| blob | f22b75218da20a7821349885c943e836e5a2af0b |
1 ;;; cc-langs.el --- language specific settings for CC Mode
3 ;; Copyright (C) 1985, 1987, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
4 ;; 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
5 ;; Free Software Foundation, Inc.
7 ;; Authors: 2002- Alan Mackenzie
8 ;; 1998- Martin Stjernholm
9 ;; 1992-1999 Barry A. Warsaw
10 ;; 1987 Dave Detlefs and Stewart Clamen
11 ;; 1985 Richard M. Stallman
12 ;; Maintainer: bug-cc-mode@gnu.org
13 ;; Created: 22-Apr-1997 (split from cc-mode.el)
14 ;; Version: See cc-mode.el
15 ;; Keywords: c languages oop
17 ;; This file is part of GNU Emacs.
19 ;; GNU Emacs is free software; you can redistribute it and/or modify
20 ;; it under the terms of the GNU General Public License as published by
21 ;; the Free Software Foundation; either version 3, or (at your option)
22 ;; any later version.
24 ;; GNU Emacs is distributed in the hope that it will be useful,
25 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
26 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
27 ;; GNU General Public License for more details.
29 ;; You should have received a copy of the GNU General Public License
30 ;; along with this program; see the file COPYING. If not, write to
31 ;; the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
32 ;; Boston, MA 02110-1301, USA.
34 ;;; Commentary:
36 ;; HACKERS NOTE: There's heavy macro magic here. If you need to make
37 ;; changes in this or other files containing `c-lang-defconst' but
38 ;; don't want to read through the longer discussion below then read
39 ;; this:
40 ;;
41 ;; o A change in a `c-lang-defconst' or `c-lang-defvar' will not take
42 ;; effect if the file containing the mode init function (typically
43 ;; cc-mode.el) is byte compiled.
44 ;; o To make changes show in font locking you need to reevaluate the
45 ;; `*-font-lock-keywords-*' constants, which normally is easiest to
46 ;; do with M-x eval-buffer in cc-fonts.el.
47 ;; o In either case it's necessary to reinitialize the mode to make
48 ;; the changes show in an existing buffer.
50 ;;; Introduction to the language dependent variable system:
51 ;;
52 ;; This file contains all the language dependent variables, except
53 ;; those specific for font locking which reside in cc-fonts.el. As
54 ;; far as possible, all the differences between the languages that CC
55 ;; Mode supports are described with these variables only, so that the
56 ;; code can be shared.
57 ;;
58 ;; The language constant system (see cc-defs.el) is used to specify
59 ;; various language dependent info at a high level, such as lists of
60 ;; keywords, and then from them generate - at compile time - the
61 ;; various regexps and other low-level structures actually employed in
62 ;; the code at runtime.
63 ;;
64 ;; This system is also designed to make it easy for developers of
65 ;; derived modes to customize the source constants for new language
66 ;; variants, without having to keep up with the exact regexps etc that
67 ;; are used in each CC Mode version. It's possible from an external
68 ;; package to add a new language by inheriting an existing one, and
69 ;; then change specific constants as necessary for the new language.
70 ;; The old values for those constants (and the values of all the other
71 ;; high-level constants) may be used to build the new ones, and those
72 ;; new values will in turn be used by the low-level definitions here
73 ;; to build the runtime constants appropriately for the new language
74 ;; in the current version of CC Mode.
75 ;;
76 ;; Like elsewhere in CC Mode, the existence of a doc string signifies
77 ;; that a language constant is part of the external API, and that it
78 ;; therefore can be used with a high confidence that it will continue
79 ;; to work with future versions of CC Mode. Even so, it's not
80 ;; unlikely that such constants will change meaning slightly as this
81 ;; system is refined further; a certain degree of dependence on the CC
82 ;; Mode version is unavoidable when hooking in at this level. Also
83 ;; note that there's still work to be done to actually use these
84 ;; constants everywhere inside CC Mode; there are still hardcoded
85 ;; values in many places in the code.
86 ;;
87 ;; Separate packages will also benefit from the compile time
88 ;; evaluation; the byte compiled file(s) for them will contain the
89 ;; compiled runtime constants ready for use by (the byte compiled) CC
90 ;; Mode, and the source definitions in this file don't have to be
91 ;; loaded then. However, if a byte compiled package is loaded that
92 ;; has been compiled with a different version of CC Mode than the one
93 ;; currently loaded, then the compiled-in values will be discarded and
94 ;; new ones will be built when the mode is initialized. That will
95 ;; automatically trig a load of the file(s) containing the source
96 ;; definitions (i.e. this file and/or cc-fonts.el) if necessary.
