ;; accessible part of the buffer at the next scan operation.
2690 ;; In the next pair of nested loops, the inner one moves back past a
2691 ;; pair of (mis-)matching parens or brackets; the outer one moves
2692 ;; back over a sequence of unmatched close brace/paren/bracket each
2693 ;; time round.
2709 ; brace, ignore it.
2712 ;; We've found the desired brace-pair.
2724 ;; We haven't found a brace pair. Record this in the cache.
2727 bra
2732 ;; If BRA+1 is nil, do nothing. Otherwise, BRA+1 is the buffer position
2733 ;; following a {, and that brace has a (mis-)matching } (or ]), and we
2734 ;; "push" "a" brace pair onto `c-state-cache'.
2735 ;;
2736 ;; Here "push" means overwrite the top element if it's itself a brace-pair,
2737 ;; otherwise push it normally.
2738 ;;
2739 ;; The brace pair we push is normally the one surrounding BRA+1, but if the
2740 ;; latter is inside a macro, not being a macro containing
2741 ;; MACRO-START-OR-HERE, we scan backwards through the buffer for a non-macro
2742 ;; base pair. This latter case is assumed to be rare.
2743 ;;
2744 ;; Note: POINT is not preserved in this routine.
2754 c-state-cache)))
2755 ;; N.B. This defsubst codes one method for the simple, normal case,
2756 ;; and a more sophisticated, slower way for the general case. Don't
2757 ;; eliminate this defsubst - it's a speed optimization.
2761 ;; Scan the buffer from FROM to HERE, adding elements into `c-state-cache'
2762 ;; for braces etc. Return a candidate for `c-state-cache-good-pos'.
2763 ;;
2764 ;; FROM must be after the latest brace/paren/bracket in `c-state-cache', if
2765 ;; any. Typically, it is immediately after it. It must not be inside a
2766 ;; literal.
2772 here)))
2775 ; Is actually the pos. to scan for a (/{/[ from,
2776 ; which sometimes is after a silly )/}/].
2779 ; good-pos.
2787 ;; Each time round the following loop, we enter a successively deeper
2788 ;; level of brace/paren nesting. (Except sometimes we "continue at
2789 ;; the existing level".) `pa+1' is a pos inside an opening
2790 ;; brace/paren/bracket, usually just after it.
2793 ;; Each time round the next loop moves forward over an opening then
2794 ;; a closing brace/bracket/paren. This loop is white hot, so it
2795 ;; plays ugly tricks to go fast. DON'T PUT ANYTHING INTO THIS
2796 ;; LOOP WHICH ISN'T ABSOLUTELY NECESSARY!!! It terminates when a
2797 ;; call of `scan-lists' signals an error, which happens when there
2798 ;; are no more b/b/p's to scan.
2809 ;; We've just entered a deeper nesting level.
2811 ;; Insert the brace pair (if present) and the single open
2812 ;; paren/brace/bracket into `c-state-cache' It cannot be
2813 ;; inside a macro, except one around point, because of what
2814 ;; `c-neutralize-syntax-in-CPP' has done.
2816 ;; Insert the opening brace/bracket/paren position.
2818 ;; Clear admin stuff for the next more nested part of the scan.
2822 ;; All open p/b/b's at this nesting level, if any, have probably
2823 ;; been closed by matching/mismatching ones. We're probably
2824 ;; finished - we just need to check for having found an
2825 ;; unmatched )/}/], which we ignore. Such a )/}/] can't be in a
2826 ;; macro, due the action of `c-neutralize-syntax-in-CPP'.
2829 ;; Record the final, innermost, brace-pair if there is one.
2832 ;; Determine a good pos
2840 here-bol))
2849 ;; Remove stale entries from the `c-cache-state', i.e. those which will
2850 ;; not be in it when it is amended for position HERE. This may involve
2851 ;; replacing a CONS element for a brace pair containing HERE with its car.
2852 ;; Additionally, the "outermost" open-brace entry before HERE will be
2853 ;; converted to a cons if the matching close-brace is below HERE.
2854 ;;
2855 ;; START-POINT is a "maximal" "safe position" - there must be no open
2856 ;; parens/braces/brackets between START-POINT and HERE.
2857 ;;
2858 ;; As a second thing, calculate the result of parse-partial-sexp at
2859 ;; PPS-POINT, w.r.t. START-POINT. The motivation here is that
2860 ;; `c-state-cache-good-pos' may become PPS-POINT, but the caller may need to
2861 ;; adjust it to get outside a string/comment. (Sorry about this! The code
2862 ;; needs to be FAST).
2863 ;;
2864 ;; Return a list (GOOD-POS SCAN-BACK-POS CONS-SEPARATED PPS-STATE), where
2865 ;; o - GOOD-POS is a position where the new value `c-state-cache' is known
2866 ;; to be good (we aim for this to be as high as possible);
2867 ;; o - SCAN-BACK-POS, if not nil, indicates there may be a brace pair
2868 ;; preceding POS which needs to be recorded in `c-state-cache'. It is a
2869 ;; position to scan backwards from. It is the position of the "{" of the
2870 ;; last element to be removed from `c-state-cache', when that elt is a
2871 ;; cons, otherwise nil.
2872 ;; o - CONS-SEPARATED is t when a cons element in `c-state-cache' has been
2873 ;; replaced by its car because HERE lies inside the brace pair represented
2874 ;; by the cons.
2875 ;; o - PPS-STATE is the parse-partial-sexp state at PPS-POINT.
2885 ; start-point or nil
2892 ; HERE), or nil.
2898 pps-state ; Will be 9 or 10 elements long.
2899 pos
2900 upper-lim ; ,beyond which `c-state-cache' entries are removed
2901 scan-back-pos
2902 cons-separated
2903 pair-beg pps-point-state target-depth)
2905 ;; Remove entries beyond HERE. Also remove any entries inside
2906 ;; a macro, unless HERE is in the same macro.
2911 here
2917 ;; If `upper-lim' is inside the last recorded brace pair, remove its
2918 ;; RBrace and indicate we'll need to search backwards for a previous
2919 ;; brace pair.
2925 cons-separated t))
2927 ;; The next loop jumps forward out of a nested level of parens each
2928 ;; time round; the corresponding elements in `c-state-cache' are
2929 ;; removed. `pos' is just after the brace-pair or the open paren at
2930 ;; (car c-state-cache). There can be no open parens/braces/brackets
2931 ;; between `start-point'/`start-point-actual-macro-start' and HERE,
2932 ;; due to the interface spec to this function.
2935 in-macro-start)))
2937 ; marked by a `category' text property.
2938 start-point))
2949 ;; Do nothing when we've merely reached pps-point.
2950 )
2952 ;; Scan!
2956 target-depth
2957 nil pps-state))
2967 ;; We've just moved out of the paren pair containing the brace-pair
2968 ;; at (car c-state-cache). `pair-beg' is where the open paren is,
2969 ;; and is potentially where the open brace of a cons in
2970 ;; c-state-cache will be.
2976 ; containing Lparen
2980 ; together with enclosed brace pair.
2981 ;; (t nil) ; Ignore an unmated Rparen.
2982 )))
2986 nil nil ; TARGETDEPTH, STOPBEFORE
2987 pps-state)))
2989 ;; If the last paren pair we moved out of was actually a brace pair,
2990 ;; insert it into `c-state-cache'.
2995 c-state-cache)))
3000 ;; Strip stale elements of `c-state-cache' by moving backwards through the
3001 ;; buffer, and inform the caller of the scenario detected.
3002 ;;
3003 ;; HERE is the position we're setting `c-state-cache' for.
3004 ;; CACHE-POS (a locally bound variable) is just after the latest recorded
3005 ;; position in `c-state-cache' before HERE, or a position at or near
3006 ;; point-min which isn't in a literal.
3007 ;;
3008 ;; This function must only be called only when (> `c-state-cache-good-pos'
3009 ;; HERE). Usually the gap between CACHE-POS and HERE is large. It is thus
3010 ;; optimized to eliminate (or minimize) scanning between these two
3011 ;; positions.
3012 ;;
3013 ;; Return a three element list (GOOD-POS SCAN-BACK-POS FWD-FLAG), where:
3014 ;; o - GOOD-POS is a "good position", where `c-state-cache' is valid, or
3015 ;; could become so after missing elements are inserted into
3016 ;; `c-state-cache'. This is JUST AFTER an opening or closing
3017 ;; brace/paren/bracket which is already in `c-state-cache' or just before
3018 ;; one otherwise. exceptionally (when there's no such b/p/b handy) the BOL
3019 ;; before `here''s line, or the start of the literal containing it.
3020 ;; o - SCAN-BACK-POS, if non-nil, indicates there may be a brace pair
3021 ;; preceding POS which isn't recorded in `c-state-cache'. It is a position
3022 ;; to scan backwards from.
3023 ;; o - FWD-FLAG, if non-nil, indicates there may be parens/braces between
3024 ;; POS and HERE which aren't recorded in `c-state-cache'.
3025 ;;
3026 ;; The comments in this defun use "paren" to mean parenthesis or square
3027 ;; bracket (as contrasted with a brace), and "(" and ")" likewise.
3028 ;;
3029 ;; . {..} (..) (..) ( .. { } ) (...) ( .... . ..)
3030 ;; | | | | | |
3031 ;; CP E here D C good
3034 pa ren ; positions of "(" and ")"
3035 dropped-cons ; whether the last element dropped from `c-state-cache'
3036 ; was a cons (representing a brace-pair)
3037 good-pos ; see above.
3038 lit ; (START . END) of a literal containing some point.
3039 here-lit-start here-lit-end ; bounds of literal containing `here'
3040 ; or `here' itself.
3045 ;; Remove completely irrelevant entries from `c-state-cache'.
3051 ;; At this stage, (> pos here);
3052 ;; (< (c-state-cache-top-lparen) here) (or is nil).
3057 ;; CASE 1: The top of the cache is a brace pair which now encloses
3058 ;; `here'. As good-pos, return the address. of the "{". Since we've no
3059 ;; knowledge of what's inside these braces, we have no alternative but
3060 ;; to direct the caller to scan the buffer from the opening brace.
3064 ; entry into a { entry, so the caller needs to
3065 ; search for a brace pair before the {.
3067 ;; `here' might be inside a literal. Check for this.
3072 ;; Has `here' just "newly entered" a macro?
3085 ;; `here' might be nested inside any depth of parens (or brackets but
3086 ;; not braces). Scan backwards to find the outermost such opening
3087 ;; paren, if there is one. This will be the scan position to return.
3094 ;; CASE 2: Address of outermost ( or [ which now encloses `here', but
3095 ;; didn't enclose the (previous) `c-state-cache-good-pos'. If there is
3096 ;; a brace pair preceding this, it will already be in `c-state-cache',
3097 ;; unless there was a brace pair after it, i.e. there'll only be one to
3098 ;; scan for if we've just deleted one.
3101 ;; `here' isn't enclosed in a (previously unrecorded) bracket/paren.
3102 ;; Further forward scanning isn't needed, but we still need to find a
3103 ;; GOOD-POS. Step out of all enclosing "("s on HERE's line.
3113 ;; CASE 3: After a }/)/] before `here''s BOL.
3117 ;; CASE 4; Best of a bad job: BOL before `here-bol', or beginning of
3118 ;; literal containing it.
3123 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3124 ;; Externally visible routines.
3128 c-state-cache-good-pos 1
3129 c-state-nonlit-pos-cache nil
3130 c-state-nonlit-pos-cache-limit 1
3131 c-state-semi-nonlit-pos-cache nil
3132 c-state-semi-nonlit-pos-cache-limit 1
3133 c-state-brace-pair-desert nil
3134 c-state-point-min 1
3135 c-state-point-min-lit-type nil
3136 c-state-point-min-lit-start nil
3137 c-state-min-scan-pos 1
3138 c-state-old-cpp-beg nil
3139 c-state-old-cpp-end nil)
3142 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3143 ;; Debugging routines to dump `c-state-cache' in a "replayable" form.
3144 ;; (defmacro c-sc-de (elt) ; "c-state-cache-dump-element"
3145 ;; `(format ,(concat "(setq " (symbol-name elt) " %s) ") ,elt))
3146 ;; (defmacro c-sc-qde (elt) ; "c-state-cache-quote-dump-element"
3147 ;; `(format ,(concat "(setq " (symbol-name elt) " '%s) ") ,elt))
3148 ;; (defun c-state-dump ()
3149 ;; ;; For debugging.
3150 ;; ;(message
3151 ;; (concat
3152 ;; (c-sc-qde c-state-cache)
3153 ;; (c-sc-de c-state-cache-good-pos)
3154 ;; (c-sc-qde c-state-nonlit-pos-cache)
3155 ;; (c-sc-de c-state-nonlit-pos-cache-limit)
3156 ;; (c-sc-qde c-state-brace-pair-desert)
3157 ;; (c-sc-de c-state-point-min)
3158 ;; (c-sc-de c-state-point-min-lit-type)
3159 ;; (c-sc-de c-state-point-min-lit-start)
3160 ;; (c-sc-de c-state-min-scan-pos)
3161 ;; (c-sc-de c-state-old-cpp-beg)
3162 ;; (c-sc-de c-state-old-cpp-end)))
3163 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
3166 ;; Invalidate all info on `c-state-cache' that applies to the buffer at HERE
3167 ;; or higher and set `c-state-cache-good-pos' accordingly. The cache is
3168 ;; left in a consistent state.
3169 ;;
3170 ;; This is much like `c-whack-state-after', but it never changes a paren
3171 ;; pair element into an open paren element. Doing that would mean that the
3172 ;; new open paren wouldn't have the required preceding paren pair element.
3173 ;;
3174 ;; This function is called from c-after-change.
3176 ;; The caches of non-literals:
3177 ;; Note that we use "<=" for the possibility of the second char of a two-char
3178 ;; comment opener being typed; this would invalidate any cache position at
3179 ;; HERE.
3185 ;; `c-state-cache':
3186 ;; Case 1: if `here' is in a literal containing point-min, everything
3187 ;; becomes (or is already) nil.
3191 c-state-cache-good-pos nil
3192 c-state-min-scan-pos nil)
3194 ;; Truncate `c-state-cache' and set `c-state-cache-good-pos' to a value
3195 ;; below `here'. To maintain its consistency, we may need to insert a new
3196 ;; brace pair.
3199 too-high-pa ; recorded {/(/[ next above here, or nil.
3200 dropped-cons ; was the last removed element a brace pair?
3201 pa)
3202 ;; The easy bit - knock over-the-top bits off `c-state-cache'.
3209 ;; Do we need to add in an earlier brace pair, having lopped one off?
3216 ;; The brace-pair desert marker:
3224 ;; Find and record all noteworthy parens between some good point earlier in
3225 ;; the file and point. That good point is at least the beginning of the
3226 ;; top-level construct we are in, or the beginning of the preceding
3227 ;; top-level construct if we aren't in one.
3228 ;;
3229 ;; The returned value is a list of the noteworthy parens with the last one
3230 ;; first. If an element in the list is an integer, it's the position of an
3231 ;; open paren (of any type) which has not been closed before the point. If
3232 ;; an element is a cons, it gives the position of a closed BRACE paren
3233 ;; pair[*]; the car is the start brace position and the cdr is the position
3234 ;; following the closing brace. Only the last closed brace paren pair
3235 ;; before each open paren and before the point is recorded, and thus the
3236 ;; state never contains two cons elements in succession. When a close brace
3237 ;; has no matching open brace (e.g., the matching brace is outside the
3238 ;; visible region), it is not represented in the returned value.
3239 ;;
3240 ;; [*] N.B. The close "brace" might be a mismatching close bracket or paren.
3241 ;; This defun explicitly treats mismatching parens/braces/brackets as
3242 ;; matching. It is the open brace which makes it a "brace" pair.
3243 ;;
3244 ;; If POINT is within a macro, open parens and brace pairs within
3245 ;; THIS macro MIGHT be recorded. This depends on whether their
3246 ;; syntactic properties have been suppressed by
3247 ;; `c-neutralize-syntax-in-CPP'. This might need fixing (2008-12-11).
3248 ;;
3249 ;; Currently no characters which are given paren syntax with the
3250 ;; syntax-table property are recorded, i.e. angle bracket arglist
3251 ;; parens are never present here. Note that this might change.
3252 ;;
3253 ;; BUG: This function doesn't cope entirely well with unbalanced
3254 ;; parens in macros. (2008-12-11: this has probably been resolved
3255 ;; by the function `c-neutralize-syntax-in-CPP'.) E.g. in the
3256 ;; following case the brace before the macro isn't balanced with the
3257 ;; one after it:
3258 ;;
3259 ;; {
3260 ;; #define X {
3261 ;; }
3262 ;;
3263 ;; Note to maintainers: this function DOES get called with point
3264 ;; within comments and strings, so don't assume it doesn't!
3265 ;;
3266 ;; This function might do hidden buffer changes.
3269 open-paren-in-column-0-is-defun-start
3270 strategy ; 'forward, 'backward etc..
3271 ;; Candidate positions to start scanning from:
3272 cache-pos ; highest position below HERE already existing in
3273 ; cache (or 1).
3274 good-pos
3275 start-point ; (when scanning forward) a place below HERE where there
3276 ; are no open parens/braces between it and HERE.
3277 bopl-state
3278 res
3279 cons-separated
3281 ;; If POINT-MIN has changed, adjust the cache
3285 ;; Strategy?
3290 ;; SCAN!
3298 ; start-point)
3308 good-pos)))
3320 good-pos)))
3324 c-state-cache-good-pos nil))))
3326 c-state-cache)
3329 ;; This is a wrapper over `c-invalidate-state-cache-1'.
3330 ;;
3331 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3332 ;; of all parens in preprocessor constructs, except for any such construct
3333 ;; containing point. We can then call `c-invalidate-state-cache-1' without
3334 ;; worrying further about macros and template delimiters.
3339 c-state-old-cpp-beg
3346 ;; If PLACE is non-nil, return a marker marking it, otherwise nil.
3347 ;; We (re)use MARKER.
3353 ;; This is a wrapper over `c-parse-state-1'. See that function for a
3354 ;; description of the functionality and return value.
3355 ;;
3356 ;; It suppresses the syntactic effect of the < and > (template) brackets and
3357 ;; of all parens in preprocessor constructs, except for any such construct
3358 ;; containing point. We can then call `c-parse-state-1' without worrying
3359 ;; further about macros and template delimiters.
3366 here-cpp-beg)
3368 ;; FIXME!!! Put in a `condition-case' here to protect the integrity of the
3369 ;; subsystem.
3374 here-cpp-beg here-cpp-end
3380 c-state-old-cpp-end
3383 ;; Debug tool to catch cache inconsistencies. This is called from
3384 ;; 000tests.el.
3401 val))))
3403 c-state-cache-good-pos
3404 c-state-nonlit-pos-cache
3405 c-state-nonlit-pos-cache-limit
3406 c-state-semi-nonlit-pos-cache
3407 c-state-semi-nonlit-pos-cache-limit
3408 c-state-brace-pair-desert
3409 c-state-point-min
3410 c-state-point-min-lit-type
3411 c-state-point-min-lit-start
3412 c-state-min-scan-pos
3413 c-state-old-cpp-beg
3414 c-state-old-cpp-end
3415 c-parse-state-point))))
3433 conses-not-ok))
3450 ;; The cache can actually go further back due to the ad-hoc way
3451 ;; the first paren is found, so try to whack off a bit of its
3452 ;; start before complaining.
3453 ;; (save-excursion
3454 ;; (goto-char (or (c-least-enclosing-brace res2) (point)))
3455 ;; (c-beginning-of-defun-1)
3456 ;; (while (not (or (bobp) (eq (char-after) ?{)))
3457 ;; (c-beginning-of-defun-1))
3458 ;; (unless (equal (c-whack-state-before (point) res1) res2)
3459 ;; (message (concat "c-parse-state inconsistency at %s: "
3460 ;; "using cache: %s, from scratch: %s")
3461 ;; here res1 res2)))
3464 here res1 res2)
3470 here res1)
3475 res2 ; res1 correct a cascading series of errors ASAP
3476 ))
3482 'c-debug-parse-state
3490 \f
3492 ;; Whack off any state information from PAREN-STATE which lies
3493 ;; before BUFPOS. Not destructive on PAREN-STATE.
3496 car)
3507 ;; Whack off any state information from PAREN-STATE which lies at or
3508 ;; after BUFPOS. Not destructive on PAREN-STATE.
3513 ;; just check the car, because in a balanced brace
3514 ;; expression, it must be impossible for the corresponding
3515 ;; close brace to be before point, but the open brace to
3516 ;; be after.
3518 nil ; whack it off
3520 ;; its possible that the open brace is before
3521 ;; bufpos, but the close brace is after. In that
3522 ;; case, convert this to a non-cons element. The
3523 ;; rest of the state is before bufpos, so we're
3524 ;; done.
3526 ;; we know that both the open and close braces are
3527 ;; before bufpos, so we also know that everything else
3528 ;; on state is before bufpos.
3531 nil ; whack it off
3532 ;; it's before bufpos, so everything else should too.
3535 nil)))
3538 ;; Return the bufpos of the innermost enclosing open paren before
3539 ;; bufpos, or nil if none was found.
3549 enclosingp))
3552 ;; Return the bufpos of the outermost enclosing open paren, or nil
3553 ;; if none was found.