97 ;;
98 ;; A small example of a derived mode is available at
99 ;; <http://cc-mode.sourceforge.net/derived-mode-ex.el>. It also
100 ;; contains some useful hints for derived mode developers.
102 ;;; Using language variables:
103 ;;
104 ;; The `c-lang-defvar' forms in this file comprise the language
105 ;; variables that CC Mode uses. It does not work to use
106 ;; `c-lang-defvar' anywhere else (which isn't much of a limitation
107 ;; since these variables sole purpose is to interface with the CC Mode
108 ;; core functions). The values in these `c-lang-defvar's are not
109 ;; evaluated right away but instead collected to a single large `setq'
110 ;; that can be inserted for a particular language with the
111 ;; `c-init-language-vars' macro.
113 ;; This file is only required at compile time, or when not running
114 ;; from byte compiled files, or when the source definitions for the
115 ;; language constants are requested.
117 ;;; Code:
124 load-path)))
131 ;; This file is not always loaded. See note above.
134 \f
135 ;;; Setup for the `c-lang-defvar' system.
138 ;; These are used to collect the init forms from the subsequent
139 ;; `c-lang-defvar' and `c-lang-setvar'. They are used to build the
140 ;; lambda in `c-make-init-lang-vars-fun' below, and to build `defvar's
141 ;; and `make-variable-buffer-local's in cc-engine and
142 ;; `make-local-variable's in `c-init-language-vars-for'.
146 c-lang-variable-inits-tail c-lang-variable-inits)
150 c-emacs-variable-inits-tail c-emacs-variable-inits))
153 "Declares the buffer local variable VAR to get the value VAL. VAL is
154 evaluated and assigned at mode initialization. More precisely, VAL is
155 evaluated and bound to VAR when the result from the macro
156 `c-init-language-vars' is evaluated.
158 `c-lang-const' is typically used in VAL to get the right value for the
159 language being initialized, and such calls will be macro expanded to
160 the evaluated constant value at compile time."
166 ;; Special case: If there's no docstring and the value is a
167 ;; simple (c-lang-const foo) where foo is the same name as VAR
168 ;; then take the docstring from the language constant foo.
180 ;; Return the symbol, like the other def* forms.
184 "Causes the variable VAR to be made buffer local and to get set to the
185 value VAL. VAL is evaluated and assigned at mode initialization. More
186 precisely, VAL is evaluated and bound to VAR when the result from the
187 macro `c-init-language-vars' is evaluated. VAR is typically a standard
188 Emacs variable like `comment-start'.
190 `c-lang-const' is typically used in VAL to get the right value for the
191 language being initialized, and such calls will be macro expanded to
192 the evaluated constant value at compile time."
199 ;; Return the symbol, like the other def* forms.
203 ; (eval-after-load "edebug" ; 2006-07-09: def-edebug-spec is now in subr.el.
204 ; '
209 ;; Some helper functions used when building the language constants.
212 ;; Extract a subset of the operators in the list OPS in a DWIM:ey
213 ;; way. The return value is a plain list of operators:
214 ;;
215 ;; OPS either has the structure of `c-operators', is a single
216 ;; group in `c-operators', or is a plain list of operators.
217 ;;
218 ;; OPGROUP-FILTER specifies how to select the operator groups. It
219 ;; can be t to choose all groups, a list of group type symbols
220 ;; (such as 'prefix) to accept, or a function which will be called
221 ;; with the group symbol for each group and should return non-nil
222 ;; if that group is to be included.
223 ;;
224 ;; If XLATE is given, it's a function which is called for each
225 ;; matching operator and its return value is collected instead.
226 ;; If it returns a list, the elements are spliced directly into
227 ;; the final result, which is returned as a list with duplicates
228 ;; removed using `equal'.
229 ;;
230 ;; `c-mode-syntax-table' for the current mode is in effect during
231 ;; the whole procedure.
252 t)
253 t)
258 opgroup)))
259 ops)
262 \f
263 ;;; Various mode specific values that aren't language related.
266 ;; The definition for the mode menu. The menu title is prepended to
267 ;; this before it's fed to `easy-menu-define'.
277 "----"
281 ;; Only applicable if there's a cpp preprocessor.
285 "----"
290 "----"
302 c-subword-mode)])))
304 \f
305 ;;; Syntax tables.
308 "Populate the given syntax table as necessary for a C-like language.
326 ;; Set up block and line oriented comments. The new C
327 ;; standard mandates both comment styles even in C, so since
328 ;; all languages now require dual comments, we make this the
329 ;; default.