3560 pos))
3563 ;; Return the closest "safe" position recorded on PAREN-STATE that
3564 ;; is higher up than BUFPOS. Return nil if PAREN-STATE doesn't
3565 ;; contain any. Return nil if BUFPOS is nil, which is useful to
3566 ;; find the closest limit before a given limit that might be nil.
3567 ;;
3568 ;; A "safe" position is a position at or after a recorded open
3569 ;; paren, or after a recorded close paren. The returned position is
3570 ;; thus either the first position after a close brace, or the first
3571 ;; position after an enclosing paren, or at the enclosing paren in
3572 ;; case BUFPOS is immediately after it.
3582 ;; See below.
3585 ;; elem is the position at and not after the opening paren, so
3586 ;; we can go forward one more step unless it's equal to
3587 ;; bufpos. This is useful in some cases avoid an extra paren
3588 ;; level between the safe position and bufpos.
3593 ;; This is used for `font-lock-beginning-of-syntax-function'. It
3594 ;; goes to the closest previous point that is known to be outside
3595 ;; any string literal or comment. `c-state-cache' is used if it has
3596 ;; a position in the vicinity.
3598 elem
3601 ;; Note: Similar code in `c-safe-position'. The
3602 ;; difference is that we accept a safe position at
3603 ;; the point and don't bother to go forward past open
3604 ;; parens.
3619 ;; The position is far back. Try `c-beginning-of-defun-1'
3620 ;; (although we can't be entirely sure it will go to a position
3621 ;; outside a comment or string in current emacsen). FIXME:
3622 ;; Consult `syntax-ppss' here.
3627 \f
3628 ;; Tools for scanning identifiers and other tokens.
3631 "Return non-nil if the point is on or directly after an identifier.
3632 Keywords are recognized and not considered identifiers. If an
3633 identifier is detected, the returned value is its starting position.
3634 If an identifier ends at the point and another begins at it \(can only
3635 happen in Pike) then the point for the preceding one is returned.
3637 Note that this function might do hidden buffer changes. See the
3638 comment at the start of cc-engine.el for more info."
3640 ;; FIXME: Shouldn't this function handle "operator" in C++?
3647 ;; Check for a normal (non-keyword) identifier.
3653 ;; Handle the `<operator> syntax in Pike.
3657 t
3664 ;; Handle the "operator +" syntax in C++.
3681 )))
3684 ;; If the point is at the end of a symbol then skip backward to the
3685 ;; beginning of it. Don't move otherwise. Return non-nil if point
3686 ;; moved.
3687 ;;
3688 ;; This function might do hidden buffer changes.
3691 ;; Handle the `<operator> syntax in Pike.
3697 t
3699 nil)))))
3702 ;; Move to the beginning of the current token. Do not move if not
3703 ;; in the middle of one. BACK-LIMIT may be used to bound the
3704 ;; backward search; if given it's assumed to be at the boundary
3705 ;; between two tokens. Return non-nil if the point is moved, nil
3706 ;; otherwise.
3707 ;;
3708 ;; This function might do hidden buffer changes.
3715 ;; `c-nonsymbol-token-regexp' should always match
3716 ;; since we've skipped backward over punctuation
3717 ;; or paren syntax, but consume one char in case
3718 ;; it doesn't so that we don't leave point before
3719 ;; some earlier incorrect token.
3726 ;; Move to the end of the current token. Do not move if not in the
3727 ;; middle of one. BACK-LIMIT may be used to bound the backward
3728 ;; search; if given it's assumed to be at the boundary between two
3729 ;; tokens. Return non-nil if the point is moved, nil otherwise.
3730 ;;
3731 ;; This function might do hidden buffer changes.
3739 ;; `c-nonsymbol-token-regexp' should always match since
3740 ;; we've skipped backward over punctuation or paren
3741 ;; syntax, but move forward in case it doesn't so that
3742 ;; we don't leave point earlier than we started with.
3758 "Move forward by tokens.
3759 A token is defined as all symbols and identifiers which aren't
3761 treated properly). Point is always either left at the beginning of a
3762 token or not moved at all. COUNT specifies the number of tokens to
3763 move; a negative COUNT moves in the opposite direction. A COUNT of 0
3764 moves to the next token beginning only if not already at one. If
3765 BALANCED is true, move over balanced parens, otherwise move into them.
3766 Also, if BALANCED is true, never move out of an enclosing paren.
3768 LIMIT sets the limit for the movement and defaults to the point limit.
3769 The case when LIMIT is set in the middle of a token, comment or macro
3770 is handled correctly, i.e. the point won't be left there.
3772 Return the number of tokens left to move \(positive or negative). If
3773 BALANCED is true, a move over a balanced paren counts as one. Note
3774 that if COUNT is 0 and no appropriate token beginning is found, 1 will
3775 be returned. Thus, a return value of 0 guarantees that point is at
3776 the requested position and a return value less \(without signs) than
3777 COUNT guarantees that point is at the beginning of some token.
3779 Note that this function might do hidden buffer changes. See the
3780 comment at the start of cc-engine.el for more info."
3787 c-jump-syntax-balanced
3788 c-jump-syntax-unbalanced))
3793 ;; If count is zero we should jump if in the middle of a token.
3800 ;; Skip whitespace. Count this as a move if we did in
3801 ;; fact move.
3805 ;; Moved out of bounds. Make sure the returned count isn't zero.
3810 ;; Use `condition-case' to avoid having the limit tests
3811 ;; inside the loop.
3819 t)
3822 t)
3823 ;; `c-nonsymbol-token-regexp' above should always
3824 ;; match if there are correct tokens. Try to
3825 ;; widen to see if the limit was set in the
3826 ;; middle of one, else fall back to treating
3827 ;; the offending thing as a one character token.
3832 nil)
3835 t))))
3845 count)))
3848 "Move backward by tokens.
3849 See `c-forward-token-2' for details."
3857 c-jump-syntax-balanced
3858 c-jump-syntax-unbalanced))
3862 ;; The count is zero so try to skip to the beginning of the
3863 ;; current token.
3867 ;; The limit is inside the same token, so return 1.
3870 ;; We're not in the middle of a token. If there's
3871 ;; whitespace after the point then we must move backward,
3872 ;; so set count to 1 in that case.
3874 ;; If we're looking at a '#' that might start a cpp
3875 ;; directive then we have to do a more elaborate check.
3888 ;; Use `condition-case' to avoid having to check for buffer
3889 ;; limits in `backward-char', `scan-sexps' and `goto-char' below.
3898 ;; This can be very inefficient if there's a long
3899 ;; sequence of operator tokens without any separation.
3900 ;; That doesn't happen in practice, anyway.
3910 count)))
3913 "Like `c-forward-token-2' but doesn't treat multicharacter operator
3915 characters are jumped over character by character. This function is
3916 for compatibility only; it's only a wrapper over `c-forward-token-2'."
3921 "Like `c-backward-token-2' but doesn't treat multicharacter operator
3923 characters are jumped over character by character. This function is
3924 for compatibility only; it's only a wrapper over `c-backward-token-2'."
3928 \f
3929 ;; Tools for doing searches restricted to syntactically relevant text.
3932 paren-level not-inside-token
3933 lookbehind-submatch)
3934 "Like `re-search-forward', but only report matches that are found
3935 in syntactically significant text. I.e. matches in comments, macros
3936 or string literals are ignored. The start point is assumed to be
3937 outside any comment, macro or string literal, or else the content of
3938 that region is taken as syntactically significant text.
3940 If PAREN-LEVEL is non-nil, an additional restriction is added to
3941 ignore matches in nested paren sexps. The search will also not go
3942 outside the current list sexp, which has the effect that if the point
3943 should be moved to BOUND when no match is found \(i.e. NOERROR is
3944 neither nil nor t), then it will be at the closing paren if the end of
3945 the current list sexp is encountered first.
3947 If NOT-INSIDE-TOKEN is non-nil, matches in the middle of tokens are
3948 ignored. Things like multicharacter operators and special symbols
3950 constants.
3952 If LOOKBEHIND-SUBMATCH is non-nil, it's taken as a number of a
3953 subexpression in REGEXP. The end of that submatch is used as the
3954 position to check for syntactic significance. If LOOKBEHIND-SUBMATCH
3955 isn't used or if that subexpression didn't match then the start
3957 subexpression is never tested before the starting position, so it
3961 subexpression can match the end of nonwhite syntactic whitespace,
3962 i.e. the end of comments or cpp directives. This since the function
3963 skips over such things before resuming the search. It's on the other
3965 matches syntactic whitespace.
3967 Bug: Unbalanced parens inside cpp directives are currently not handled
3968 correctly \(i.e. they don't get ignored as they should) when
3969 PAREN-LEVEL is set.
3971 Note that this function might do hidden buffer changes. See the
3972 comment at the start of cc-engine.el for more info."
3977 ;;(message "c-syntactic-re-search-forward %s %s %S" (point) bound regexp)
3980 tmp
3981 ;; Start position for the last search.
3982 search-pos
3983 ;; The `parse-partial-sexp' state between the start position
3984 ;; and the point.
3985 state
3986 ;; The current position after the last state update. The next
3987 ;; `parse-partial-sexp' continues from here.
3989 ;; The position at which to check the state and the state
3990 ;; there. This is separate from `state-pos' since we might
3991 ;; need to back up before doing the next search round.
3992 check-pos check-state
3993 ;; Last position known to end a token.
3995 ;; Set when a valid match is found.
3996 found)
4014 state-pos check-pos paren-level nil state))
4016 check-state state))
4018 ;; NOTE: If we got a look behind subexpression and get
4019 ;; an insignificant match in something that isn't
4020 ;; syntactic whitespace (i.e. strings or in nested
4021 ;; parentheses), then we can never skip more than a
4022 ;; single character from the match start position
4023 ;; (i.e. `state-pos' here) before continuing the
4024 ;; search. That since the look behind subexpression
4025 ;; might match the end of the insignificant region in
4026 ;; the next search.
4030 ;; Match inside a line comment. Skip to eol. Use
4031 ;; `re-search-forward' instead of `skip-chars-forward' to get
4032 ;; the right bound behavior.
4036 ;; Match inside a block comment. Skip to the '*/'.
4043 ;; Match in the middle of the opener of a block or line
4044 ;; comment.
4049 ;; The last `parse-partial-sexp' above might have
4050 ;; stopped short of the real check position if the end
4051 ;; of the current sexp was encountered in paren-level
4052 ;; mode. The checks above are always false in that
4053 ;; case, and since they can do better skipping in
4054 ;; lookbehind-submatch mode, we do them before
4055 ;; checking the paren level.
4059 ;; Check the paren level first since we're short of the
4060 ;; syntactic checking position if the end of the
4061 ;; current sexp was encountered by `parse-partial-sexp'.
4064 ;; Inside a nested paren sexp.
4066 ;; See the NOTE above.
4068 ;; Skip out of the paren quickly.
4072 ;; Have exited the current paren sexp.
4075 ;; The last `parse-partial-sexp' call above
4076 ;; has left us just after the closing paren
4077 ;; in this case, so we can modify the bound
4078 ;; to leave the point at the right position
4079 ;; upon return.
4081 nil)
4085 ;; Match inside a string.
4088 ;; See the NOTE above.
4090 ;; Skip to the end of the string before continuing.
4099 continue)))
4104 ;; Match inside a macro. Skip to the end of it.
4112 (< check-pos
4118 ;; Inside a token.
4120 ;; See the NOTE above.
4125 ;; A real match.
4127 nil)))
4129 ;; Should loop to search again, but take care to avoid
4130 ;; looping on the same spot.
4134 nil
4137 t))))
4143 ;;(message "c-syntactic-re-search-forward done %s" (or (match-end 0) (point)))
4150 ;; Search failed. Set point as appropriate.
4154 nil)))
4159 ;; Return the start of the literal point is in, or nil.
4160 ;; We read and write the variables `safe-pos', `safe-pos-list', `state'
4161 ;; bound in the caller.
4163 ;; Use `parse-partial-sexp' from a safe position down to the point to check
4164 ;; if it's outside comments and strings.
4167 ;; Pick a safe position as close to the point as possible.
4168 ;;
4169 ;; FIXME: Consult `syntax-ppss' here if our cache doesn't give a good
4170 ;; position.
4183 ;; Cache positions along the way to use if we have to back up more. We
4184 ;; cache every closing paren on the same level. If the paren cache is
4185 ;; relevant in this region then we're typically already on the same
4186 ;; level as the target position. Note that we might cache positions
4187 ;; after opening parens in case safe-pos is in a nested list. That's
4188 ;; both uncommon and harmless.
4196 ;; If the state contains the start of the containing sexp we cache that
4197 ;; position too, so that parse-partial-sexp in the next run has a bigger
4198 ;; chance of starting at the same level as the target position and thus
4199 ;; will get more good safe positions into the list.
4205 ;; Inside string or comment. Continue search at the
4206 ;; beginning of it.
4210 "Like `skip-chars-backward' but only look at syntactically relevant chars,
4211 i.e. don't stop at positions inside syntactic whitespace or string
4212 literals. Preprocessor directives are also ignored, with the exception
4213 of the one that the point starts within, if any. If LIMIT is given,
4214 it's assumed to be at a syntactically relevant position.
4216 If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4217 sexps, and the search will also not go outside the current paren sexp.
4218 However, if LIMIT or the buffer limit is reached inside a nested paren
4219 then the point will be left at the limit.
4221 Non-nil is returned if the point moved, nil otherwise.
4223 Note that this function might do hidden buffer changes. See the
4224 comment at the start of cc-engine.el for more info."
4227 state-2
4228 ;; A list of syntactically relevant positions in descending
4229 ;; order. It's used to avoid scanning repeatedly over
4230 ;; potentially large regions with `parse-partial-sexp' to verify
4231 ;; each position. Used in `c-ssb-lit-begin'
4232 safe-pos-list
4233 ;; The result from `c-beginning-of-macro' at the start position or the
4234 ;; start position itself if it isn't within a macro. Evaluated on
4235 ;; demand.
4236 start-macro-beg
4237 ;; The earliest position after the current one with the same paren
4238 ;; level. Used only when `paren-level' is set.
4239 lit-beg
4244 ;; The next loop "tries" to find the end point each time round,
4245 ;; loops when it hasn't succeeded.
4253 ;; Don't stop inside a literal
4256 t)
4264 ;; Not at the right level.
4267 ;; We stop above if we go out of a paren.
4268 ;; Now check whether it precedes or is
4269 ;; nested in the starting sexp.
4273 pps-end-pos paren-level-pos
4274 nil nil state-2))
4277 ;; We've stopped short of the starting position
4278 ;; so the hit was inside a nested list. Go up
4279 ;; until we are at the right level.
4288 nil)
4289 t))
4292 nil))
4294 ;; The hit was outside the list at the start
4295 ;; position. Go to the start of the list and exit.
4297 nil))
4300 ;; Inside a macro.
4308 t
4310 ;; It's inside the same macro we started in so it's
4311 ;; a relevant match.
4313 nil))))))
4317 ;; Skip syntactic ws afterwards so that we don't stop at the
4318 ;; end of a comment if `skip-chars' is something like "^/".
4322 ;; We might want to extend this with more useful return values in
4323 ;; the future.
4326 ;; The following is an alternative implementation of
4327 ;; `c-syntactic-skip-backward' that uses backward movement to keep
4328 ;; track of the syntactic context. It turned out to be generally
4329 ;; slower than the one above which uses forward checks from earlier
4330 ;; safe positions.
4331 ;;
4332 ;;(defconst c-ssb-stop-re
4333 ;; ;; The regexp matching chars `c-syntactic-skip-backward' needs to
4334 ;; ;; stop at to avoid going into comments and literals.
4335 ;; (concat
4336 ;; ;; Match comment end syntax and string literal syntax. Also match
4337 ;; ;; '/' for block comment endings (not covered by comment end
4338 ;; ;; syntax).
4339 ;; "\\s>\\|/\\|\\s\""
4340 ;; (if (memq 'gen-string-delim c-emacs-features)
4341 ;; "\\|\\s|"
4342 ;; "")
4343 ;; (if (memq 'gen-comment-delim c-emacs-features)
4344 ;; "\\|\\s!"
4345 ;; "")))
4346 ;;
4347 ;;(defconst c-ssb-stop-paren-re
4348 ;; ;; Like `c-ssb-stop-re' but also stops at paren chars.
4349 ;; (concat c-ssb-stop-re "\\|\\s(\\|\\s)"))
4350 ;;
4351 ;;(defconst c-ssb-sexp-end-re
4352 ;; ;; Regexp matching the ending syntax of a complex sexp.
4353 ;; (concat c-string-limit-regexp "\\|\\s)"))
4354 ;;
4355 ;;(defun c-syntactic-skip-backward (skip-chars &optional limit paren-level)
4356 ;; "Like `skip-chars-backward' but only look at syntactically relevant chars,
4357 ;;i.e. don't stop at positions inside syntactic whitespace or string
4358 ;;literals. Preprocessor directives are also ignored. However, if the
4359 ;;point is within a comment, string literal or preprocessor directory to
4360 ;;begin with, its contents is treated as syntactically relevant chars.
4361 ;;If LIMIT is given, it limits the backward search and the point will be
4362 ;;left there if no earlier position is found.
4363 ;;
4364 ;;If PAREN-LEVEL is non-nil, the function won't stop in nested paren
4365 ;;sexps, and the search will also not go outside the current paren sexp.
4366 ;;However, if LIMIT or the buffer limit is reached inside a nested paren
4367 ;;then the point will be left at the limit.
4368 ;;
4369 ;;Non-nil is returned if the point moved, nil otherwise.
4370 ;;
4371 ;;Note that this function might do hidden buffer changes. See the
4372 ;;comment at the start of cc-engine.el for more info."
4373 ;;
4374 ;; (save-restriction
4375 ;; (when limit
4376 ;; (narrow-to-region limit (point-max)))
4377 ;;
4378 ;; (let ((start (point)))
4379 ;; (catch 'done
4380 ;; (while (let ((last-pos (point))
4381 ;; (stop-pos (progn
4382 ;; (skip-chars-backward skip-chars)
4383 ;; (point))))
4384 ;;
4385 ;; ;; Skip back over the same region as
4386 ;; ;; `skip-chars-backward' above, but keep to
4387 ;; ;; syntactically relevant positions.
4388 ;; (goto-char last-pos)
4389 ;; (while (and
4390 ;; ;; `re-search-backward' with a single char regexp
4391 ;; ;; should be fast.
4392 ;; (re-search-backward
4393 ;; (if paren-level c-ssb-stop-paren-re c-ssb-stop-re)
4394 ;; stop-pos 'move)
4395 ;;
4396 ;; (progn
4397 ;; (cond
4398 ;; ((looking-at "\\s(")
4399 ;; ;; `paren-level' is set and we've found the
4400 ;; ;; start of the containing paren.
4401 ;; (forward-char)
4402 ;; (throw 'done t))
4403 ;;
4404 ;; ((looking-at c-ssb-sexp-end-re)
4405 ;; ;; We're at the end of a string literal or paren
4406 ;; ;; sexp (if `paren-level' is set).
4407 ;; (forward-char)
4408 ;; (condition-case nil
4409 ;; (c-backward-sexp)
4410 ;; (error
4411 ;; (goto-char limit)
4412 ;; (throw 'done t))))
4413 ;;
4414 ;; (t
4415 ;; (forward-char)
4416 ;; ;; At the end of some syntactic ws or possibly
4417 ;; ;; after a plain '/' operator.
4418 ;; (let ((pos (point)))
4419 ;; (c-backward-syntactic-ws)
4420 ;; (if (= pos (point))
4421 ;; ;; Was a plain '/' operator. Go past it.
4422 ;; (backward-char)))))
4423 ;;
4424 ;; (> (point) stop-pos))))
4425 ;;
4426 ;; ;; Now the point is either at `stop-pos' or at some
4427 ;; ;; position further back if `stop-pos' was at a
4428 ;; ;; syntactically irrelevant place.
4429 ;;
4430 ;; ;; Skip additional syntactic ws so that we don't stop
4431 ;; ;; at the end of a comment if `skip-chars' is
4432 ;; ;; something like "^/".
4433 ;; (c-backward-syntactic-ws)
4434 ;;
4435 ;; (< (point) stop-pos))))
4436 ;;
4437 ;; ;; We might want to extend this with more useful return values
4438 ;; ;; in the future.
4439 ;; (/= (point) start))))
4441 \f
4442 ;; Tools for handling comments and string literals.
4445 "Return the type of literal point is in, if any.
4446 The return value is `c' if in a C-style comment, `c++' if in a C++
4447 style comment, `string' if in a string literal, `pound' if DETECT-CPP
4448 is non-nil and in a preprocessor line, or nil if somewhere else.
4449 Optional LIM is used as the backward limit of the search. If omitted,
4450 or nil, `c-beginning-of-defun' is used.
4452 The last point calculated is cached if the cache is enabled, i.e. if
4453 `c-in-literal-cache' is bound to a two element vector.
4455 Note that this function might do hidden buffer changes. See the
4456 comment at the start of cc-engine.el for more info."
4467 "Return a cons of the beginning and end positions of the comment or
4468 string surrounding point (including both delimiters), or nil if point
4470 to start parsing from. If NEAR is non-nil, then the limits of any
4472 spaces and tabs between point and the literal. The search for such a
4473 literal is done first in forward direction. If NOT-IN-DELIMITER is
4474 non-nil, the case when point is inside a starting delimiter won't be
4475 recognized. This only has effect for comments which have starting
4476 delimiters with more than one character.
4478 Note that this function might do hidden buffer changes. See the
4479 comment at the start of cc-engine.el for more info."
4495 ;; Search forward for a literal.
4506 ;; Search backward.
4517 ;; Block comment. Due to the nature of line
4518 ;; comments, they will always be covered by the
4519 ;; normal case above.
4522 ;; If LIM is bogus, beg will be bogus.
4526 ))))
4528 ;; In case external callers use this; it did have a docstring.
4532 "If the argument is a cons of two buffer positions (such as returned by
4533 `c-literal-limits'), and that range contains a C++ style line comment,
4534 then an extended range is returned that contains all adjacent line
4535 comments (i.e. all comments that starts in the same column with no
4536 empty lines or non-whitespace characters between them). Otherwise the
4537 argument is returned.
4539 Note that this function might do hidden buffer changes. See the
4540 comment at the start of cc-engine.el for more info."
4551 ;; Got to take care in the backward direction to handle
4552 ;; comments which are preceded by code.
4566 range)
4570 "Convenience function that given the result of `c-literal-limits',
4571 returns nil or the type of literal that the range surrounds, one
4572 of the symbols 'c, 'c++ or 'string. It's much faster than using
4573 `c-in-literal' and is intended to be used when you need both the
4574 type of a literal and its limits.
4576 Note that this function might do hidden buffer changes. See the
4577 comment at the start of cc-engine.el for more info."
4588 range))
4591 ;; Get a "safe place" approximately TRY-SIZE characters before START.
4592 ;; This doesn't preserve point.
4601 ;; Return a buffer position HOW-FAR-BACK non-literal characters from START
4602 ;; (default point). This is done by going back further in the buffer then
4603 ;; searching forward for literals. The position found won't be in a
4604 ;; literal. We start searching for the sought position TRY-SIZE (default
4605 ;; twice HOW-FAR-BACK) bytes back from START. This function must be fast.
4606 ;; :-)
4614 stack elt size
4617 ;; Move forward one literal each time round this loop.
4618 ;; Move forward to the start of a comment or string.
4620 pos
4621 start
4622 nil ; target-depth
4623 nil ; stop-before
4624 s ; state
4627 ;; Gather details of the non-literal-bit - starting pos and size.
4631 pos))
4635 ;; Move forward to the end of the comment/string.
4639 start
4640 nil ; target-depth
4641 nil ; stop-before
4642 s ; state
4646 ;; Now try and find enough non-literal characters recorded on the stack.
4647 ;; Go back one recorded literal each time round this loop.
4649 stack)
4654 ;; Have we found enough yet?
4664 ;; Return a buffer position about HOW-FAR non-literal characters forward
4665 ;; from START-POS (default point), which must not be inside a literal.
4672 ;; Scan over counted characters.
4674 pos
4676 nil ; target-depth
4677 nil ; stop-before
4678 s ; state
4682 ;; Scan over literal characters.
4685 pos
4687 nil ; target-depth
4688 nil ; stop-before
4689 s ; state
4694 \f
4695 ;; `c-find-decl-spots' and accompanying stuff.
4697 ;; Variables used in `c-find-decl-spots' to cache the search done for
4698 ;; the first declaration in the last call. When that function starts,
4699 ;; it needs to back up over syntactic whitespace to look at the last
4700 ;; token before the region being searched. That can sometimes cause
4701 ;; moves back and forth over a quite large region of comments and
4702 ;; macros, which would be repeated for each changed character when
4703 ;; we're called during fontification, since font-lock refontifies the
4704 ;; current line for each change. Thus it's worthwhile to cache the
4705 ;; first match.
4706 ;;
4707 ;; `c-find-decl-syntactic-pos' is a syntactically relevant position in
4708 ;; the syntactic whitespace less or equal to some start position.
4709 ;; There's no cached value if it's nil.
4710 ;;
4711 ;; `c-find-decl-match-pos' is the match position if
4712 ;; `c-find-decl-prefix-search' matched before the syntactic whitespace
4713 ;; at `c-find-decl-syntactic-pos', or nil if there's no such match.
4724 ; (defface c-debug-decl-spot-face
4725 ; '((t (:background "Turquoise")))
4726 ; "Debug face to mark the spots where `c-find-decl-spots' stopped.")
4727 ; (defface c-debug-decl-sws-face
4728 ; '((t (:background "Khaki")))
4729 ; "Debug face to mark the syntactic whitespace between the declaration
4730 ; spots and the preceding token end.")
4746 ;; Macro used inside `c-find-decl-spots'. It ought to be a defun,
4747 ;; but it contains lots of free variables that refer to things
4748 ;; inside `c-find-decl-spots'. The point is left at `cfd-match-pos'
4749 ;; if there is a match, otherwise at `cfd-limit'.
4750 ;;
4751 ;; The macro moves point forward to the next putative start of a declaration
4752 ;; or cfd-limit. This decl start is the next token after a "declaration
4753 ;; prefix". The declaration prefix is the earlier of `cfd-prop-match' and
4754 ;; `cfd-re-match'. `cfd-match-pos' is set to the decl prefix.
4755 ;;
4756 ;; This macro might do hidden buffer changes.
4759 ;; Find the next property match position if we haven't got one already.
4770 ;; Find the next `c-decl-prefix-or-start-re' match if we haven't
4771 ;; got one already.
4777 ;; Each time round, the next `while' moves forward over a pseudo match
4778 ;; of `c-decl-prefix-or-start-re' which is either inside a literal, or
4779 ;; is a ":" not preceded by "public", etc.. `cfd-re-match' and
4780 ;; `cfd-re-match-end' get set.
4787 ;; No match. Finish up and exit the loop.
4789 nil)
4792 ;; Matched the end of a token preceding a decl spot.
4796 ;; Matched a token that start a decl spot.
4799 c-literal-faces)
4800 ;; Pseudo match inside a comment or string literal. Skip out
4801 ;; of comments and string literals.
4813 ;; Found a ":" which isn't part of "public:", etc.
4814 t)
4817 ;; If our match was at the decl start, we have to back up over the
4818 ;; preceding syntactic ws to set `cfd-match-pos' and to catch
4819 ;; any decl spots in the syntactic ws.
4824 ;; Choose whichever match is closer to the start.
4827 cfd-re-match nil)
4829 cfd-prop-match nil))
4834 ;; Skip forward past comments only so we don't skip macros.
4836 ;; Set the position to continue at. We can avoid going over
4837 ;; the comments skipped above a second time, but it's possible
4838 ;; that the comment skipping has taken us past `cfd-prop-match'
4839 ;; since the property might be used inside comments.
4845 ;; Call CFD-FUN for each possible spot for a declaration, cast or
4846 ;; label from the point to CFD-LIMIT.
4847 ;;
4848 ;; CFD-FUN is called with point at the start of the spot. It's passed two
4849 ;; arguments: The first is the end position of the token preceding the spot,
4850 ;; or 0 for the implicit match at bob. The second is a flag that is t when
4851 ;; the match is inside a macro. Point should be moved forward by at least
4852 ;; one token.
4853 ;;
4854 ;; If CFD-FUN adds `c-decl-end' properties somewhere below the current spot,
4855 ;; it should return non-nil to ensure that the next search will find them.
4856 ;;
4857 ;; Such a spot is:
4858 ;; o The first token after bob.
4859 ;; o The first token after the end of submatch 1 in
4860 ;; `c-decl-prefix-or-start-re' when that submatch matches.
4861 ;; o The start of each `c-decl-prefix-or-start-re' match when
4862 ;; submatch 1 doesn't match.
4863 ;; o The first token after the end of each occurrence of the
4864 ;; `c-type' text property with the value `c-decl-end', provided
4865 ;; `c-type-decl-end-used' is set.
4866 ;;
4867 ;; Only a spot that match CFD-DECL-RE and whose face is in the
4868 ;; CFD-FACE-CHECKLIST list causes CFD-FUN to be called. The face
4869 ;; check is disabled if CFD-FACE-CHECKLIST is nil.
4870 ;;
4871 ;; If the match is inside a macro then the buffer is narrowed to the
4872 ;; end of it, so that CFD-FUN can investigate the following tokens
4873 ;; without matching something that begins inside a macro and ends
4874 ;; outside it. It's to avoid this work that the CFD-DECL-RE and
4875 ;; CFD-FACE-CHECKLIST checks exist.
4876 ;;
4877 ;; The spots are visited approximately in order from top to bottom.
4878 ;; It's however the positions where `c-decl-prefix-or-start-re'
4879 ;; matches and where `c-decl-end' properties are found that are in
4880 ;; order. Since the spots often are at the following token, they
4881 ;; might be visited out of order insofar as more spots are reported
4882 ;; later on within the syntactic whitespace between the match
4883 ;; positions and their spots.
4884 ;;
4885 ;; It's assumed that comments and strings are fontified in the
4886 ;; searched range.
4887 ;;
4888 ;; This is mainly used in fontification, and so has an elaborate
4889 ;; cache to handle repeated calls from the same start position; see
4890 ;; the variables above.
4891 ;;
4892 ;; All variables in this function begin with `cfd-' to avoid name
4893 ;; collision with the (dynamically bound) variables used in CFD-FUN.
4894 ;;
4895 ;; This function might do hidden buffer changes.
4899 ;; The end of the token preceding the decl spot last found
4900 ;; with `c-decl-prefix-or-start-re'. `cfd-limit' if there's
4901 ;; no match.
4902 cfd-re-match
4903 ;; The end position of the last `c-decl-prefix-or-start-re'
4904 ;; match. If this is greater than `cfd-continue-pos', the
4905 ;; next regexp search is started here instead.
4907 ;; The end of the last `c-decl-end' found by
4908 ;; `c-find-decl-prefix-search'. `cfd-limit' if there's no
4909 ;; match. If searching for the property isn't needed then we
4910 ;; disable it by setting it to `cfd-limit' directly.
4912 ;; The end of the token preceding the decl spot last found by
4913 ;; `c-find-decl-prefix-search'. 0 for the implicit match at
4914 ;; bob. `cfd-limit' if there's no match. In other words,
4915 ;; this is the minimum of `cfd-re-match' and `cfd-prop-match'.
4917 ;; The position to continue searching at.
4918 cfd-continue-pos
4919 ;; The position of the last "real" token we've stopped at.
4920 ;; This can be greater than `cfd-continue-pos' when we get
4921 ;; hits inside macros or at `c-decl-end' positions inside
4922 ;; comments.
4924 ;; The end position of the last entered macro.
4927 ;; Initialize by finding a syntactically relevant start position
4928 ;; before the point, and do the first `c-decl-prefix-or-start-re'
4929 ;; search unless we're at bob.
4932 ;; Must back up a bit since we look for the end of the previous
4933 ;; statement or declaration, which is earlier than the first
4934 ;; returned match.
4937 ;; First we need to move to a syntactically relevant position.
4938 ;; Begin by backing out of comment or string literals.
4941 ;; Try to use the faces to back up to the start of the
4942 ;; literal. FIXME: What if the point is on a declaration
4943 ;; inside a comment?
4949 ;; XEmacs doesn't fontify the quotes surrounding string
4950 ;; literals.
4959 ;; Don't trust the literal to contain only literal faces
4960 ;; (the font lock package might not have fontified the
4961 ;; start of it at all, for instance) so check that we have
4962 ;; arrived at something that looks like a start or else
4963 ;; resort to `c-literal-limits'.
4970 ;; The start is in a literal. If the limit is in the same
4971 ;; one we don't have to find a syntactic position etc. We
4972 ;; only check that if the limit is at or before bonl to save
4973 ;; time; it covers the by far most common case when font-lock
4974 ;; refontifies the current line only.
4985 ;; Completely inside a literal. Set up variables to trig the
4986 ;; (< cfd-continue-pos cfd-start-pos) case below and it'll
4987 ;; find a suitable start position.
4990 ;; Check if the region might be completely inside a macro, to
4991 ;; optimize that like the completely-inside-literal above.
4998 ;; (Maybe) completely inside a macro. Only need to trig the
4999 ;; (< cfd-continue-pos cfd-start-pos) case below to make it
5000 ;; set things up.
5002 start-in-macro t))
5005 ;; Back out of any macro so we don't miss any declaration
5006 ;; that could follow after it.
5010 ;; Now we're at a proper syntactically relevant position so we
5011 ;; can use the cache. But first clear it if it applied
5012 ;; further down.
5017 ;; Don't have to do this if the cache is relevant here,
5018 ;; typically if the same line is refontified again. If
5019 ;; we're just some syntactic whitespace further down we can
5020 ;; still use the cache to limit the skipping.
5023 ;; If we hit `c-find-decl-syntactic-pos' and
5024 ;; `c-find-decl-match-pos' is set then we install the cached
5025 ;; values. If we hit `c-find-decl-syntactic-pos' and
5026 ;; `c-find-decl-match-pos' is nil then we know there's no decl
5027 ;; prefix in the whitespace before `c-find-decl-syntactic-pos'
5028 ;; and so we can continue the search from this point. If we
5029 ;; didn't hit `c-find-decl-syntactic-pos' then we're now in
5030 ;; the right spot to begin searching anyway.
5032 c-find-decl-match-pos)
5034 cfd-continue-pos syntactic-pos)
5039 ;; Always consider bob a match to get the first
5040 ;; declaration in the file. Do this separately instead of
5041 ;; letting `c-decl-prefix-or-start-re' match bob, so that
5042 ;; regexp always can consume at least one character to
5043 ;; ensure that we won't get stuck in an infinite loop.
5048 ;; Do an initial search now. In the bob case above it's
5049 ;; only done to search for a `c-decl-end' spot.
5053 cfd-match-pos)))))
5055 ;; Advance `cfd-continue-pos' if it's before the start position.
5056 ;; The closest continue position that might have effect at or
5057 ;; after the start depends on what we started in. This also
5058 ;; finds a suitable start position in the special cases when the
5059 ;; region is completely within a literal or macro.
5064 ;; If we're in a macro then it's the closest preceding token
5065 ;; in the macro. Check this before `start-in-literal',
5066 ;; since if we're inside a literal in a macro, the preceding
5067 ;; token is earlier than any `c-decl-end' spot inside the
5068 ;; literal (comment).
5070 ;; The only syntactic ws in macros are comments.
5076 ;; If we're in a comment it can only be the closest
5077 ;; preceding `c-decl-end' position within that comment, if
5078 ;; any. Go back to the beginning of such a property so that
5079 ;; `c-find-decl-prefix-search' will find the end of it.
5080 ;; (Can't stop at the end and install it directly on
5081 ;; `cfd-prop-match' since that variable might be cleared
5082 ;; after `cfd-fun' below.)
5083 ;;
5084 ;; Note that if the literal is a string then the property
5085 ;; search will simply skip to the beginning of it right
5086 ;; away.
5098 ;; Didn't find any property inside the comment, so we can
5099 ;; skip it entirely. (This won't skip past a string, but
5100 ;; that'll be handled quickly by the next
5101 ;; `c-find-decl-prefix-search' anyway.)
5107 ;; If we started in normal code, the only match that might
5108 ;; apply before the start is what we already got in
5109 ;; `cfd-match-pos' so we can continue at the start position.
5110 ;; (Note that we don't get here if the first match is below
5111 ;; it.)
5114 ;; Delete found matches if they are before our new continue
5115 ;; position, so that `c-find-decl-prefix-search' won't back up
5116 ;; to them later on.
5124 ;; This is the normal case and we got a proper syntactic
5125 ;; position. If there's a match then it's always outside
5126 ;; macros and comments, so advance to the next token and set
5127 ;; `cfd-token-pos'. The loop below will later go back using
5128 ;; `cfd-continue-pos' to fix declarations inside the
5129 ;; syntactic ws.
5137 ;; Have one of the special cases when the region is completely
5138 ;; within a literal or macro. `cfd-continue-pos' is set to a
5139 ;; good start position for the search, so do it.
5142 ;; Now loop. Round what? (ACM, 2006/7/5). We already got the first match.
5149 ;; Kludge to filter out matches on the "<" that
5150 ;; aren't open parens, for the sake of languages
5151 ;; that got `c-recognize-<>-arglists' set.
5156 ;; If `cfd-continue-pos' is less or equal to
5157 ;; `cfd-token-pos', we've got a hit inside a macro
5158 ;; that's in the syntactic whitespace before the last
5159 ;; "real" declaration we've checked. If they're equal
5160 ;; we've arrived at the declaration a second time, so
5161 ;; there's nothing to do.
5165 ;; If `cfd-continue-pos' is less than `cfd-token-pos'
5166 ;; we're still searching for declarations embedded in
5167 ;; the syntactic whitespace. In that case we need
5168 ;; only to skip comments and not macros, since they
5169 ;; can't be nested, and that's already been done in
5170 ;; `c-find-decl-prefix-search'.
5175 ;; Continue if the following token fails the
5176 ;; CFD-DECL-RE and CFD-FACE-CHECKLIST checks.
5183 t)))
5194 ;; Narrow to the end of the macro if we got a hit inside
5195 ;; one, to avoid recognizing things that start inside the
5196 ;; macro and end outside it.
5198 ;; Not in the same macro as in the previous round.
5209 t
5212 t)
5213 ;; The matched token was the last thing in the macro,
5214 ;; so the whole match is bogus.
5216 nil))))
5223 ;; Restore limits if we did macro narrowing above.
5230 ; cfd-match-pos, etc.
5232 \f
5233 ;; A cache for found types.
5235 ;; Buffer local variable that contains an obarray with the types we've
5236 ;; found. If a declaration is recognized somewhere we record the
5237 ;; fully qualified identifier in it to recognize it as a type
5238 ;; elsewhere in the file too. This is not accurate since we do not
5239 ;; bother with the scoping rules of the languages, but in practice the
5240 ;; same name is seldom used as both a type and something else in a
5241 ;; file, and we only use this as a last resort in ambiguous cases (see
5242 ;; `c-forward-decl-or-cast-1').
5243 ;;
5244 ;; Not every type need be in this cache. However, things which have
5245 ;; ceased to be types must be removed from it.
5246 ;;
5247 ;; Template types in C++ are added here too but with the template
5248 ;; arglist replaced with "<>" in references or "<" for the one in the
5249 ;; primary type. E.g. the type "Foo<A,B>::Bar<C>" is stored as
5250 ;; "Foo<>::Bar<". This avoids storing very long strings (since C++
5251 ;; template specs can be fairly sized programs in themselves) and
5252 ;; improves the hit ratio (it's a type regardless of the template
5253 ;; args; it's just not the same type, but we're only interested in
5254 ;; recognizing types, not telling distinct types apart). Note that
5255 ;; template types in references are added here too; from the example
5256 ;; above there will also be an entry "Foo<".
5261 ;; Clears `c-found-types'.
5265 ;; Add the given region as a type in `c-found-types'. If the region
5266 ;; doesn't match an existing type but there is a type which is equal
5267 ;; to the given one except that the last character is missing, then
5268 ;; the shorter type is removed. That's done to avoid adding all
5269 ;; prefixes of a type as it's being entered and font locked. This
5270 ;; doesn't cover cases like when characters are removed from a type
5271 ;; or added in the middle. We'd need the position of point when the
5272 ;; font locking is invoked to solve this well.
5273 ;;
5274 ;; This function might do hidden buffer changes.
5281 ;; Remove the "NAME" from c-found-types, if present.
5285 ;; Return non-nil if the given region contains a type in
5286 ;; `c-found-types'.
5287 ;;
5288 ;; This function might do hidden buffer changes.
5290 c-found-types))
5293 ;; Return all the types in `c-found-types' as a sorted list of
5294 ;; strings.
5298 type-list)))
5299 c-found-types)
5302 ;; Shut up the byte compiler.
5306 ;; An after change function which, in conjunction with the info in
5307 ;; c-maybe-stale-found-type (set in c-before-change), removes a type
5308 ;; from `c-found-types', should this type have become stale. For
5309 ;; example, this happens to "foo" when "foo \n bar();" becomes
5310 ;; "foo(); \n bar();". Such stale types, if not removed, foul up
5311 ;; the fontification.
5312 ;;
5313 ;; Have we, perhaps, added non-ws characters to the front/back of a found
5314 ;; type?
5334 ;; Changing the amount of (already existing) whitespace - don't do anything.
5339 ;; The syntactic relationship which defined a "found type" has been
5340 ;; destroyed.
5343 ;; ((eq (car c-maybe-stale-found-type) 'c-decl-type-start) FIXME!!!
5344 )))
5346 \f
5347 ;; Setting and removing syntax properties on < and > in languages (C++
5348 ;; and Java) where they can be template/generic delimiters as well as
5349 ;; their normal meaning of "less/greater than".
5351 ;; Normally, < and > have syntax 'punctuation'. When they are found to
5352 ;; be delimiters, they are marked as such with the category properties
5353 ;; c-<-as-paren-syntax, c->-as-paren-syntax respectively.
5355 ;; STRATEGY:
5356 ;;
5357 ;; It is impossible to determine with certainty whether a <..> pair in
5358 ;; C++ is two comparison operators or is template delimiters, unless
5359 ;; one duplicates a lot of a C++ compiler. For example, the following
5360 ;; code fragment:
5361 ;;
5362 ;; foo (a < b, c > d) ;
5363 ;;
5364 ;; could be a function call with two integer parameters (each a
5365 ;; relational expression), or it could be a constructor for class foo
5366 ;; taking one parameter d of templated type "a < b, c >". They are
5367 ;; somewhat easier to distinguish in Java.