331 ;; XEmacs
335 ;; Emacs
339 ;; incompatible
343 ;; Give CR the same syntax as newline, for selective-display
347 "Functions that generates the mode specific syntax tables.
348 The syntax tables aren't stored directly since they're quite large."
352 ;; Mode specific syntaxes.
354 ;; Let '@' be part of symbols in ObjC to cope with
355 ;; its compiler directives as single keyword tokens.
356 ;; This is then necessary since it's assumed that
357 ;; every keyword is a single symbol.
361 table)))
364 ;; The syntax tables in evaluated form. Only used temporarily when
365 ;; the constants in this file are evaluated.
369 ;; A variant of `c++-mode-syntax-table' that defines `<' and `>' as
370 ;; parenthesis characters. Used temporarily when template argument
371 ;; lists are parsed. Note that this encourages incorrect parsing of
372 ;; templates since they might contain normal operators that uses the
373 ;; '<' and '>' characters. Therefore this syntax table might go
374 ;; away when CC Mode handles templates correctly everywhere.
375 t nil
380 table)))
386 "A list that describes the modifications that should be done to the
387 mode syntax table to get a syntax table that matches all identifiers
388 and keywords as words.
390 The list is just like the one used in `font-lock-defaults': Each
391 element is a cons where the car is the character to modify and the cdr
392 the new syntax, as accepted by `modify-syntax-entry'."
393 ;; The $ character is not allowed in most languages (one exception
394 ;; is Java which allows it for legacy reasons) but we still classify
395 ;; it as an indentifier character since it's often used in various
396 ;; machine generated identifiers.
407 mod)
412 table)
413 "Syntax table built on the mode syntax table but additionally
414 classifies symbol constituents like '_' and '$' as word constituents,
418 "If non-nil, a function called from c-before-change-hook.
419 Typically it will record enough state to allow
420 `c-before-font-lock-function' to extend the region to fontify,
421 and may do such things as removing text-properties which must be
422 recalculated.
424 It takes 2 parameters, the BEG and END supplied to every
425 before-change function; on entry, the buffer will have been
426 widened and match-data will have been saved; point is undefined
427 on both entry and exit; the return value is ignored.
429 When the mode is initialized, this function is called with
430 parameters \(point-min) and \(point-max)."
431 t nil
438 "If non-nil, a function called just before font locking.
439 Typically it will extend the region about to be fontified \(see
440 below) and will set `syntax-table' text properties on the region.
442 It takes 3 parameters, the BEG, END, and OLD-LEN supplied to
443 every after-change function; point is undefined on both entry and
444 exit; on entry, the buffer will have been widened and match-data
445 will have been saved; the return value is ignored.
447 The function may extend the region to be fontified by setting the
448 buffer local variables c-new-BEG and c-new-END.
450 The function is called even when font locking is disabled.
452 When the mode is initialized, this function is called with
453 parameters \(point-min), \(point-max) and <buffer size>."
454 t nil
460 \f
461 ;;; Lexer-level syntax (identifiers, tokens etc).
464 "Regexp that matches the start of a symbol, i.e. any identifier or
466 operator at the top level."
473 "Set of characters that can be part of a symbol.
474 This is on the form that fits inside [ ] in a regexp."
475 ;; Pike note: With the backquote identifiers this would include most
476 ;; operator chars too, but they are handled with other means instead.
481 "Regexp matching identifiers and keywords (with submatch 0). Assumed
482 to match if `c-symbol-start' matches on the same position."
486 ;; Use the value from C here since the operator backquote is
487 ;; covered by the other alternative.
495 ;; Number of regexp grouping parens in `c-symbol-key'.
499 "This is the set of chars that can't be part of a symbol, i.e. the
500 negation of `c-symbol-chars'."
505 "Regexp that matches any character that can't be part of a symbol.
506 It's usually appended to other regexps to avoid matching a prefix.
507 It's assumed to not contain any submatchers."
508 ;; The same thing regarding Unicode identifiers applies here as to
509 ;; `c-symbol-key'.
513 "The operators that make up fully qualified identifiers. nil in
514 languages that don't have such things. See `c-operators' for a
515 description of the format. Binary operators can concatenate symbols,
519 This value is by default merged into `c-operators'."
520 t nil
524 ;; Java has "." to concatenate identifiers but it's also used for
525 ;; normal indexing. There's special code in the Java font lock
526 ;; rules to fontify qualified identifiers based on the standard
527 ;; naming conventions. We still define "." here to make
528 ;; `c-forward-name' move over as long names as possible which is
529 ;; necessary to e.g. handle throws clauses correctly.