5368 ;;
5369 ;; The strategy now (2010-01) adopted is to mark and unmark < and
5370 ;; > IN MATCHING PAIRS ONLY. [Previously, they were marked
5371 ;; individually when their context so indicated. This gave rise to
5372 ;; intractable problems when one of a matching pair was deleted, or
5373 ;; pulled into a literal.]
5374 ;;
5375 ;; At each buffer change, the syntax-table properties are removed in a
5376 ;; before-change function and reapplied, when needed, in an
5377 ;; after-change function. It is far more important that the
5378 ;; properties get removed when they they are spurious than that they
5379 ;; be present when wanted.
5380 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
5382 ;; POS (default point) is at a < character. If it is marked with
5383 ;; open paren syntax-table text property, remove the property,
5384 ;; together with the close paren property on the matching > (if
5385 ;; any).
5400 ;; POS (default point) is at a > character. If it is marked with
5401 ;; close paren syntax-table property, remove the property, together
5402 ;; with the open paren property on the matching < (if any).
5417 ;; POS (default point) is at a < or > character. If it has an
5418 ;; open/close paren syntax-table property, remove this property both
5419 ;; from the current character and its partner (which will also be
5420 ;; thusly marked).
5430 ;; POS (default point) is at a < character. If it is both marked
5431 ;; with open/close paren syntax-table property, and has a matching >
5432 ;; (also marked) which is after LIM, remove the property both from
5433 ;; the current > and its partner. Return t when this happens, nil
5434 ;; when it doesn't.
5448 t)))
5451 ;; POS (default point) is at a > character. If it is both marked
5452 ;; with open/close paren syntax-table property, and has a matching <
5453 ;; (also marked) which is before LIM, remove the property both from
5454 ;; the current < and its partner. Return t when this happens, nil
5455 ;; when it doesn't.
5469 t)))
5471 ;; Set by c-common-init in cc-mode.el.
5476 ;; Unmark certain pairs of "< .... >" which are currently marked as
5477 ;; template/generic delimiters. (This marking is via syntax-table
5478 ;; text properties).
5479 ;;
5480 ;; These pairs are those which are in the current "statement" (i.e.,
5481 ;; the region between the {, }, or ; before BEG and the one after
5482 ;; END), and which enclose any part of the interval (BEG END).
5483 ;;
5484 ;; Note that in C++ (?and Java), template/generic parens cannot
5485 ;; enclose a brace or semicolon, so we use these as bounds on the
5486 ;; region we must work on.
5487 ;;
5488 ;; This function is called from before-change-functions (via
5489 ;; c-get-state-before-change-functions). Thus the buffer is widened,
5490 ;; and point is undefined, both at entry and exit.
5491 ;;
5492 ;; FIXME!!! This routine ignores the possibility of macros entirely.
5493 ;; 2010-01-29.
5497 new-beg new-end need-new-beg need-new-end)
5498 ;; Locate the barrier before the changed region
5503 ;; Remove the syntax-table properties from each pertinent <...> pair.
5504 ;; Firsly, the ones with the < before beg and > after beg.
5509 ;; Locate the barrier after END.
5514 ;; Remove syntax-table properties from the remaining pertinent <...>
5515 ;; pairs, those with a > after end and < before end.
5520 ;; Extend the fontification region, if needed.
5532 ;; This is called from `after-change-functions' when
5533 ;; c-recognize-<>-arglists' is set. It ensures that no "<" or ">"
5534 ;; chars with paren syntax become part of another operator like "<<"
5535 ;; or ">=".
5536 ;;
5537 ;; This function might do hidden buffer changes.
5570 \f
5571 ;; Handling of small scale constructs like types and names.
5573 ;; Dynamically bound variable that instructs `c-forward-type' to also
5574 ;; treat possible types (i.e. those that it normally returns 'maybe or
5575 ;; 'found for) as actual types (and always return 'found for them).
5576 ;; This means that it records them in `c-record-type-identifiers' if
5577 ;; that is set, and that it adds them to `c-found-types'.
5580 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5581 ;; mark up successfully parsed arglists with paren syntax properties on
5582 ;; the surrounding angle brackets and with `c-<>-arg-sep' in the
5583 ;; `c-type' property of each argument separating comma.
5584 ;;
5585 ;; Setting this variable also makes `c-forward-<>-arglist' recurse into
5586 ;; all arglists for side effects (i.e. recording types), otherwise it
5587 ;; exploits any existing paren syntax properties to quickly jump to the
5588 ;; end of already parsed arglists.
5589 ;;
5590 ;; Marking up the arglists is not the default since doing that correctly
5591 ;; depends on a proper value for `c-restricted-<>-arglists'.
5594 ;; Dynamically bound variable that instructs `c-forward-<>-arglist' to
5595 ;; not accept arglists that contain binary operators.
5596 ;;
5597 ;; This is primarily used to handle C++ template arglists. C++
5598 ;; disambiguates them by checking whether the preceding name is a
5599 ;; template or not. We can't do that, so we assume it is a template
5600 ;; if it can be parsed as one. That usually works well since
5601 ;; comparison expressions on the forms "a < b > c" or "a < b, c > d"
5602 ;; in almost all cases would be pointless.
5603 ;;
5604 ;; However, in function arglists, e.g. in "foo (a < b, c > d)", we
5605 ;; should let the comma separate the function arguments instead. And
5606 ;; in a context where the value of the expression is taken, e.g. in
5607 ;; "if (a < b || c > d)", it's probably not a template.
5610 ;; Dynamically bound variables that instructs
5611 ;; `c-forward-keyword-clause', `c-forward-<>-arglist',
5612 ;; `c-forward-name', `c-forward-type', `c-forward-decl-or-cast-1', and
5613 ;; `c-forward-label' to record the ranges of all the type and
5614 ;; reference identifiers they encounter. They will build lists on
5615 ;; these variables where each element is a cons of the buffer
5616 ;; positions surrounding each identifier. This recording is only
5617 ;; activated when `c-record-type-identifiers' is non-nil.
5618 ;;
5619 ;; All known types that can't be identifiers are recorded, and also
5620 ;; other possible types if `c-promote-possible-types' is set.
5621 ;; Recording is however disabled inside angle bracket arglists that
5622 ;; are encountered inside names and other angle bracket arglists.
5623 ;; Such occurrences are taken care of by `c-font-lock-<>-arglists'
5624 ;; instead.
5625 ;;
5626 ;; Only the names in C++ template style references (e.g. "tmpl" in
5627 ;; "tmpl<a,b>::foo") are recorded as references, other references
5628 ;; aren't handled here.
5629 ;;
5630 ;; `c-forward-label' records the label identifier(s) on
5631 ;; `c-record-ref-identifiers'.
5635 ;; This variable will receive a cons cell of the range of the last
5636 ;; single identifier symbol stepped over by `c-forward-name' if it's
5637 ;; successful. This is the range that should be put on one of the
5638 ;; record lists above by the caller. It's assigned nil if there's no
5639 ;; such symbol in the name.
5644 ;; Always true.
5654 ;; Always true.
5662 ;; Dynamically bound variable that instructs `c-forward-type' to
5663 ;; record the ranges of types that only are found. Behaves otherwise
5664 ;; like `c-record-type-identifiers'.
5668 ;; Used internally in `c-forward-keyword-clause' to move forward
5669 ;; over a type (if TYPE is 'type) or a name (otherwise) which
5670 ;; possibly is prefixed by keywords and their associated clauses.
5671 ;; Try with a type/name first to not trip up on those that begin
5672 ;; with a keyword. Return t if a known or found type is moved
5673 ;; over. The point is clobbered if nil is returned. If range
5674 ;; recording is enabled, the identifier is recorded on as a type
5675 ;; if TYPE is 'type or as a reference if TYPE is 'ref.
5676 ;;
5677 ;; This macro might do hidden buffer changes.
5682 nil
5689 t)))
5692 ;; Used internally in `c-forward-keyword-clause' to move forward
5693 ;; over a comma separated list of types or names using
5694 ;; `c-forward-keyword-prefixed-id'.
5695 ;;
5696 ;; This macro might do hidden buffer changes.
5707 ;; Submatch MATCH in the current match data is assumed to surround a
5708 ;; token. If it's a keyword, move over it and any immediately
5709 ;; following clauses associated with it, stopping at the start of
5710 ;; the next token. t is returned in that case, otherwise the point
5711 ;; stays and nil is returned. The kind of clauses that are
5712 ;; recognized are those specified by `c-type-list-kwds',
5713 ;; `c-ref-list-kwds', `c-colon-type-list-kwds',
5714 ;; `c-paren-nontype-kwds', `c-paren-type-kwds', `c-<>-type-kwds',
5715 ;; and `c-<>-arglist-kwds'.
5716 ;;
5717 ;; This function records identifier ranges on
5718 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5719 ;; `c-record-type-identifiers' is non-nil.
5720 ;;
5721 ;; Note that for `c-colon-type-list-kwds', which doesn't necessary
5722 ;; apply directly after the keyword, the type list is moved over
5723 ;; only when there is no unaccounted token before it (i.e. a token
5724 ;; that isn't moved over due to some other keyword list). The
5725 ;; identifier ranges in the list are still recorded if that should
5726 ;; be done, though.
5727 ;;
5728 ;; This function might do hidden buffer changes.
5731 ;; The call to `c-forward-<>-arglist' below is made after
5732 ;; `c-<>-sexp-kwds' keywords, so we're certain they actually
5733 ;; are angle bracket arglists and `c-restricted-<>-arglists'
5734 ;; should therefore be nil.
5746 ;; There's a type directly after a keyword in `c-type-list-kwds'.
5751 ;; There's a name directly after a keyword in `c-ref-list-kwds'.
5756 ;; There's an open paren after a keyword in `c-paren-any-kwds'.
5763 ;; Use `c-forward-type' on every identifier we can find
5764 ;; inside the paren, to record the types.
5789 ;; If we are at the colon already, we move over the type
5790 ;; list after it.
5796 ;; Not at the colon, so stop here. But the identifier
5797 ;; ranges in the type list later on should still be
5798 ;; recorded.
5801 ;; If a keyword matched both one of the types above and
5802 ;; this one, we match `c-colon-type-list-re' after the
5803 ;; clause matched above.
5810 ;; There's a type after the `c-colon-type-list-re' match
5811 ;; after a keyword in `c-colon-type-list-kwds'.
5815 t)))
5817 ;; cc-mode requires cc-fonts.
5821 ;; The point is assumed to be at a "<". Try to treat it as the open
5822 ;; paren of an angle bracket arglist and move forward to the
5823 ;; corresponding ">". If successful, the point is left after the
5824 ;; ">" and t is returned, otherwise the point isn't moved and nil is
5825 ;; returned. If ALL-TYPES is t then all encountered arguments in
5826 ;; the arglist that might be types are treated as found types.
5827 ;;
5828 ;; The variable `c-parse-and-markup-<>-arglists' controls how this
5829 ;; function handles text properties on the angle brackets and argument
5830 ;; separating commas.
5831 ;;
5832 ;; `c-restricted-<>-arglists' controls how lenient the template
5833 ;; arglist recognition should be.
5834 ;;
5835 ;; This function records identifier ranges on
5836 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
5837 ;; `c-record-type-identifiers' is non-nil.
5838 ;;
5839 ;; This function might do hidden buffer changes.
5842 ;; If `c-record-type-identifiers' is set then activate
5843 ;; recording of any found types that constitute an argument in
5844 ;; the arglist.
5852 ;; `nconc' doesn't mind that the tail of
5853 ;; `c-record-found-types' is t.
5856 t)
5859 nil)))
5862 ;; Recursive part of `c-forward-<>-arglist'.
5863 ;;
5864 ;; This function might do hidden buffer changes.
5867 ;; Cover this so that any recorded found type ranges are
5868 ;; automatically lost if it turns out to not be an angle
5869 ;; bracket arglist. It's propagated through the return value
5870 ;; on successful completion.
5872 ;; List that collects the positions after the argument
5873 ;; separating ',' in the arglist.
5874 arg-start-pos)
5875 ;; If the '<' has paren open syntax then we've marked it as an angle
5876 ;; bracket arglist before, so skip to the end.
5884 t
5885 ;; Got unmatched paren angle brackets. We don't clear the paren
5886 ;; syntax properties and retry, on the basis that it's very
5887 ;; unlikely that paren angle brackets become operators by code
5888 ;; manipulation. It's far more likely that it doesn't match due
5889 ;; to narrowing or some temporary change.
5891 nil))
5896 ;; go forward one non-alphanumeric character (group) per iteration of
5897 ;; this loop.
5904 ;; All encountered identifiers are types, so set the
5905 ;; promote flag and parse the type.
5928 ;; Note: These regexps exploit the match order in \| so
5929 ;; that "<>" is matched by "<" rather than "[^>:-]>".
5931 ;; Stop on ',', '|', '&', '+' and '-' to catch
5932 ;; common binary operators that could be between
5933 ;; two comparison expressions "a<b" and "c>d".
5935 nil t t))
5939 ;; Either an operator starting with '>' or the end of
5940 ;; the angle bracket arglist.
5947 ;; The angle bracket arglist is finished.
5959 ;; Either an operator starting with '<' or a nested arglist.
5963 ;; The '<' begins a multi-char operator.
5967 ;; We're at a nested <.....>
5973 ;; There's always an identifier before an angle
5974 ;; bracket arglist, or a keyword in `c-<>-type-kwds'
5975 ;; or `c-<>-arglist-kwds'.
5992 ;; It was an angle bracket arglist.
5995 ;; Record the identifier before the template as a type
5996 ;; or reference depending on whether the arglist is last
5997 ;; in a qualified identifier.
6007 ;; At a "less than" operator.
6010 )))
6017 ;; Just another argument. Record the position. The
6018 ;; type check stuff that made us stop at it is at
6019 ;; the top of the loop.
6023 ;; Got a character that can't be in an angle bracket
6024 ;; arglist argument. Abort using `throw', since
6025 ;; it's useless to try to find a surrounding arglist
6026 ;; if we're nested.
6032 ;; The point is assumed to be directly after a ">". Try to treat it
6033 ;; as the close paren of an angle bracket arglist and move back to
6034 ;; the corresponding "<". If successful, the point is left at
6035 ;; the "<" and t is returned, otherwise the point isn't moved and
6036 ;; nil is returned. ALL-TYPES is passed on to
6037 ;; `c-forward-<>-arglist'.
6038 ;;
6039 ;; If the optional LIMIT is given, it bounds the backward search.
6040 ;; It's then assumed to be at a syntactically relevant position.
6041 ;;
6042 ;; This is a wrapper around `c-forward-<>-arglist'. See that
6043 ;; function for more details.
6052 t
6053 ;; See corresponding note in `c-forward-<>-arglist'.
6055 nil)
6062 t
6063 ;; Stopped at bob or a char that isn't allowed in an
6064 ;; arglist, so we've failed.
6066 nil)
6071 ;; If we moved then the "<" was part of some
6072 ;; multicharacter token.
6073 t
6079 ;; Matched the arglist. Break the while.
6081 nil)
6083 ;; We started from a non-paren ">" inside an
6084 ;; arglist.
6086 nil)
6088 ;; Matched a shorter arglist. Can be a nested
6089 ;; one so continue looking.
6091 t))
6092 t))))))
6097 ;; Move forward over a complete name if at the beginning of one,
6098 ;; stopping at the next following token. A keyword, as such,
6099 ;; doesn't count as a name. If the point is not at something that
6100 ;; is recognized as a name then it stays put.
6101 ;;
6102 ;; A name could be something as simple as "foo" in C or something as
6103 ;; complex as "X<Y<class A<int>::B, BIT_MAX >> b>, ::operator<> ::
6104 ;; Z<(a>b)> :: operator const X<&foo>::T Q::G<unsigned short
6105 ;; int>::*volatile const" in C++ (this function is actually little
6106 ;; more than a `looking-at' call in all modes except those that,
6107 ;; like C++, have `c-recognize-<>-arglists' set).
6108 ;;
6109 ;; Return
6110 ;; o - nil if no name is found;
6111 ;; o - 'template if it's an identifier ending with an angle bracket
6112 ;; arglist;
6113 ;; o - 'operator of it's an operator identifier;
6114 ;; o - t if it's some other kind of name.
6115 ;;
6116 ;; This function records identifier ranges on
6117 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6118 ;; `c-record-type-identifiers' is non-nil.
6119 ;;
6120 ;; This function might do hidden buffer changes.
6123 ;; Turn off `c-promote-possible-types' here since we might
6124 ;; call `c-forward-<>-arglist' and we don't want it to promote
6125 ;; every suspect thing in the arglist to a type. We're
6126 ;; typically called from `c-forward-type' in this case, and
6127 ;; the caller only wants the top level type that it finds to
6128 ;; be promoted.
6129 c-promote-possible-types)
6135 ;; Check for keyword. We go to the last symbol in
6136 ;; `c-identifier-key' first.
6149 ;; "template" is only valid inside an
6150 ;; identifier if preceded by "::".
6155 t))
6157 ;; Handle a C++ operator or template identifier.
6161 ;; Got "... ::template".
6165 res subres))))
6168 ;; Got a cast operator.
6172 ;; Now we should match a sequence of either
6173 ;; '*', '&' or a name followed by ":: *",
6174 ;; where each can be followed by a sequence
6175 ;; of `c-opt-type-modifier-key'.
6178 t)
6188 t))))
6196 ;; Got some other operator.
6204 nil)
6206 ;; `id-start' is equal to `id-end' if we've jumped over
6207 ;; an identifier that doesn't end with a symbol token.
6208 ;; That can occur e.g. for Java import directives on the
6209 ;; form "foo.bar.*".
6216 res t)))
6226 ;; Got a concatenated identifier. This handles the
6227 ;; cases with tricky syntactic whitespace that aren't
6228 ;; covered in `c-identifier-key'.
6231 t)
6235 ;; Maybe an angle bracket arglist.
6243 c-last-identifier-range nil)
6248 ;; Continue if there's an identifier concatenation
6249 ;; operator after the template argument.
6255 t)
6260 nil)))
6261 )))))
6264 res))
6267 ;; Move forward over a type spec if at the beginning of one,
6268 ;; stopping at the next following token. The keyword "typedef"
6269 ;; isn't part of a type spec here.
6270 ;;
6271 ;; BRACE-BLOCK-TOO, when non-nil, means move over the brace block in
6272 ;; constructs like "struct foo {...} bar ;" or "struct {...} bar;".
6273 ;; The current (2009-03-10) intention is to convert all uses of
6274 ;; `c-forward-type' to call with this parameter set, then to
6275 ;; eliminate it.
6276 ;;
6277 ;; Return
6278 ;; o - t if it's a known type that can't be a name or other
6279 ;; expression;
6280 ;; o - 'known if it's an otherwise known type (according to
6281 ;; `*-font-lock-extra-types');
6282 ;; o - 'prefix if it's a known prefix of a type;
6283 ;; o - 'found if it's a type that matches one in `c-found-types';
6284 ;; o - 'maybe if it's an identifier that might be a type; or
6285 ;; o - nil if it can't be a type (the point isn't moved then).
6286 ;;
6287 ;; The point is assumed to be at the beginning of a token.
6288 ;;
6289 ;; Note that this function doesn't skip past the brace definition
6290 ;; that might be considered part of the type, e.g.
6291 ;; "enum {a, b, c} foo".
6292 ;;
6293 ;; This function records identifier ranges on
6294 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6295 ;; `c-record-type-identifiers' is non-nil.
6296 ;;
6297 ;; This function might do hidden buffer changes.
6305 ;; Skip leading type modifiers. If any are found we know it's a
6306 ;; prefix of a type.
6315 ; "typedef".
6323 ;; In many languages the name can be used without the
6324 ;; prefix, so we add it to `c-found-types'.
6327 c-last-identifier-range)
6349 id-range c-last-identifier-range))))
6361 id-end)
6363 ;; Looking at a primitive or known type identifier. We've
6364 ;; checked for a name first so that we don't go here if the
6365 ;; known type match only is a prefix of another name.
6376 ;; There might be more keywords for the type.
6404 ;; A normal identifier.
6414 ;; It's an identifier that has been used as
6415 ;; a type somewhere else.
6416 'found
6417 ;; It's an identifier that might be a type.
6420 ;; A template is a type.
6424 ;; Otherwise it's an operator identifier, which is not a type.
6429 ;; Skip trailing type modifiers. If any are found we know it's
6430 ;; a type.
6436 ;; Step over any type suffix operator. Do not let the existence
6437 ;; of these alter the classification of the found type, since
6438 ;; these operators typically are allowed in normal expressions
6439 ;; too.
6446 ;; Look for a trailing operator that concatenates the type
6447 ;; with a following one, and if so step past that one through
6448 ;; a recursive call. Note that we don't record concatenated
6449 ;; types in `c-found-types' - it's the component types that
6450 ;; are recorded when appropriate.
6453 c-promote-possible-types))
6454 ;; If we can't promote then set `c-record-found-types' so that
6455 ;; we can merge in the types from the second part afterwards if
6456 ;; it turns out to be a known type there.
6459 subres)
6468 ;; If either operand certainly is a type then both are, but we
6469 ;; don't let the existence of the operator itself promote two
6470 ;; uncertain types to a certain one.
6489 ;; Merge in the ranges of any types found by the second
6490 ;; `c-forward-type'.
6492 ;; `nconc' doesn't mind that the tail of
6493 ;; `c-record-found-types' is t.
6495 c-record-type-identifiers))))
6503 ;;(message "c-forward-type %s -> %s: %s" start (point) res)
6505 res))
6508 ;; Used for Java code only at the moment. Assumes point is on the
6509 ;; @, moves forward an annotation. returns nil if there is no
6510 ;; annotation at point.
6517 t)))
6520 ;; Pull the next open brace from PS (which has the form of paren-state),
6521 ;; skipping over any brace pairs. Returns NIL when PS is exhausted.
6529 ;; Test whether we are in a C++ member initializer list, and if so, go back
6530 ;; to the introducing ":", returning the position of the opening paren of
6531 ;; the function's arglist. Otherwise return nil, leaving point unchanged.
6534 res)
6572 res))
6574 \f
6575 ;; Handling of large scale constructs like statements and declarations.
6577 ;; Macro used inside `c-forward-decl-or-cast-1'. It ought to be a
6578 ;; defsubst or perhaps even a defun, but it contains lots of free
6579 ;; variables that refer to things inside `c-forward-decl-or-cast-1'.
6581 ;; `c-forward-decl-or-cast-1' can consume an arbitrary length list
6582 ;; of types when parsing a declaration, which means that it
6583 ;; sometimes consumes the identifier in the declaration as a type.
6584 ;; This is used to "backtrack" and make the last type be treated as
6585 ;; an identifier instead.
6588 ;; These identifiers are bound only in the inner let.
6590 identifier-start type-start
6591 got-parens nil
6592 got-identifier t
6593 got-suffix t
6594 got-suffix-after-parens id-start
6598 t
6599 backup-at-type))
6601 id-start backup-id-start)
6603 id-start start-pos))
6605 ;; When these flags already are set we've found specifiers that
6606 ;; unconditionally signal these attributes - backtracking doesn't
6607 ;; change that. So keep them set in that case.
6614 ;; This identifier is bound only in the inner let.
6618 ;; Move forward over a declaration or a cast if at the start of one.
6619 ;; The point is assumed to be at the start of some token. Nil is
6620 ;; returned if no declaration or cast is recognized, and the point
6621 ;; is clobbered in that case.
6622 ;;
6623 ;; If a declaration is parsed:
6624 ;;
6625 ;; The point is left at the first token after the first complete
6626 ;; declarator, if there is one. The return value is a cons where
6627 ;; the car is the position of the first token in the declarator. (See
6628 ;; below for the cdr.)
6629 ;; Some examples:
6630 ;;
6631 ;; void foo (int a, char *b) stuff ...
6632 ;; car ^ ^ point
6633 ;; float (*a)[], b;
6634 ;; car ^ ^ point
6635 ;; unsigned int a = c_style_initializer, b;
6636 ;; car ^ ^ point
6637 ;; unsigned int a (cplusplus_style_initializer), b;
6638 ;; car ^ ^ point (might change)
6639 ;; class Foo : public Bar {}
6640 ;; car ^ ^ point
6641 ;; class PikeClass (int a, string b) stuff ...
6642 ;; car ^ ^ point
6643 ;; enum bool;
6644 ;; car ^ ^ point
6645 ;; enum bool flag;
6646 ;; car ^ ^ point
6647 ;; void cplusplus_function (int x) throw (Bad);
6648 ;; car ^ ^ point
6649 ;; Foo::Foo (int b) : Base (b) {}
6650 ;; car ^ ^ point
6651 ;;
6652 ;; The cdr of the return value is non-nil when a
6653 ;; `c-typedef-decl-kwds' specifier is found in the declaration.
6654 ;; Specifically it is a dotted pair (A . B) where B is t when a
6655 ;; `c-typedef-kwds' ("typedef") is present, and A is t when some
6656 ;; other `c-typedef-decl-kwds' (e.g. class, struct, enum)
6657 ;; specifier is present. I.e., (some of) the declared
6658 ;; identifier(s) are types.
6659 ;;
6660 ;; If a cast is parsed:
6661 ;;
6662 ;; The point is left at the first token after the closing paren of
6663 ;; the cast. The return value is `cast'. Note that the start
6664 ;; position must be at the first token inside the cast parenthesis
6665 ;; to recognize it.
6666 ;;
6667 ;; PRECEDING-TOKEN-END is the first position after the preceding
6668 ;; token, i.e. on the other side of the syntactic ws from the point.
6669 ;; Use a value less than or equal to (point-min) if the point is at
6670 ;; the first token in (the visible part of) the buffer.
6671 ;;
6672 ;; CONTEXT is a symbol that describes the context at the point:
6673 ;; 'decl In a comma-separated declaration context (typically
6674 ;; inside a function declaration arglist).
6675 ;; '<> In an angle bracket arglist.
6676 ;; 'arglist Some other type of arglist.
6677 ;; nil Some other context or unknown context. Includes
6678 ;; within the parens of an if, for, ... construct.
6679 ;;
6680 ;; LAST-CAST-END is the first token after the closing paren of a
6681 ;; preceding cast, or nil if none is known. If
6682 ;; `c-forward-decl-or-cast-1' is used in succession, it should be
6683 ;; the position after the closest preceding call where a cast was
6684 ;; matched. In that case it's used to discover chains of casts like
6685 ;; "(a) (b) c".
6686 ;;
6687 ;; This function records identifier ranges on
6688 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
6689 ;; `c-record-type-identifiers' is non-nil.
6690 ;;
6691 ;; This function might do hidden buffer changes.
6694 ;; first type, i.e. after any leading specifiers. It might
6695 ;; also point at the beginning of the preceding syntactic
6696 ;; whitespace.
6698 ;; Set to the result of `c-forward-type'.
6699 at-type
6700 ;; The position of the first token in what we currently
6701 ;; believe is the type in the declaration or cast, after any
6702 ;; specifiers and their associated clauses.
6703 type-start
6704 ;; The position of the first token in what we currently
6705 ;; believe is the declarator for the first identifier. Set
6706 ;; when the type is found, and moved forward over any
6707 ;; `c-decl-hangon-kwds' and their associated clauses that
6708 ;; occurs after the type.
6709 id-start
6710 ;; These store `at-type', `type-start' and `id-start' of the
6711 ;; identifier before the one in those variables. The previous
6712 ;; identifier might turn out to be the real type in a
6713 ;; declaration if the last one has to be the declarator in it.
6714 ;; If `backup-at-type' is nil then the other variables have
6715 ;; undefined values.
6716 backup-at-type backup-type-start backup-id-start
6717 ;; Set if we've found a specifier (apart from "typedef") that makes
6718 ;; the defined identifier(s) types.
6719 at-type-decl
6720 ;; Set if we've a "typedef" keyword.
6721 at-typedef
6722 ;; Set if we've found a specifier that can start a declaration
6723 ;; where there's no type.
6724 maybe-typeless
6725 ;; If a specifier is found that also can be a type prefix,
6726 ;; these flags are set instead of those above. If we need to
6727 ;; back up an identifier, they are copied to the real flag
6728 ;; variables. Thus they only take effect if we fail to
6729 ;; interpret it as a type.
6730 backup-at-type-decl backup-maybe-typeless
6731 ;; Whether we've found a declaration or a cast. We might know
6732 ;; this before we've found the type in it. It's 'ids if we've
6733 ;; found two consecutive identifiers (usually a sure sign, but
6734 ;; we should allow that in labels too), and t if we've found a
6735 ;; specifier keyword (a 100% sure sign).
6736 at-decl-or-cast
6737 ;; Set when we need to back up to parse this as a declaration
6738 ;; but not as a cast.
6739 backup-if-not-cast
6740 ;; For casts, the return position.
6741 cast-end
6742 ;; Save `c-record-type-identifiers' and
6743 ;; `c-record-ref-identifiers' since ranges are recorded
6744 ;; speculatively and should be thrown away if it turns out
6745 ;; that it isn't a declaration or cast.
6752 ;; Check for a type. Unknown symbols are treated as possible
6753 ;; types, but they could also be specifiers disguised through
6754 ;; macros like __INLINE__, so we recognize both types and known
6755 ;; specifiers after them too.
6759 ;; Look for a specifier keyword clause.
6771 ;; Found a known or possible type or a prefix of a known type.
6774 ;; Got two identifiers with nothing but whitespace
6775 ;; between them. That can only happen in declarations.
6779 ;; If the previous identifier is a found type we
6780 ;; record it as a real one; it might be some sort of
6781 ;; alias for a prefix like "unsigned".
6788 backup-type-start type-start
6789 backup-id-start id-start
6790 at-type found-type
6791 type-start start
6793 ;; The previous ambiguous specifier/type turned out
6794 ;; to be a type since we've parsed another one after
6795 ;; it, so clear these backup flags.
6796 backup-at-type-decl nil
6797 backup-maybe-typeless nil))
6801 ;; Handle known specifier keywords and
6802 ;; `c-decl-hangon-kwds' which can occur after known
6803 ;; types.
6806 ;; It's a hang-on keyword that can occur anywhere.
6810 ;; Move the identifier start position if
6811 ;; we've passed a type.
6813 ;; Otherwise treat this as a specifier and
6814 ;; move the fallback position.
6818 ;; It's an ordinary specifier so we know that
6819 ;; anything before this can't be the type.
6821 start-pos kwd-clause-end)
6824 ;; It's ambiguous whether this keyword is a
6825 ;; specifier or a type prefix, so set the backup
6826 ;; flags. (It's assumed that `c-forward-type'
6827 ;; moved further than `c-forward-keyword-clause'.)
6835 ;; This test only happens after we've scanned a type.
6836 ;; So, with valid syntax, kwd-sym can't be 'typedef.
6841 ;; Haven't matched a type so it's an unambiguous
6842 ;; specifier keyword and we know we're in a
6843 ;; declaration.
6848 ;; If the type isn't known we continue so that we'll jump
6849 ;; over all specifiers and type identifiers. The reason
6850 ;; to do this for a known type prefix is to make things
6851 ;; like "unsigned INT16" work.
6856 ;; If a known type was found, we still need to skip over any
6857 ;; hangon keyword clauses after it. Otherwise it has already
6858 ;; been done in the loop above.
6864 ;; A prefix type is itself a primitive type when it's not
6865 ;; followed by another type.
6869 ;; Got no type but set things up to continue anyway to handle
6870 ;; the various cases when a declaration doesn't start with a
6871 ;; type.
6876 ;; If it's C++ then check if the last "type" ends on the form
6877 ;; "foo::foo" or "foo::~foo", i.e. if it's the name of a
6878 ;; (con|de)structor.
6889 ;; Cheating in the handling of syntactic ws below.
6896 name))
6897 ;; It is a (con|de)structor name. In that case the
6898 ;; declaration is typeless so zap out any preceding
6899 ;; identifier(s) that we might have taken as types.
6902 backup-at-type nil
6903 id-start type-start))))))
6905 ;; Check for and step over a type decl expression after the thing
6906 ;; that is or might be a type. This can't be skipped since we
6907 ;; need the correct end position of the declarator for
6908 ;; `max-type-decl-end-*'.
6910 ;; True if there's a non-open-paren match of
6911 ;; `c-type-decl-prefix-key'.
6912 got-prefix
6913 ;; True if the declarator is surrounded by a parenthesis pair.
6914 got-parens
6915 ;; True if there is an identifier in the declarator.
6916 got-identifier
6917 ;; True if there's a non-close-paren match of
6918 ;; `c-type-decl-suffix-key'.
6919 got-suffix
6920 ;; True if there's a prefix match outside the outermost
6921 ;; paren pair that surrounds the declarator.
6922 got-prefix-before-parens
6923 ;; True if there's a suffix match outside the outermost
6924 ;; paren pair that surrounds the declarator. The value is
6925 ;; the position of the first suffix match.
6926 got-suffix-after-parens
6927 ;; True if we've parsed the type decl to a token that is
6928 ;; known to end declarations in this context.
6929 at-decl-end
6930 ;; The earlier values of `at-type' and `type-start' if we've
6931 ;; shifted the type backwards.
6932 identifier-type identifier-start
6933 ;; If `c-parse-and-markup-<>-arglists' is set we need to
6934 ;; turn it off during the name skipping below to avoid
6935 ;; getting `c-type' properties that might be bogus. That
6936 ;; can happen since we don't know if
6937 ;; `c-restricted-<>-arglists' will be correct inside the
6938 ;; arglist paren that gets entered.
6939 c-parse-and-markup-<>-arglists
6940 ;; Start of the identifier for which `got-identifier' was set.
6941 name-start)
6945 ;; Skip over type decl prefix operators. (Note similar code in
6946 ;; `c-font-lock-declarators'.)
6956 ;; If the third submatch matches in C++ then
6957 ;; we're looking at an identifier that's a
6958 ;; prefix only if it specifies a member pointer.
6963 ;; We only check for a trailing "::" and
6964 ;; let the "*" that should follow be
6965 ;; matched in the next round.
6967 ;; It turned out to be the real identifier,
6968 ;; so stop.
6969 nil))
6970 t))
6984 ;; Skip over an identifier.
6991 ;; Skip over type decl suffix operators.
6998 t)
7002 t)
7008 ;; No suffix matched. We might have matched the
7009 ;; identifier as a type and the open paren of a
7010 ;; function arglist as a type decl prefix. In that
7011 ;; case we should "backtrack": Reinterpret the last
7012 ;; type as the identifier, move out of the arglist and
7013 ;; continue searching for suffix operators.
7014 ;;
7015 ;; Do this even if there's no preceding type, to cope
7016 ;; with old style function declarations in K&R C,
7017 ;; (con|de)structors in C++ and `c-typeless-decl-kwds'
7018 ;; style declarations. That isn't applicable in an
7019 ;; arglist context, though.
7024 maybe-typeless
7025 backup-maybe-typeless
7032 t))
7039 at-type)
7040 ;; Have found no identifier but `c-typeless-decl-kwds' has
7041 ;; matched so we know we're inside a declaration. The
7042 ;; preceding type must be the identifier instead.
7046 at-decl-or-cast
7049 ;; CASE 1
7051 ;; Encountered something inside parens that isn't matched by
7052 ;; the `c-type-decl-*' regexps, so it's not a type decl
7053 ;; expression. Try to skip out to the same paren depth to
7054 ;; not confuse the cast check below.
7056 ;; If we've found a specifier keyword then it's a
7057 ;; declaration regardless.
7065 ;; Now we've collected info about various characteristics of
7066 ;; the construct we're looking at. Below follows a decision
7067 ;; tree based on that. It's ordered to check more certain
7068 ;; signs before less certain ones.
7073 ;; CASE 2
7076 ;; Got another identifier directly after the type, so it's a
7077 ;; declaration.
7083 ;; (not got-suffix-after-parens)
7085 maybe-typeless
7086 backup-maybe-typeless
7091 ;; Got a declaration of the form "foo bar (gnu);" or "bar
7092 ;; (gnu);" where we've recognized "bar" as the type and "gnu"
7093 ;; as the declarator. In this case it's however more likely
7094 ;; that "bar" is the declarator and "gnu" a function argument
7095 ;; or initializer (if `c-recognize-paren-inits' is set),
7096 ;; since the parens around "gnu" would be superfluous if it's
7097 ;; a declarator. Shift the type one step backward.
7100 ;; Found no identifier.
7105 ;; CASE 3
7107 ;; Got a plain list of identifiers. If a colon follows it's
7108 ;; a valid label, or maybe a bitfield. Otherwise the last
7109 ;; one probably is the declared identifier and we should
7110 ;; back up to the previous type, providing it isn't a cast.
7114 ;; If we've found a specifier keyword then it's a
7115 ;; declaration regardless.
7126 ;; CASE 4
7130 ;; Got a plain list of identifiers followed by some suffix.
7131 ;; If this isn't a cast then the last identifier probably is
7132 ;; the declared one and we should back up to the previous
7133 ;; type.
7137 ;; CASE 5
7139 ;; If the type is known we know that there can't be any
7140 ;; identifier somewhere else, and it's only in declarations in
7141 ;; e.g. function prototypes and in casts that the identifier may
7142 ;; be left out.
7146 ;; Only got a single identifier (parsed as a type so far).
7147 ;; CASE 6
7149 ;; Check that the identifier isn't at the start of an
7150 ;; expression.
7151 at-decl-end
7154 ;; Inside an arglist that contains declarations. If K&R
7155 ;; style declarations and parenthesis style initializers
7156 ;; aren't allowed then the single identifier must be a
7157 ;; type, else we require that it's known or found
7158 ;; (primitive types are handled above).
7163 ;; Inside a template arglist. Accept known and found
7164 ;; types; other identifiers could just as well be
7165 ;; constants in C++.
7168 ;; CASE 7
7169 ;; Can't be a valid declaration or cast, but if we've found a
7170 ;; specifier it can't be anything else either, so treat it as
7171 ;; an invalid/unfinished declaration or cast.
7179 maybe-typeless
7180 backup-maybe-typeless
7184 c-after-suffixed-type-maybe-decl-key))))))
7185 ;; Got an empty paren pair and a preceding type that probably
7186 ;; really is the identifier. Shift the type backwards to make
7187 ;; the last one the identifier. This is analogous to the
7188 ;; "backtracking" done inside the `c-type-decl-suffix-key' loop
7189 ;; above.
7190 ;;
7191 ;; Exception: In addition to the conditions in that
7192 ;; "backtracking" code, do not shift backward if we're not
7193 ;; looking at either `c-after-suffixed-type-decl-key' or "[;,]".
7194 ;; Since there's no preceding type, the shift would mean that
7195 ;; the declaration is typeless. But if the regexp doesn't match
7196 ;; then we will simply fall through in the tests below and not
7197 ;; recognize it at all, so it's better to try it as an abstract
7198 ;; declarator instead.
7201 ;; Still no identifier.
7202 ;; CASE 8
7204 ;; Require `got-prefix' together with either `got-parens' or
7205 ;; `got-suffix' to recognize it as an abstract declarator:
7206 ;; `got-parens' only is probably an empty function call.
7207 ;; `got-suffix' only can build an ordinary expression together
7208 ;; with the preceding identifier which we've taken as a type.
7209 ;; We could actually accept on `got-prefix' only, but that can
7210 ;; easily occur temporarily while writing an expression so we
7211 ;; avoid that case anyway. We could do a better job if we knew
7212 ;; the point when the fontification was invoked.
7215 ;; CASE 9
7219 got-suffix-after-parens
7221 ;; Got a type, no declarator but a paren suffix. I.e. it's a
7222 ;; normal function call after all (or perhaps a C++ style object
7223 ;; instantiation expression).
7226 ;; CASE 10
7228 ;; By now we've located the type in the declaration that we know
7229 ;; we're in.
7232 ;; CASE 11
7240 ;; Shift the type backward in the case that there's a
7241 ;; single identifier inside parens. That can only
7242 ;; occur in K&R style function declarations so it's
7243 ;; more likely that it really is a function call.
7244 ;; Therefore we only do this after
7245 ;; `c-after-suffixed-type-decl-key' has matched.
7247 got-suffix-after-parens))
7248 ;; A declaration according to `c-after-suffixed-type-decl-key'.
7251 ;; CASE 12
7254 ;; It's a declaration if a known type precedes it and it can't be a
7255 ;; function call.
7258 ;; If we get here we can't tell if this is a type decl or a normal
7259 ;; expression by looking at it alone. (That's under the assumption
7260 ;; that normal expressions always can look like type decl expressions,
7261 ;; which isn't really true but the cases where it doesn't hold are so
7262 ;; uncommon (e.g. some placements of "const" in C++) it's not worth
7263 ;; the effort to look for them.)
7265 ;;; 2008-04-16: commented out the next form, to allow the function to recognize
7266 ;;; "foo (int bar)" in CC (an implicit type (in class foo) without a semicolon)
7267 ;;; as a(n almost complete) declaration, enabling it to be fontified.
7268 ;; CASE 13
7269 ;; (unless (or at-decl-end (looking-at "=[^=]"))
7270 ;; If this is a declaration it should end here or its initializer(*)
7271 ;; should start here, so check for allowed separation tokens. Note
7272 ;; that this rule doesn't work e.g. with a K&R arglist after a
7273 ;; function header.
7274 ;;
7275 ;; *) Don't check for C++ style initializers using parens
7276 ;; since those already have been matched as suffixes.
7277 ;;
7278 ;; If `at-decl-or-cast' is then we've found some other sign that
7279 ;; it's a declaration or cast, so then it's probably an
7280 ;; invalid/unfinished one.
7281 ;; (throw 'at-decl-or-cast at-decl-or-cast))
7283 ;; Below are tests that only should be applied when we're certain to
7284 ;; not have parsed halfway through an expression.
7286 ;; CASE 14
7288 ;; The expression starts with a known type so treat it as a
7289 ;; declaration.
7292 ;; CASE 15
7294 ;; In C++ we check if the identifier is a known type, since
7295 ;; (con|de)structors use the class name as identifier.
7296 ;; We've always shifted over the identifier as a type and
7297 ;; then backed up again in this case.
7298 identifier-type
7301 ;; `at-type' probably won't be 'found for
7302 ;; destructors since the "~" is then part of the
7303 ;; type name being checked against the list of
7304 ;; known types, so do a check without that
7305 ;; operator.
7310 c-identifier-syntax-table
7314 ;; We have already parsed the type earlier,
7315 ;; so it'd be possible to cache the end
7316 ;; position instead of redoing it here, but
7317 ;; then we'd need to keep track of another
7318 ;; position everywhere.
7326 ;; CASE 16
7328 at-type
7332 ;; Got something like "foo * bar;". Since we're not inside an
7333 ;; arglist it would be a meaningless expression because the
7334 ;; result isn't used. We therefore choose to recognize it as
7335 ;; a declaration. Do not allow a suffix since it could then
7336 ;; be a function call.
7339 ;; CASE 17
7344 ;; Got something like "a (*b) (c);" or "a (b) = c;". It could
7345 ;; be an odd expression or it could be a declaration. Treat
7346 ;; it as a declaration if "a" has been used as a type
7347 ;; somewhere else (if it's a known type we won't get here).
7350 ;; CASE 18
7356 ;; Got a type followed by an abstract declarator. If `got-prefix'
7357 ;; is set it's something like "a *" without anything after it. If
7358 ;; `got-parens' or `got-suffix' is set it's "a()", "a[]", "a()[]",
7359 ;; or similar, which we accept only if the context rules out
7360 ;; expressions.
7363 ;; If we had a complete symbol table here (which rules out
7364 ;; `c-found-types') we should return t due to the disambiguation rule
7365 ;; (in at least C++) that anything that can be parsed as a declaration
7366 ;; is a declaration. Now we're being more defensive and prefer to
7367 ;; highlight things like "foo (bar);" as a declaration only if we're
7368 ;; inside an arglist that contains declarations.
7371 ;; The point is now after the type decl expression.
7374 ;; Check for a cast.
7377 c-cast-parens
7379 ;; Should be the first type/identifier in a cast paren.
7383 ;; The closing paren should follow.
7388 ;; There should be a primary expression after it.
7397 ;; Check if the expression begins with a prefix keyword.
7400 ;; Expression begins with an ambiguous operator. Treat
7401 ;; it as a cast if it's a type decl or if we've
7402 ;; recognized the type somewhere else.
7405 ;; Unless it's a keyword, it's the beginning of a primary
7406 ;; expression.
7408 ;; If `c-primary-expr-regexp' matched a nonsymbol token, check
7409 ;; that it matched a whole one so that we don't e.g. confuse
7410 ;; the operator '-' with '->'. It's ok if it matches further,
7411 ;; though, since it e.g. can match the float '.5' while the
7412 ;; operator regexp only matches '.'.
7416 ;; There should either be a cast before it or something that isn't an
7417 ;; identifier or close paren.
7425 ;; It's a symbol. Accept it only if it's one of the
7426 ;; keywords that can precede an expression (without
7427 ;; surrounding parens).
7430 ;; Check that it isn't a close paren (block close is ok,
7431 ;; though).
7433 ;; Check that it isn't a nonsymbol identifier.
7436 ;; Handle the cast.
7446 ;; We're at a declaration. Highlight the type and the following
7447 ;; declarators.
7453 ;; Make sure to propagate the `c-decl-arg-start' property to
7454 ;; the next argument if it's set in this one, to cope with
7455 ;; interactive refontification.
7459 ;; There seems no reason to exclude a token from
7460 ;; fontification just because it's "a known type that can't
7461 ;; be a name or other expression". 2013-09-18.
7462 )
7473 ;; False alarm. Restore the recorded ranges.
7475 c-record-ref-identifiers save-rec-ref-ids)
7476 nil))))
7479 ;; Assuming that point is at the beginning of a token, check if it starts a
7480 ;; label and if so move over it and return non-nil (t in default situations,
7481 ;; specific symbols (see below) for interesting situations), otherwise don't
7482 ;; move and return nil. "Label" here means "most things with a colon".
7483 ;;
7484 ;; More precisely, a "label" is regarded as one of:
7485 ;; (i) a goto target like "foo:" - returns the symbol `goto-target';
7486 ;; (ii) A case label - either the entire construct "case FOO:", or just the
7487 ;; bare "case", should the colon be missing. We return t;
7488 ;; (iii) a keyword which needs a colon, like "default:" or "private:"; We
7489 ;; return t;
7490 ;; (iv) One of QT's "extended" C++ variants of
7491 ;; "private:"/"protected:"/"public:"/"more:" looking like "public slots:".
7492 ;; Returns the symbol `qt-2kwds-colon'.
7493 ;; (v) QT's construct "signals:". Returns the symbol `qt-1kwd-colon'.
7494 ;; (vi) One of the keywords matched by `c-opt-extra-label-key' (without any
7495 ;; colon). Currently (2006-03), this applies only to Objective C's
7496 ;; keywords "@private", "@protected", and "@public". Returns t.
7497 ;;
7498 ;; One of the things which will NOT be recognized as a label is a bit-field
7499 ;; element of a struct, something like "int foo:5".
7500 ;;
7501 ;; The end of the label is taken to be just after the colon, or the end of
7502 ;; the first submatch in `c-opt-extra-label-key'. The point is directly
7503 ;; after the end on return. The terminating char gets marked with
7504 ;; `c-decl-end' to improve recognition of the following declaration or
7505 ;; statement.
7506 ;;
7507 ;; If ASSUME-MARKUP is non-nil, it's assumed that the preceding
7508 ;; label, if any, has already been marked up like that.
7509 ;;
7510 ;; If PRECEDING-TOKEN-END is given, it should be the first position
7511 ;; after the preceding token, i.e. on the other side of the
7512 ;; syntactic ws from the point. Use a value less than or equal to
7513 ;; (point-min) if the point is at the first token in (the visible
7514 ;; part of) the buffer.
7515 ;;
7516 ;; The optional LIMIT limits the forward scan for the colon.
7517 ;;
7518 ;; This function records the ranges of the label symbols on
7519 ;; `c-record-ref-identifiers' if `c-record-type-identifiers' (!) is
7520 ;; non-nil.
7521 ;;
7522 ;; This function might do hidden buffer changes.
7525 label-end
7526 qt-symbol-idx
7527 macro-start ; if we're in one.
7528 label-type
7529 kwd)
7531 ;; "case" or "default" (Doesn't apply to AWK).
7534 ;; Record only the keyword itself for fontification, since in
7535 ;; case labels the following is a constant expression and not
7536 ;; a label.
7540 ;; Find the label end.
7544 ;; Stop on chars that aren't allowed in expressions,
7545 ;; and on operator chars that would be meaningless
7546 ;; there. FIXME: This doesn't cope with ?: operators.
7554 t)
7556 ;; It's an unfinished label. We consider the keyword enough
7557 ;; to recognize it as a label, so that it gets fontified.
7558 ;; Leave the point at the end of it, but don't put any
7559 ;; `c-decl-end' marker.
7561 t))))
7563 ;; @private, @protected, @public, in Objective C, or similar.
7566 ;; For a `c-opt-extra-label-key' match, we record the whole
7567 ;; thing for fontification. That's to get the leading '@' in
7568 ;; Objective-C protection labels fontified.
7575 ;; All other cases of labels.
7578 ;; A colon label must have something before the colon.
7581 ;; Check that we're not after a token that can't precede a label.
7583 ;; Trivially succeeds when there's no preceding token.
7584 ;; Succeeds when we're at a virtual semicolon.
7593 ;; Check if we're after a label, if we're after a closing
7594 ;; paren that belong to statement, and with
7595 ;; `c-label-prefix-re'. It's done in different order
7596 ;; depending on `assume-markup' since the checks have
7597 ;; different expensiveness.
7624 ;; Might be after another label, so check it recursively.
7628 ;; Essentially the same as the
7629 ;; `c-syntactic-re-search-forward' regexp below.
7635 ;; Note: the following should work instead of the
7636 ;; narrow-to-region above. Investigate why not,
7637 ;; sometime. ACM, 2006-03-31.
7638 ;; (c-syntactic-skip-backward "^-]:?;}=*/%&|,<>!@+"
7639 ;; macro-start t)
7641 ;; If the caller turned on recording for us,
7642 ;; it shouldn't apply when we check the
7643 ;; preceding label.
7644 c-record-type-identifiers)
7645 ;; A label can't start at a cpp directive. Check for
7646 ;; this, since c-forward-syntactic-ws would foul up on it.
7651 ;; Point is still at the beginning of the possible label construct.
7652 ;;
7653 ;; Check that the next nonsymbol token is ":", or that we're in one
7654 ;; of QT's "slots" declarations. Allow '(' for the sake of macro
7655 ;; arguments. FIXME: Should build this regexp from the language
7656 ;; constants.
7658 ;; public: protected: private:
7668 ;; QT double keyword like "protected slots:" or goto target.
7687 'qt-1kwd-colon
7699 ;; Check that `c-nonlabel-token-key' doesn't match anywhere.
7709 t)
7712 ;; Record the identifiers in the label for fontification, unless
7713 ;; it begins with `c-label-kwds' in which case the following
7714 ;; identifiers are part of a (constant) expression that
7715 ;; shouldn't be fontified.
7727 ;; Not a label.
7729 label-type))
7732 ;; Assuming the point is at the beginning of a token, try to move
7733 ;; forward to the end of the Objective-C directive that starts
7734 ;; there. Return t if a directive was fully recognized, otherwise
7735 ;; the point is moved as far as one could be successfully parsed and
7736 ;; nil is returned.
7737 ;;
7738 ;; This function records identifier ranges on
7739 ;; `c-record-type-identifiers' and `c-record-ref-identifiers' if
7740 ;; `c-record-type-identifiers' is non-nil.
7741 ;;
7742 ;; This function might do hidden buffer changes.
7745 start-char
7747 lim
7748 ;; Turn off recognition of angle bracket arglists while parsing
7749 ;; types here since the protocol reference list might then be
7750 ;; considered part of the preceding name or superclass-name.
7761 t)
7770 ;; Handle the name of the class itself.
7772 ; (c-forward-token-2) ; 2006/1/13 This doesn't move if the token's
7773 ; at EOB.
7780 ;; Look for ": superclass-name" or "( category-name )".
7791 ;; Look for a protocol reference list.
7797 t))))
7803 t)
7806 nil)))
7809 ;; Go to the first non-whitespace after the colon that starts a
7810 ;; multiple inheritance introduction. Optional LIM is the farthest
7811 ;; back we should search.
7812 ;;
7813 ;; This function might do hidden buffer changes.
7821 ;; Return nil if we aren't in a method definition, otherwise the
7822 ;; position of the initial [+-].
7823 ;;
7824 ;; This function might do hidden buffer changes.
7830 ))
7832 ;; Contributed by Kevin Ryde <user42@zip.com.au>.
7834 ;; Return non-nil if point is within a gcc \"asm\" block.
7835 ;;
7836 ;; This should be called with point inside an argument list.
7837 ;;
7838 ;; Only one level of enclosing parentheses is considered, so for
7839 ;; instance `nil' is returned when in a function call within an asm
7840 ;; operand.
7841 ;;
7842 ;; This function might do hidden buffer changes.
7854 a function.
7856 More precisely, being at the top-level means that point is either
7857 outside any enclosing block (such as a function definition), or
7858 directly inside a class, namespace or other block that contains
7859 another declaration level.
7861 If point is not at the top-level (e.g. it is inside a method
7862 definition), then nil is returned. Otherwise, if point is at a
7863 top-level not enclosed within a class definition, t is returned.
7864 Otherwise, a 2-vector is returned where the zeroth element is the
7865 buffer position of the start of the class declaration, and the first
7866 element is the buffer position of the enclosing class's opening
7867 brace.
7869 Note that this function might do hidden buffer changes. See the
7870 comment at the start of cc-engine.el for more info."
7876 ;; Return non-nil if the point is in the region after the argument
7877 ;; list of a function and its opening brace (or semicolon in case it
7878 ;; got no body). If there are K&R style argument declarations in
7879 ;; that region, the point has to be inside the first one for this
7880 ;; function to recognize it.
7881 ;;
7882 ;; If successful, the point is moved to the first token after the
7883 ;; function header (see `c-forward-decl-or-cast-1' for details) and
7884 ;; the position of the opening paren of the function arglist is
7885 ;; returned.
7886 ;;
7887 ;; The point is clobbered if not successful.
7888 ;;
7889 ;; LIM is used as bound for backward buffer searches.
7890 ;;
7891 ;; This function might do hidden buffer changes.
7904 ;; There should not be a '=' or ',' between beg and the
7905 ;; start of the declaration since that means we were in the
7906 ;; "expression part" of the declaration.
7911 ;; Check that there's an arglist paren in the
7912 ;; declaration.
7915 ;; The declarator is a paren expression, so skip past it
7916 ;; so that we don't get stuck on that instead of the
7917 ;; function arglist.
7921 ;; Don't trip up on "operator ()".
7928 ;; Return the position of the first argument declaration if point is
7929 ;; inside a K&R style argument declaration list, nil otherwise.
7930 ;; `c-recognize-knr-p' is not checked. If LIM is non-nil, it's a
7931 ;; position that bounds the backward search for the argument list.
7932 ;;
7933 ;; Point must be within a possible K&R region, e.g. just before a top-level
7934 ;; "{". It must be outside of parens and brackets. The test can return
7935 ;; false positives otherwise.
7936 ;;
7937 ;; This function might do hidden buffer changes.
7941 ;; If we're in a macro, our search range is restricted to it. Narrow to
7942 ;; the searchable range.
7946 before-lparen after-rparen
7948 ; we give up
7951 ;; Search backwards for the defun's argument list. We give up if we
7952 ;; encounter a "}" (end of a previous defun) an "=" (which can't be in
7953 ;; a knr region) or BOB.
7954 ;;
7955 ;; The criterion for a paren structure being the arg list is:
7956 ;; o - there is non-WS stuff after it but before any "{"; AND
7957 ;; o - the token after it isn't a ";" AND
7958 ;; o - it is preceded by either an identifier (the function name) or
7959 ;; a macro expansion like "DEFUN (...)"; AND
7960 ;; o - its content is a non-empty comma-separated list of identifiers
7961 ;; (an empty arg list won't have a knr region).
7962 ;;
7963 ;; The following snippet illustrates these rules:
7964 ;; int foo (bar, baz, yuk)
7965 ;; int bar [] ;
7966 ;; int (*baz) (my_type) ;
7967 ;; int (*) (void) (*yuk) (void) ;
7968 ;; {
7979 ; parens/brackets.
7983 ;; We're inside a paren. Could it be our argument list....?
7989 ;; FIXME!!! What about macros between the parens? 2007/01/20
7992 ;; It can't be the arg list if next token is ; or {
7997 ;; Is the thing preceding the list an identifier (the
7998 ;; function name), or a macro expansion?
8008 ;; Have we got a non-empty list of comma-separated
8009 ;; identifiers?
8023 ;; ...Yes. We've identified the function's argument list.
8029 ;; ...No. The current parens aren't the function's arg list.
8036 ;; skip forward over conditional at point, including any predicate
8037 ;; statements in parentheses. No error checking is performed.
8038 ;;
8039 ;; This function might do hidden buffer changes.
8041 ;; else if()
8046 ;; do, else, try, finally
8051 ;; for, if, while, switch, catch, synchronized, foreach
8055 ;; If looking at the token after a conditional then return the
8056 ;; position of its start, otherwise return nil.
8057 ;;
8058 ;; This function might do hidden buffer changes.
8069 ;; If the point is after an operator identifier that isn't handled
8070 ;; like an ordinary symbol (i.e. like "operator =" in C++) then the
8071 ;; position of the start of that identifier is returned. nil is
8072 ;; returned otherwise. The point may be anywhere in the syntactic
8073 ;; whitespace after the last token of the operator identifier.
8074 ;;
8075 ;; This function might do hidden buffer changes.
8087 ;; Assuming point is at a brace that opens a statement block of some
8088 ;; kind, move to the proper anchor point for that block. It might
8089 ;; need to be adjusted further by c-add-stmt-syntax, but the
8090 ;; position at return is suitable as start position for that
8091 ;; function.
8092 ;;
8093 ;; This function might do hidden buffer changes.
8100 ;; Assuming point is at a brace that opens the block of a top level
8101 ;; declaration of some kind, move to the proper anchor point for
8102 ;; that block.
8103 ;;
8104 ;; This function might do hidden buffer changes.
8109 ;; Search forward for the end of the "header" of the current
8110 ;; declaration. That's the position where the definition body
8111 ;; starts, or the first variable initializer, or the ending
8112 ;; semicolon. I.e. search forward for the closest following
8113 ;; (syntactically relevant) '{', '=' or ';' token. Point is left
8114 ;; _after_ the first found token, or at point-max if none is found.
8115 ;;
8116 ;; This function might do hidden buffer changes.
8121 ;; In C++ we need to take special care to handle operator
8122 ;; tokens and those pesky template brackets.
8127 ;; Handle operator identifiers, i.e. ignore any
8128 ;; operator token preceded by "operator".
8135 t
8137 nil)))))
8146 ;; Go to the beginning of the current declaration, or the beginning
8147 ;; of the previous one if already at the start of it. Point won't
8148 ;; be moved out of any surrounding paren. Return a cons cell of the
8149 ;; form (MOVE . KNR-POS). MOVE is like the return value from
8150 ;; `c-beginning-of-statement-1'. If point skipped over some K&R
8151 ;; style argument declarations (and they are to be recognized) then
8152 ;; KNR-POS is set to the start of the first such argument
8153 ;; declaration, otherwise KNR-POS is nil. If LIM is non-nil, it's a
8154 ;; position that bounds the backward search.
8155 ;;
8156 ;; NB: Cases where the declaration continues after the block, as in
8157 ;; "struct foo { ... } bar;", are currently recognized as two
8158 ;; declarations, e.g. "struct foo { ... }" and "bar;" in this case.
8159 ;;
8160 ;; This function might do hidden buffer changes.
8166 ;; `c-beginning-of-statement-1' stops at a block start, but we
8167 ;; want to continue if the block doesn't begin a top level
8168 ;; construct, i.e. if it isn't preceded by ';', '}', ':', bob,
8169 ;; or an open paren.
8171 ;; Go back one "statement" each time round the loop until we're just
8172 ;; after a ;, }, or :, or at BOB or the start of a macro or start of
8173 ;; an ObjC method. This will move over a multiple declaration whose
8174 ;; components are comma separated.
8176 ;; Must check with c-opt-method-key in ObjC mode.
8186 ;; Check that we don't move from the first thing in a
8187 ;; macro to its header.
8193 move tentative-move))
8198 ;; Handle K&R argdecls. Back up after the "statement" jumped
8199 ;; over by `c-beginning-of-statement-1', unless it was the
8200 ;; function body, in which case we're sitting on the opening
8201 ;; brace now. Then test if we're in a K&R argdecl region and
8202 ;; that we started at the other side of the first argdecl in
8203 ;; it.
8213 'previous
8215 knr-argdecl-start))
8218 ;; `c-beginning-of-statement-1' counts each brace block as a separate
8219 ;; statement, so the result will be 'previous if we've moved over any.
8220 ;; So change our result back to 'same if necessary.
8221 ;;
8222 ;; If they were brace list initializers we might not have moved over a
8223 ;; declaration boundary though, so change it to 'same if we've moved
8224 ;; past a '=' before '{', but not ';'. (This ought to be integrated
8225 ;; into `c-beginning-of-statement-1', so we avoid this extra pass which
8226 ;; potentially can search over a large amount of text.). Take special
8227 ;; pains not to get mislead by C++'s "operator=", and the like.
8230 c++-template-syntax-table
8236 ; no more, or get to one which isn't an "operator ="
8239 c-overloadable-operators-regexp
8240 c-opt-op-identifier-prefix
8255 ;; Assuming point is at the start of a declaration (as detected by
8256 ;; e.g. `c-beginning-of-decl-1'), go to the end of it. Unlike
8257 ;; `c-beginning-of-decl-1', this function handles the case when a
8258 ;; block is followed by identifiers in e.g. struct declarations in C
8259 ;; or C++. If a proper end was found then t is returned, otherwise
8260 ;; point is moved as far as possible within the current sexp and nil
8261 ;; is returned. This function doesn't handle macros; use
8262 ;; `c-end-of-macro' instead in those cases.
8263 ;;
8264 ;; This function might do hidden buffer changes.
8267 c++-template-syntax-table
8275 ;; Stopped at the ';' in a K&R argdecl section which is
8276 ;; detected using the same criteria as in
8277 ;; `c-beginning-of-decl-1'. Move to the following block
8278 ;; start.
8282 ;; Encountered a block in the declaration. Jump over it.
8293 ;; Check for `c-opt-block-decls-with-vars-key'
8294 ;; before the first paren.
8297 c-opt-block-decls-with-vars-key
8299 lim t t t)
8302 ;; Check that the first following paren is
8303 ;; the block.
8305 lim t t t)
8307 ;; The declaration doesn't have any of the
8308 ;; `c-opt-block-decls-with-vars' keywords in the
8309 ;; beginning, so it ends here at the end of the block.
8317 nil)))
8320 ;; Assuming the point is at an open brace, check if it starts a
8321 ;; block that contains another declaration level, i.e. that isn't a
8322 ;; statement block or a brace list, and if so return non-nil.
8323 ;;
8324 ;; If the check is successful, the return value is the start of the
8325 ;; keyword that tells what kind of construct it is, i.e. typically
8326 ;; what `c-decl-block-key' matched. Also, if GOTO-START is set then
8327 ;; the point will be at the start of the construct, before any
8328 ;; leading specifiers, otherwise it's at the returned position.
8329 ;;
8330 ;; The point is clobbered if the check is unsuccessful.
8331 ;;
8332 ;; CONTAINING-SEXP is the position of the open of the surrounding
8333 ;; paren, or nil if none.
8334 ;;
8335 ;; The optional LIMIT limits the backward search for the start of
8336 ;; the construct. It's assumed to be at a syntactically relevant
8337 ;; position.
8338 ;;
8339 ;; If any template arglists are found in the searched region before
8340 ;; the open brace, they get marked with paren syntax.
8341 ;;
8342 ;; This function might do hidden buffer changes.
8349 ;; Could be at the end of a template arglist.
8360 ;; Note: Can't get bogus hits inside template arglists below since they
8361 ;; have gotten paren syntax above.
8363 ;; If `goto-start' is set we begin by searching for the
8364 ;; first possible position of a leading specifier list.
8365 ;; The `c-decl-block-key' search continues from there since
8366 ;; we know it can't match earlier.
8369 open-brace t t)
8371 t)
8372 t)
8379 ;; Found a keyword that can't be a type?
8382 ;; Can be a type too, in which case it's the return type of a
8383 ;; function (under the assumption that no declaration level
8384 ;; block construct starts with a type).
8387 ;; Jumped over a type, but it could be a declaration keyword
8388 ;; followed by the declared identifier that we've jumped over
8389 ;; instead (e.g. in "class Foo {"). If it indeed is a type
8390 ;; then we should be at the declarator now, so check for a
8391 ;; valid declarator start.
8392 ;;
8393 ;; Note: This doesn't cope with the case when a declared
8394 ;; identifier is followed by e.g. '(' in a language where '('
8395 ;; also might be part of a declarator expression. Currently
8396 ;; there's no such language.
8400 ;; In Pike a list of modifiers may be followed by a brace
8401 ;; to make them apply to many identifiers. Note that the
8402 ;; match data will be empty on return in this case.
8408 ;; Use this variant to avoid yet another special regexp.
8412 t)))
8414 ;; Got a match.
8417 ;; Back up over any preceding specifiers and their clauses
8418 ;; by going forward from `first-specifier-pos', which is the
8419 ;; earliest possible position where the specifier list can
8420 ;; start.
8426 ;; Accept any plain symbol token on the ground that
8427 ;; it's a specifier masked through a macro (just
8428 ;; like `c-forward-decl-or-cast-1' skip forward over
8429 ;; such tokens).
8430 ;;
8431 ;; Could be more restrictive wrt invalid keywords,
8432 ;; but that'd only occur in invalid code so there's
8433 ;; no use spending effort on it.
8439 ;; Can't parse a declaration preamble and is still
8440 ;; before `kwd-start'. That means `first-specifier-pos'
8441 ;; was in some earlier construct. Search again.
8445 ;; Got no preamble before the block declaration keyword.
8451 kwd-start)))
8454 ;; Check if the closest containing paren sexp is a declaration
8455 ;; block, returning a 2 element vector in that case. Aref 0
8456 ;; contains the bufpos at boi of the class key line, and aref 1
8457 ;; contains the bufpos of the open brace. This function is an
8458 ;; obsolete wrapper for `c-looking-at-decl-block'.
8459 ;;
8460 ;; This function might do hidden buffer changes.
8468 nil))
8473 ;; Return the buffer position of the most enclosing decl-block brace (in the
8474 ;; sense of c-looking-at-decl-block) in the PAREN-STATE structure, or nil if
8475 ;; none was found.
8484 open-brace))
8487 ;; Return the position of the L-brace if point is inside a brace list
8488 ;; initialization of an array, etc. This is an approximate function,
8489 ;; designed for speed over accuracy. It will not find every bracelist, but
8490 ;; a non-nil result is reliable. We simply search for "= {" (naturally with
8491 ;; syntactic whitespace allowed). PAREN-STATE is the normal thing that it
8492 ;; is everywhere else.
8501 b-pos)))
8504 ;; We're at a "{". Move back to the enum-like keyword that starts this
8505 ;; declaration and return t, otherwise don't move and return nil.
8507 up-sexp-pos before-identifier)
8521 t)
8526 t)
8532 ;; return the buffer position of the beginning of the brace list
8533 ;; statement if we're inside a brace list, otherwise return nil.
8534 ;; CONTAINING-SEXP is the buffer pos of the innermost containing
8535 ;; paren. PAREN-STATE is the remainder of the state of enclosing
8536 ;; braces
8537 ;;
8538 ;; N.B.: This algorithm can potentially get confused by cpp macros
8539 ;; placed in inconvenient locations. It's a trade-off we make for
8540 ;; speed.
8541 ;;
8542 ;; This function might do hidden buffer changes.
8544 ;; This will pick up brace list declarations.
8548 ;; this will pick up array/aggregate init lists, even if they are nested.
8551 ;; Pike can have class definitions anywhere, so we must
8552 ;; check for the class key here.
8554 c-decl-block-key))
8555 bufpos braceassignp lim next-containing macro-start)
8557 containing-sexp)
8568 ;; We're in an in-expression block of some kind. Do not
8569 ;; check nesting. We deliberately set the limit to the
8570 ;; containing sexp, so that c-looking-at-inexpr-block
8571 ;; doesn't check for an identifier before it.
8573 ;; see if the open brace is preceded by = or [...] in
8574 ;; this statement, but watch out for operator=
8577 ;; Checks to do only on the first sexp before the brace.
8580 ;; In Java, an initialization brace list may follow
8581 ;; directly after "new Foo[]", so check for a "new"
8582 ;; earlier.
8588 ;; Carry on looking if this is an
8589 ;; identifier (may contain "." in Java)
8590 ;; or another "[]" sexp.
8593 ;; Checks to do on all sexps before the brace, up to the
8594 ;; beginning of the statement.
8602 ;; We've seen a =, but must check earlier tokens so
8603 ;; that it isn't something that should be ignored.
8609 ;; Check for operator =
8612 nil)
8613 ;; Check for `<opchar>= in Pike.
8616 ;; Special case for Pikes
8617 ;; `[]=, since '[' is not in
8618 ;; the punctuation class.
8621 nil)
8623 ;; make sure we're not in a C++ template
8624 ;; argument assignment
8636 ))))
8637 nil)
8644 ;; We've hit the beginning of the aggregate list.
8649 ;; Brace lists can't contain a semicolon, so we're done.
8654 ;; We've a macro whose expansion starts with the '{'.
8655 ;; Heuristically, if we have a ';' in it we've not got a
8656 ;; brace list, otherwise we have.
8663 containing-sexp nil)
8666 ;; Go up one level
8668 lim nil
8669 next-containing nil)))))
8671 bufpos))
8672 ))
8675 ;; If we're looking at the start of a pike-style list, i.e., `({Â })',
8676 ;; `([Â ])', `(<Â >)', etc., a cons of a cons of its starting and ending
8677 ;; positions and its entry in c-special-brace-lists is returned, nil
8678 ;; otherwise. The ending position is nil if the list is still open.
8679 ;; LIM is the limit for forward search. The point may either be at
8680 ;; the `(' or at the following paren character. Tries to check the
8681 ;; matching closer, but assumes it's correct if no balanced paren is
8682 ;; found (i.e. the case `({ ... } ... )' is detected as _not_ being
8683 ;; a special brace list).
8684 ;;
8685 ;; This function might do hidden buffer changes.
8690 inner-beg end type)
8705 nil)))))
8715 ;; Check balancing of the inner paren
8716 ;; below.
8719 t)
8720 ;; If the inner char isn't a paren then
8721 ;; we can't check balancing, so just
8722 ;; check the char before the outer
8723 ;; closing paren.
8738 ;; Return non-nil if between two statements or declarations, assuming
8739 ;; point is not inside a literal or comment.
8740 ;;
8741 ;; Obsolete - `c-at-statement-start-p' or `c-at-expression-start-p'
8742 ;; are recommended instead.
8743 ;;
8744 ;; This function might do hidden buffer changes.
8749 ;; Return non-nil if we're looking at the beginning of a block
8750 ;; inside an expression. The value returned is actually a cons of
8751 ;; either 'inlambda, 'inexpr-statement or 'inexpr-class and the
8752 ;; position of the beginning of the construct.
8753 ;;
8754 ;; LIM limits the backward search. CONTAINING-SEXP is the start
8755 ;; position of the closest containing list. If it's nil, the
8756 ;; containing paren isn't used to decide whether we're inside an
8757 ;; expression or not. If both LIM and CONTAINING-SEXP are used, LIM
8758 ;; needs to be farther back.
8759 ;;
8760 ;; If CHECK-AT-END is non-nil then extra checks at the end of the
8761 ;; brace block might be done. It should only be used when the
8762 ;; construct can be assumed to be complete, i.e. when the original
8763 ;; starting position was further down than that.
8764 ;;
8765 ;; This function might do hidden buffer changes.
8770 ;; Look at the character after point only as a last resort
8771 ;; when we can't disambiguate.
8791 ;; If the class definition is at the start of
8792 ;; a statement, we don't consider it an
8793 ;; in-expression class.
8801 ;; Also, in Pike we treat it as an
8802 ;; in-expression class if it's used in an
8803 ;; object clone expression.
8819 nil)
8827 ;; Accept several square bracket sexps for
8828 ;; Java array initializations.
8836 block-follows
8837 containing-sexp
8846 nil
8849 res))))
8852 ;; Returns non-nil if we're looking at the end of an in-expression
8853 ;; block, otherwise the same as `c-looking-at-inexpr-block'.
8854 ;; PAREN-STATE is the paren state relevant at the current position.
8855 ;;
8856 ;; This function might do hidden buffer changes.
8858 ;; We currently only recognize a block.
8861 containing-sexp)
8870 paren-state)
8871 containing-sexp)))))
8874 ;; Is there a "virtual semicolon" at POS or point?
8875 ;; (See cc-defs.el for full details of "virtual semicolons".)
8876 ;;
8877 ;; This is true when point is at the last non syntactic WS position on the
8878 ;; line, there is a macro call last on the line, and this particular macro's
8879 ;; name is defined by the regexp `c-vs-macro-regexp' as not needing a
8880 ;; semicolon.
8888 c-macro-with-semi-re
8891 ;; Check we've got nothing after this except comments and empty lines
8892 ;; joined by escaped EOLs.
8905 nil)
8913 ;; Check for one of the listed macros being before point.
8917 t))
8925 \f
8926 ;; `c-guess-basic-syntax' and the functions that precedes it below
8927 ;; implements the main decision tree for determining the syntactic
8928 ;; analysis of the current line of code.
8930 ;; Dynamically bound to t when `c-guess-basic-syntax' is called during
8931 ;; auto newline analysis.
8935 ;; BRACEPOS is the position of a brace in a construct like "namespace
8936 ;; Bar {". Return the anchor point in this construct; this is the
8937 ;; earliest symbol on the brace's line which isn't earlier than
8938 ;; "namespace".
8939 ;;
8940 ;; Currently (2007-08-17), "like namespace" means "matches
8941 ;; c-other-block-decl-kwds". It doesn't work with "class" or "struct"
8942 ;; or anything like that.
8952 ;; A simple function to prepend a new syntax element to
8953 ;; `c-syntactic-context'. Using `setq' on it is unsafe since it
8954 ;; should always be dynamically bound but since we read it first
8955 ;; we'll fail properly anyway if this function is misused.
8957 c-syntactic-context)))
8960 ;; Like `c-add-syntax' but appends to the end of the syntax list.
8961 ;; (Normally not necessary.)
8966 syntax-extra-args
8967 stop-at-boi-only
8968 containing-sexp
8969 paren-state)
8970 ;; Add the indicated SYNTAX-SYMBOL to `c-syntactic-context', extending it as
8971 ;; needed with further syntax elements of the types `substatement',
8972 ;; `inexpr-statement', `arglist-cont-nonempty', `statement-block-intro', and
8973 ;; `defun-block-intro'.
8974 ;;
8975 ;; Do the generic processing to anchor the given syntax symbol on
8976 ;; the preceding statement: Skip over any labels and containing
8977 ;; statements on the same line, and then search backward until we
8978 ;; find a statement or block start that begins at boi without a
8979 ;; label or comment.
8980 ;;
8981 ;; Point is assumed to be at the prospective anchor point for the
8982 ;; given SYNTAX-SYMBOL. More syntax entries are added if we need to
8983 ;; skip past open parens and containing statements. Most of the added
8984 ;; syntax elements will get the same anchor point - the exception is
8985 ;; for an anchor in a construct like "namespace"[*] - this is as early
8986 ;; as possible in the construct but on the same line as the {.
8987 ;;
8988 ;; [*] i.e. with a keyword matching c-other-block-decl-kwds.
8989 ;;
8990 ;; SYNTAX-EXTRA-ARGS are a list of the extra arguments for the
8991 ;; syntax symbol. They are appended after the anchor point.
8992 ;;
8993 ;; If STOP-AT-BOI-ONLY is nil, we can stop in the middle of the line
8994 ;; if the current statement starts there.
8995 ;;
8996 ;; Note: It's not a problem if PAREN-STATE "overshoots"
8997 ;; CONTAINING-SEXP, i.e. contains info about parens further down.
8998 ;;
8999 ;; This function might do hidden buffer changes.
9002 ;; This is by far the most common case, so let's give it special
9003 ;; treatment.
9008 ;; Set when we're on a label, so that we don't stop there.
9009 ;; FIXME: To be complete we should check if we're on a label
9010 ;; now at the start.
9011 on-label)
9013 ;; Use point as the anchor point for "namespace", "extern", etc.
9017 syntax-extra-args)
9019 ;; Loop while we have to back out of containing blocks.
9024 ;; Loop while we have to back up statements.
9026 on-label
9029 ;; Skip past any comments that stands between the
9030 ;; statement start and boi.
9038 ;; Skip to the beginning of this statement or backward
9039 ;; another one.
9047 ;; If we didn't move we're at the start of a block and
9048 ;; have to continue outside it.
9057 ;; Special case to avoid deeper and deeper indentation
9058 ;; of "else if" clauses.
9059 )
9064 ;; If stop-at-boi-only is nil, we shouldn't back up
9065 ;; over previous or containing statements to try to
9066 ;; reach boi, so go back to the last position and
9067 ;; exit.
9074 ;; If we've moved into another statement then we
9075 ;; should no longer try to stop in the middle of a
9076 ;; line.
9079 ;; Record this as a substatement if we skipped up one
9080 ;; level.
9083 )))
9085 containing-sexp)
9087 ;; Now we have to go out of this block.
9090 ;; Don't stop in the middle of a special brace list opener
9091 ;; like "({".
9103 ;; Analyze the construct in front of the block we've stepped out
9104 ;; from and add the right syntactic element for it.
9107 step-type)
9110 ;; Stepped out of a parenthesis block, so we're in an
9111 ;; expression now.
9123 ;; Stepped out of an in-expression statement. This
9124 ;; syntactic element won't get an anchor pos.
9127 ;; A parenthesis normally belongs to an arglist.
9135 on-label nil))
9137 ;; Stepped out of a brace block.
9149 containing-sexp)))
9151 'defun-block-intro
9153 nil))
9157 c-other-decl-block-key-in-symbols-alist))
9164 ;; Always done if the open brace was at boi. The
9165 ;; c-beginning-of-statement-1 call above is necessary
9166 ;; anyway, to decide the type of block-intro to add.
9169 ))
9171 ;; Fill in the current point as the anchor for all the symbols
9172 ;; added above.
9181 )))
9184 containing-decl-open
9185 containing-decl-start
9186 containing-decl-kwd
9187 paren-state)
9188 ;; The inclass and class-close syntactic symbols are added in
9189 ;; several places and some work is needed to fix everything.
9190 ;; Therefore it's collected here.
9191 ;;
9192 ;; This function might do hidden buffer changes.
9197 containing-decl-open)
9199 ;; Ought to use `c-add-stmt-syntax' instead of backing up to boi
9200 ;; here, but we have to do like this for compatibility.
9210 char-after-ip
9211 beg-of-same-or-containing-stmt
9212 containing-sexp
9213 paren-state)
9214 ;; This function contains the decision tree reached through both
9215 ;; cases 18 and 10. It's a continued statement or top level
9216 ;; construct of some kind.
9217 ;;
9218 ;; This function might do hidden buffer changes.
9225 ;; (CASE A removed.)
9226 ;; CASE B: open braces for class or brace-lists
9233 ;; CASE B.1: class-open
9240 ;; CASE B.2: brace-list-open
9245 indent-point t t t)))
9246 ;; The most semantically accurate symbol here is
9247 ;; brace-list-open, but we normally report it simply as a
9248 ;; statement-cont. The reason is that one normally adjusts
9249 ;; brace-list-open for brace lists as top-level constructs,
9250 ;; and brace lists inside statements is a completely different
9251 ;; context. C.f. case 5A.3.
9254 ;; Turn off the dwim above when we're
9255 ;; analyzing the nature of the brace
9256 ;; for the auto newline feature.
9257 'brace-list-open
9259 nil nil
9260 containing-sexp paren-state))
9262 ;; CASE B.3: The body of a function declared inside a normal
9263 ;; block. Can occur e.g. in Pike and when using gcc
9264 ;; extensions, but watch out for macros followed by blocks.
9265 ;; C.f. cases E, 16F and 17G.
9271 ;; Turn off recognition of constructs that lacks a
9272 ;; type in this case, since that's more likely to be
9273 ;; a macro followed by a block.
9276 containing-sexp paren-state))
9278 ;; CASE B.4: Continued statement with block open. The most
9279 ;; accurate analysis is perhaps `statement-cont' together with
9280 ;; `block-open' but we play DWIM and use `substatement-open'
9281 ;; instead. The rationale is that this typically is a macro
9282 ;; followed by a block which makes it very similar to a
9283 ;; statement with a substatement block.
9286 containing-sexp paren-state))
9287 ))
9289 ;; CASE C: iostream insertion or extraction operator
9293 ;; If there is no preceding streamop in the statement
9294 ;; then indent this line as a normal statement-cont.
9298 t))))
9300 ;; CASE E: In the "K&R region" of a function declared inside a
9301 ;; normal block. C.f. case B.3.
9303 ;; Check that the next token is a '{'. This works as
9304 ;; long as no language that allows nested function
9305 ;; definitions allows stuff like member init lists, K&R
9306 ;; declarations or throws clauses there.
9307 ;;
9308 ;; Note that we do a forward search for something ahead
9309 ;; of the indentation line here. That's not good since
9310 ;; the user might not have typed it yet. Unfortunately
9311 ;; it's exceedingly tricky to recognize a function
9312 ;; prototype in a code block without resorting to this.
9320 ;; Turn off recognition of constructs that lacks a
9321 ;; type in this case, since that's more likely to be
9322 ;; a macro followed by a block.
9325 containing-sexp paren-state))
9327 ;;CASE F: continued statement and the only preceding items are
9328 ;;annotations.
9336 t)
9343 ;; CASE G: a template list continuation?
9344 ;; Mostly a duplication of case 5D.3 to fix templates-19:
9362 ;; In a nested template arglist.
9368 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
9369 ;; template aware.
9372 ;; CASE D: continued statement.
9376 containing-sexp paren-state))
9377 )))
9379 ;; The next autoload was added by RMS on 2005/8/9 - don't know why (ACM,
9380 ;; 2005/11/29).
9381 ;;;###autoload
9383 "Return the syntactic context of the current line."
9389 open-paren-in-column-0-is-defun-start
9390 ;; A whole ugly bunch of various temporary variables. Have
9391 ;; to declare them here since it's not possible to declare
9392 ;; a variable with only the scope of a cond test and the
9393 ;; following result clauses, and most of this function is a
9394 ;; single gigantic cond. :P
9395 literal char-before-ip before-ws-ip char-after-ip macro-start
9396 in-macro-expr c-syntactic-context placeholder c-in-literal-cache
9397 step-type tmpsymbol keyword injava-inher special-brace-list tmp-pos
9398 containing-<
9399 ;; The following record some positions for the containing
9400 ;; declaration block if we're directly within one:
9401 ;; `containing-decl-open' is the position of the open
9402 ;; brace. `containing-decl-start' is the start of the
9403 ;; declaration. `containing-decl-kwd' is the keyword
9404 ;; symbol of the keyword that tells what kind of block it
9405 ;; is.
9406 containing-decl-open
9407 containing-decl-start
9408 containing-decl-kwd
9409 ;; The open paren of the closest surrounding sexp or nil if
9410 ;; there is none.
9411 containing-sexp
9412 ;; The position after the closest preceding brace sexp
9413 ;; (nested sexps are ignored), or the position after
9414 ;; `containing-sexp' if there is none, or (point-min) if
9415 ;; `containing-sexp' is nil.
9416 lim
9417 ;; The paren state outside `containing-sexp', or at
9418 ;; `indent-point' if `containing-sexp' is nil.
9420 ;; There's always at most one syntactic element which got
9421 ;; an anchor pos. It's stored in syntactic-relpos.
9422 syntactic-relpos
9425 ;; Check if we're directly inside an enclosing declaration
9426 ;; level block.
9435 containing-sexp)
9436 t)))
9439 containing-sexp nil)
9444 ;; Init some position variables.
9453 ;; Ignore balanced paren. The next entry
9454 ;; can't be another one.
9457 ;; If there is no surrounding open paren then
9458 ;; put the last balanced pair back on paren-state.
9460 containing-sexp nil)))
9464 ;; If we're in a parenthesis list then ',' delimits the
9465 ;; "statements" rather than being an operator (with the
9466 ;; exception of the "for" clause). This difference is
9467 ;; typically only noticeable when statements are used in macro
9468 ;; arglists.
9472 ;; cache char before and after indent point, and move point to
9473 ;; the most likely position to perform the majority of tests
9482 ;; are we in a literal?
9485 ;; now figure out syntactic qualities of the current line
9488 ;; CASE 1: in a string.
9492 ;; CASE 2: in a C or C++ style comment.
9494 ;; This is a kludge for XEmacs where we use
9495 ;; `buffer-syntactic-context', which doesn't correctly
9496 ;; recognize "\*/" to end a block comment.
9497 ;; `parse-partial-sexp' which is used by
9498 ;; `c-literal-limits' will however do that in most
9499 ;; versions, which results in that we get nil from
9500 ;; `c-literal-limits' even when `c-in-literal' claims
9501 ;; we're inside a comment.
9505 ;; CASE 3: in a cpp preprocessor macro continuation.
9515 ;; If at the beginning of the body of a #define
9516 ;; directive then analyze as cpp-define-intro
9517 ;; only. Go on with the syntactic analysis
9518 ;; otherwise. in-macro-expr is set if we're in a
9519 ;; cpp expression, i.e. before the #define body
9520 ;; or anywhere in a non-#define directive.
9527 nil)))))
9531 ;; CASE 11: an else clause?
9535 containing-sexp paren-state))
9537 ;; CASE 12: while closure of a do/while construct?
9545 containing-sexp paren-state))
9547 ;; CASE 13: A catch or finally clause? This case is simpler
9548 ;; than if-else and do-while, because a block is required
9549 ;; after every try, catch and finally.
9557 t)
9561 t)
9564 t))
9569 containing-sexp paren-state))
9571 ;; CASE 18: A substatement we can recognize by keyword.
9581 ;; Ought to cache the result from the
9582 ;; c-beginning-of-statement-1 calls here.
9592 ;; Step up to the containing statement if we
9593 ;; stayed in the same one.
9601 ;; There was no containing statement after all.
9603 placeholder))
9605 ;; Require a parenthesis after these keywords.
9606 ;; Necessary to catch e.g. synchronized in Java,
9607 ;; which can be used both as statement and
9608 ;; modifier.
9614 ;; CASE 18A: Simple substatement.
9620 containing-sexp paren-state))
9626 containing-sexp paren-state))
9629 containing-sexp paren-state))))
9631 ;; CASE 18B: Some other substatement. This is shared
9632 ;; with case 10.
9634 char-after-ip
9635 placeholder
9636 lim
9637 paren-state)))
9639 ;; CASE 14: A case or default label
9645 containing-sexp))
9648 ;; Got a bogus label at the top level. In lack of better
9649 ;; alternatives, anchor it on (point-min).
9652 ;; CASE 15: any other label
9659 containing-decl-open
9660 containing-decl-start
9661 containing-decl-kwd
9662 paren-state))
9663 ;; Append access-label with the same anchor point as
9664 ;; inclass gets.
9670 containing-sexp))
9675 ;; If the surrounding statement is a switch then
9676 ;; let's analyze all labels as switch labels, so
9677 ;; that they get lined up consistently.
9678 'case-label
9684 ;; A label on the top level. Treat it as a class
9685 ;; context. (point-min) is the closest we get to the
9686 ;; class open brace.
9689 ;; CASE 4: In-expression statement. C.f. cases 7B, 16A and
9690 ;; 17E.
9693 containing-sexp
9694 ;; Have to turn on the heuristics after
9695 ;; the point even though it doesn't work
9696 ;; very well. C.f. test case class-16.pike.
9697 t))
9702 ;; It's a statement block or an anonymous class.
9704 ;; It's a Pike lambda. Check whether we are between the
9705 ;; lambda keyword and the argument list or at the defun
9706 ;; opener.
9708 'inline-open
9714 paren-state)
9718 ;; CASE 5: Line is inside a declaration level block or at top level.
9722 ;; CASE 5A: we are looking at a defun, brace list, class,
9723 ;; or inline-inclass method opening brace
9730 ;; CASE 5A.1: Non-class declaration block open.
9745 ;; Special case for extern-lang-open.
9746 'extern-lang-open
9748 nil t containing-sexp paren-state))
9750 ;; CASE 5A.2: we are looking at a class opening brace
9759 ;; CASE 5A.3: brace list open
9784 ))))
9788 ;; We're in Java and have found that the open brace
9789 ;; belongs to a "new Foo[]" initialization list,
9790 ;; which means the brace list is part of an
9791 ;; expression and not a top level definition. We
9792 ;; therefore treat it as any topmost continuation
9793 ;; even though the semantically correct symbol still
9794 ;; is brace-list-open, on the same grounds as in
9795 ;; case B.2.
9801 ;; CASE 5A.4: inline defun open
9807 containing-decl-open
9808 containing-decl-start
9809 containing-decl-kwd
9810 paren-state))
9812 ;; CASE 5A.5: ordinary defun open
9820 ;; Bogus to use bol here, but it's the legacy. (Resolved,
9821 ;; 2007-11-09)
9822 ))))
9824 ;; CASE 5R: Member init list. (Used to be part of CASE 5B.1)
9825 ;; Note there is no limit on the backward search here, since member
9826 ;; init lists can, in practice, be very large.
9832 ;; There is no preceding member init clause.
9833 ;; Indent relative to the beginning of indentation
9834 ;; for the topmost-intro line that contains the
9835 ;; prototype's open paren.
9838 ;; Indent relative to the first member init clause.
9843 ;; CASE 5B: After a function header but before the body (or
9844 ;; the ending semicolon if there's no body).
9851 ;; CASE 5B.1: Member init list.
9853 ;; There is no preceding member init clause.
9854 ;; Indent relative to the beginning of indentation
9855 ;; for the topmost-intro line that contains the
9856 ;; prototype's open paren.
9860 ;; CASE 5B.2: K&R arg decl intro
9867 containing-decl-open
9868 containing-decl-start
9869 containing-decl-kwd
9870 paren-state)))
9872 ;; CASE 5B.4: Nether region after a C++ or Java func
9873 ;; decl, which could include a `throws' declaration.
9877 )))
9879 ;; CASE 5C: inheritance line. could be first inheritance
9880 ;; line, or continuation of a multiple inheritance
9888 ;; watch out for scope operator
9893 )))
9902 ;; for Java
9907 cont done)
9912 done t))
9916 )
9918 injava-inher)
9921 ))
9924 ;; CASE 5C.1: non-hanging colon on an inher intro
9928 ;; don't add inclass symbol since relative point already
9929 ;; contains any class offset
9930 )
9932 ;; CASE 5C.2: hanging colon on an inher intro
9938 containing-decl-open
9939 containing-decl-start
9940 containing-decl-kwd
9941 paren-state)))
9943 ;; CASE 5C.3: in a Java implements/extends
9957 )))
9959 ;; CASE 5C.4: a continued inheritance line
9963 ;; don't add inclass symbol since relative point already
9964 ;; contains any class offset
9965 )))
9967 ;; CASE 5P: AWK pattern or function or continuation
9968 ;; thereof.
9974 'topmost-intro-cont
9976 nil nil
9977 containing-sexp paren-state))
9979 ;; CASE 5D: this could be a top-level initialization, a
9980 ;; member init list continuation, or a template argument
9981 ;; list continuation.
9983 ;; Note: We use the fact that lim is always after any
9984 ;; preceding brace sexp.
9994 t))
9999 t)
10005 nil)
10006 t))
10008 ;; NB: No c-after-special-operator-id stuff in this
10009 ;; clause - we assume only C++ needs it.
10014 ;; CASE 5D.3: perhaps a template list continuation?
10029 ;; In a nested template arglist.
10035 ;; FIXME: Should use c-add-stmt-syntax, but it's not yet
10036 ;; template aware.
10039 ;; CASE 5D.4: perhaps a multiple inheritance line?
10051 t)
10056 ;; CASE 5D.5: Continuation of the "expression part" of a
10057 ;; top level construct. Or, perhaps, an unrecognized construct.
10069 ;; A preceding comma at the top level means that a
10070 ;; new variable declaration starts here. Use
10071 ;; topmost-intro-cont for it, for consistency with
10072 ;; the first variable declaration. C.f. case 5N.
10075 nil nil containing-sexp paren-state))
10076 ))
10078 ;; CASE 5F: Close of a non-class declaration level block.
10082 ;; This is inconsistent: Should use `containing-decl-open'
10083 ;; here if it's at boi, like in case 5J.
10087 ;; Special case for compatibility with the
10088 ;; extern-lang syntactic symbols.
10089 'extern-lang-close
10092 nil t
10094 paren-state))
10096 ;; CASE 5G: we are looking at the brace which closes the
10097 ;; enclosing nested class decl
10102 containing-decl-open
10103 containing-decl-start
10104 containing-decl-kwd
10105 paren-state))
10107 ;; CASE 5H: we could be looking at subsequent knr-argdecls
10114 ;; Do an extra check to avoid tripping up on
10115 ;; statements that occur in invalid contexts
10116 ;; (e.g. in macro bodies where we don't really
10117 ;; know the context of what we're looking at).
10124 ;; CASE 5I: ObjC method definition.
10129 ;; Handle the case when it's the first (non-comment)
10130 ;; thing in the buffer. Can't look for a 'same return
10131 ;; value from cbos1 since ObjC directives currently
10132 ;; aren't recognized fully, so that we get 'same
10133 ;; instead of 'previous if it moved over a preceding
10134 ;; directive.
10138 ;; CASE 5N: At a variable declaration that follows a class
10139 ;; definition or some other block declaration that doesn't
10140 ;; end at the closing '}'. C.f. case 5D.5.
10149 ;; Speed up the backward search a bit.
10154 ;; The '}' is unbalanced.
10155 nil
10160 containing-sexp paren-state))
10162 ;; NOTE: The point is at the end of the previous token here.
10164 ;; CASE 5J: we are at the topmost level, make
10165 ;; sure we skip back past any access specifiers
10167 ;; A macro continuation line is never at top level.
10189 nil)))))
10190 ;; For historic reasons we anchor at bol of the last
10191 ;; line of the previous declaration. That's clearly
10192 ;; highly bogus and useless, and it makes our lives hard
10193 ;; to remain compatible. :P
10204 ;; Special case for compatibility with the
10205 ;; extern-lang syntactic symbols.
10206 'inextern-lang
10209 nil t
10211 paren-state))
10213 containing-decl-open
10214 containing-decl-start
10215 containing-decl-kwd
10216 paren-state)))
10218 macro-start
10223 ;; CASE 5K: we are at an ObjC method definition
10224 ;; continuation line.
10233 ;; CASE 5L: we are at the first argument of a template
10234 ;; arglist that begins on the previous line.
10242 ;; CASE 5Q: we are at a statement within a macro.
10247 ;;CASE 5N: We are at a topmost continuation line and the only
10248 ;;preceding items are annotations.
10255 t)
10261 ;; CASE 5M: we are at a topmost continuation line
10272 ))
10275 ;; (CASE 6 has been removed.)
10277 ;; CASE 7: line is an expression, not a statement. Most
10278 ;; likely we are either in a function prototype or a function
10279 ;; call argument list
10287 ;; CASE 7A: we are looking at the arglist closing paren.
10288 ;; C.f. case 7F.
10300 paren-state))
10302 ;; CASE 7B: Looking at the opening brace of an
10303 ;; in-expression block or brace list. C.f. cases 4, 16A
10304 ;; and 17E.
10308 paren-state))
10312 placeholder
10315 paren-state)
10316 containing-sexp)))
10317 ;; placeholder is nil if it's a block directly in
10318 ;; a function arglist. That makes us skip out of
10319 ;; this case.
10320 )))
10325 paren-state)
10329 ;; CASE 7C: we are looking at the first argument in an empty
10330 ;; argument list. Use arglist-close if we're actually
10331 ;; looking at a close paren or bracket.
10343 paren-state))
10345 ;; CASE 7D: we are inside a conditional test clause. treat
10346 ;; these things as statements
10356 ))
10358 ;; CASE 7E: maybe a continued ObjC method call. This is the
10359 ;; case when we are inside a [] bracketed exp, and what
10360 ;; precede the opening bracket is not an identifier.
10368 )))
10370 ;; CASE 7F: we are looking at an arglist continuation line,
10371 ;; but the preceding argument is on the same line as the
10372 ;; opening paren. This case includes multi-line
10373 ;; mathematical paren groupings, but we could be on a
10374 ;; for-list continuation line. C.f. case 7A.
10391 paren-state))
10393 ;; CASE 7G: we are looking at just a normal arglist
10394 ;; continuation line
10397 ))
10399 ;; CASE 8: func-local multi-inheritance line
10409 ;; CASE 8A: non-hanging colon on an inher intro
10414 ;; CASE 8B: hanging colon on an inher intro
10418 ;; CASE 8C: a continued inheritance line
10422 )))
10424 ;; CASE 9: we are inside a brace-list
10434 ;; CASE 9A: In the middle of a special brace list opener.
10454 ;; CASE 9B: brace-list-close brace
10456 ;; Check special brace list closer.
10463 ;; We were between the special close char and the `)'.
10466 ;; We were before the special close char.
10470 ;; Normal brace list check.
10481 ;; Prepare for the rest of the cases below by going to the
10482 ;; token following the opening brace
10495 ;; CASE 9C: we're looking at the first line in a brace-list
10506 ;; CASE 9D: this is just a later brace-list-entry or
10507 ;; brace-entry-open
10516 ))
10517 ))))
10519 ;; CASE 10: A continued statement or top level construct.
10529 ;; This is shared with case 18.
10531 char-after-ip
10532 placeholder
10533 containing-sexp
10534 paren-state))
10536 ;; CASE 16: block close brace, possibly closing the defun or
10537 ;; the class
10539 ;; From here on we have the next containing sexp in lim.
10544 ;; CASE 16E: Closing a statement block? This catches
10545 ;; cases where it's preceded by a statement keyword,
10546 ;; which works even when used in an "invalid" context,
10547 ;; e.g. a macro argument.
10552 ;; CASE 16A: closing a lambda defun or an in-expression
10553 ;; block? C.f. cases 4, 7B and 17E.
10556 nil))
10558 'inline-close
10568 paren-state)
10572 ;; CASE 16B: does this close an inline or a function in
10573 ;; a non-class declaration level block?
10580 nil))
10590 ;; CASE 16F: Can be a defun-close of a function declared
10591 ;; in a statement block, e.g. in Pike or when using gcc
10592 ;; extensions, but watch out for macros followed by
10593 ;; blocks. Let it through to be handled below.
10594 ;; C.f. cases B.3 and 17G.
10600 ;; Turn off recognition of constructs that
10601 ;; lacks a type in this case, since that's more
10602 ;; likely to be a macro followed by a block.
10609 ;; CASE 16C: If there is an enclosing brace then this is
10610 ;; a block close since defun closes inside declaration
10611 ;; level blocks have been handled above.
10613 ;; If the block is preceded by a case/switch label on
10614 ;; the same line, we anchor at the first preceding label
10615 ;; at boi. The default handling in c-add-stmt-syntax
10616 ;; really fixes it better, but we do like this to keep
10617 ;; the indentation compatible with version 5.28 and
10618 ;; earlier. C.f. case 17H.
10625 ;; c-backward-to-block-anchor not necessary here; those
10626 ;; situations are handled in case 16E above.
10629 ;; CASE 16D: Only top level defun close left.
10635 paren-state))
10636 ))
10638 ;; CASE 19: line is an expression, not a statement, and is directly
10639 ;; contained by a template delimiter. Most likely, we are in a
10640 ;; template arglist within a statement. This case is based on CASE
10641 ;; 7. At some point in the future, we may wish to create more
10642 ;; syntactic symbols such as `template-intro',
10643 ;; `template-cont-nonempty', etc., and distinguish between them as we
10644 ;; do for `arglist-intro' etc. (2009-12-07).
10650 ; '<') before indent-point.
10658 paren-state))
10660 ;; CASE 17: Statement or defun catchall.
10663 ;; Back up statements until we find one that starts at boi.
10666 containing-sexp)))
10670 nil)
10675 ;; CASE 17B: continued statement
10679 containing-sexp paren-state))
10681 ;; CASE 17A: After a case/default label?
10689 'statement-case-open
10691 nil t containing-sexp paren-state))
10693 ;; CASE 17D: any old statement
10700 containing-sexp paren-state)
10704 ;; CASE 17I: Inside a substatement block.
10706 ;; The following tests are all based on containing-sexp.
10708 ;; From here on we have the next containing sexp in lim.
10713 lim paren-state)
10717 ;; CASE 17E: first statement in an in-expression block.
10718 ;; C.f. cases 4, 7B and 16A.
10721 nil))
10723 'defun-block-intro
10732 paren-state)
10738 ;; CASE 17F: first statement in an inline, or first
10739 ;; statement in a top-level defun. we can tell this is it
10740 ;; if there are no enclosing braces that haven't been
10741 ;; narrowed out by a class (i.e. don't use bod here).
10744 paren-state)))
10750 nil))))
10755 ;; CASE 17G: First statement in a function declared inside
10756 ;; a normal block. This can occur in Pike and with
10757 ;; e.g. the gcc extensions, but watch out for macros
10758 ;; followed by blocks. C.f. cases B.3 and 16F.
10764 ;; Turn off recognition of constructs that lacks
10765 ;; a type in this case, since that's more likely
10766 ;; to be a macro followed by a block.
10772 lim paren-state))
10774 ;; CASE 17H: First statement in a block.
10776 ;; If the block is preceded by a case/switch label on the
10777 ;; same line, we anchor at the first preceding label at
10778 ;; boi. The default handling in c-add-stmt-syntax is
10779 ;; really fixes it better, but we do like this to keep the
10780 ;; indentation compatible with version 5.28 and earlier.
10781 ;; C.f. case 16C.
10788 ;; c-backward-to-block-anchor not necessary here; those
10789 ;; situations are handled in case 17I above.
10791 lim paren-state))
10794 ))
10795 )
10797 ;; now we need to look at any modifiers
10801 ;; are we looking at a comment only line?
10806 ;; we might want to give additional offset to friends (in C++).
10811 ;; Set syntactic-relpos.
10817 nil)
10818 t))
10821 ;; Start of or a continuation of a preprocessor directive?
10836 ;; If inside a cpp expression, i.e. anywhere in a
10837 ;; cpp directive except a #define body, we only let
10838 ;; through the syntactic analysis that is internal
10839 ;; in the expression. That means the arglist
10840 ;; elements, if they are anchored inside the cpp
10841 ;; expression.
10850 ;; Inside a #define body and the syntactic analysis is
10851 ;; anchored on the start of the #define. In this case
10852 ;; we add cpp-define-intro to get the extra
10853 ;; indentation of the #define body.
10856 ;; return the syntax
10857 c-syntactic-context)))
10859 \f
10860 ;; Indentation calculation.
10863 ;; offset can be a number, a function, a variable, a list, or one of
10864 ;; the symbols + or -
10865 ;;
10866 ;; This function might do hidden buffer changes.
10885 langelem symbol))
10891 offset symbol)
10892 nil)
10906 Error evaluating offset %S for %s: \
10907 Cannot combine absolute offset %S with relative %S in `%s' method"
10913 Error evaluating offset %S for %s: \
10914 Cannot combine relative offset %S with absolute %S in `%s' method"
10919 res))
10937 Error evaluating offset %S for %s: \
10938 Cannot combine absolute offsets %S and %S in `add' method"
10942 res))
10951 res))))
10958 nil))))
10962 res
10964 offset symbol res)
10965 nil)))
10968 ;; Get offset from LANGELEM which is a list beginning with the
10969 ;; syntactic symbol and followed by any analysis data it provides.
10970 ;; That data may be zero or more elements, but if at least one is
10971 ;; given then the first is the anchor position (or nil). The symbol
10972 ;; is matched against `c-offsets-alist' and the offset calculated
10973 ;; from that is returned.
10974 ;;
10975 ;; This function might do hidden buffer changes.
10985 offset
10989 ))
10992 ;; This is a compatibility wrapper for `c-calc-offset' in case
10993 ;; someone is calling it directly. It takes an old style syntactic
10994 ;; element on the form (SYMBOL . ANCHOR-POS) and converts it to the
10995 ;; new list form.
10996 ;;
10997 ;; This function might do hidden buffer changes.
11004 ;; Calculate the syntactic indentation from a syntactic description
11005 ;; as returned by `c-guess-syntax'.
11006 ;;
11007 ;; Note that topmost-intro always has an anchor position at bol, for
11008 ;; historical reasons. It's often used together with other symbols
11009 ;; that has more sane positions. Since we always use the first
11010 ;; found anchor position, we rely on that these other symbols always
11011 ;; precede topmost-intro in the LANGELEMS list.
11012 ;;
11013 ;; This function might do hidden buffer changes.
11021 ;; Got an absolute column that overrides any indentation
11022 ;; we've collected so far, but not the relative
11023 ;; indentation we might get for the nested structures
11024 ;; further down the langelems list.
11028 ;; Got a relative change of the current calculated
11029 ;; indentation.
11032 ;; Use the anchor position from the first syntactic
11033 ;; element with one.
11043 indent)))
11045 \f
11048 ;;; cc-engine.el ends here