538 ;; Appendable adorned regexp matching the operators that join
539 ;; symbols to fully qualified identifiers, or nil in languages that
540 ;; don't have such things.
541 ;;
542 ;; This was a docstring constant in 5.30. It still works but is now
543 ;; considered internal - change `c-identifier-ops' instead.
546 t)))
554 ;; Number of regexp grouping parens in `c-opt-identifier-concat-key'.
558 ;; Appendable adorned regexp matching operators that might precede
559 ;; an identifier and that are part of the identifier in that case.
560 ;; nil in languages without such things.
563 t)))
568 "Operators that can occur after a binary operator on `c-identifier-ops'
569 in identifiers. nil in languages that don't have such things.
571 Operators here should also have appropriate entries in `c-operators' -
572 it's not taken care of by default."
573 t nil
574 ;; '~' for destructors in C++, '*' for member pointers.
576 ;; In Java we recognize '*' to deal with "foo.bar.*" that can occur
577 ;; in import declarations. (This will also match bogus things like
578 ;; "foo.*bar" but we don't bother.)
582 ;; Regexp that must match the token after
583 ;; `c-opt-identifier-concat-key' for it to be considered an
584 ;; identifier concatenation operator (which e.g. causes the
585 ;; preceding identifier to be fontified as a reference). Assumed to
586 ;; be a string if `c-opt-identifier-concat-key' is.
587 ;;
588 ;; This was a docstring constant in 5.30. It still works but is now
589 ;; considered internal - change `c-after-id-concat-ops' instead.
597 "Regexp that matches the start of an (optionally qualified) identifier.
598 It should also match all keywords. It's unspecified how far it
599 matches."
609 C++. It does not recognize the full range of syntactic whitespace
610 between the tokens; `c-forward-name' has to be used for that. It
611 should also not match identifiers containing parenthesis groupings,
613 ;; This regexp is more complex than strictly necessary to ensure
614 ;; that it can be matched with a minimum of backtracking.
635 ;; For flexibility, consider the symbol match
636 ;; optional if we've hit a
637 ;; `c-after-id-concat-ops' operator. This is
638 ;; also necessary to handle the "*" that can
639 ;; end import declaration identifiers in Java.
653 ;; This was a docstring constant in 5.30 but it's no longer used.
654 ;; It's only kept to avoid breaking third party code.
655 ;;
656 ;; Used to identify the submatch in `c-identifier-key' that
657 ;; surrounds the last symbol in the qualified identifier. It's a
658 ;; list of submatch numbers, of which the first that has a match is
659 ;; taken. It's assumed that at least one does when the regexp has
660 ;; matched.
661 t nil)
664 "Set if the language support backslash escaped newlines inside string
665 literals."
666 t nil
672 "Set if the language supports multiline string literals without escaped
673 newlines. If t, all string literals are multiline. If a character,
674 only literals where the open quote is immediately preceded by that
675 literal are multiline."
676 t nil
677 pike ?#)
682 "Regexp matching the prefix of a cpp directive in the languages that
683 normally use that macro preprocessor. Tested at bol or at boi.
685 ;; TODO (ACM, 2005-04-01). Amend the following to recognise escaped NLs;
686 ;; amend all uses of c-opt-cpp-prefix which count regexp-depth.
692 "Regexp matching the prefix of a cpp directive anchored to BOL,
693 in the languages that have a macro preprocessor."
699 "Regexp matching the prefix of a cpp directive including the directive
700 name, or nil in languages without preprocessor support. The first
701 submatch surrounds the directive name."
705 ;; Pike, being a scripting language, recognizes hash-bangs too.
711 "List of cpp directives (without the prefix) that are followed by a
712 string message."
718 "List of cpp directives (without the prefix) that are followed by a
719 file name in angle brackets or quotes."
725 "Cpp directive (without the prefix) that is followed by a macro
726 definition, or nil if the language doesn't have any."
731 ;; Regexp matching everything up to the macro body of a cpp define, or the
732 ;; end of the logical line if there is none. Submatch 1 is the name of the
733 ;; macro. Set if c-opt-cpp-macro-define is.
738 ;; ^ ^ #defined name
744 ;; Regexp matching everything up to the end of the identifier defined
745 ;; by a cpp define.
754 "List of cpp directives (without the prefix) that are followed by an
755 expression."
760 "List of functions in cpp expressions."
766 "List of all assignment operators."
772 idl nil)
775 "List describing all operators, along with their precedence and
776 associativity. The order in the list corresponds to the precedence of
777 the operators: The operators in each element are a group with the same
778 precedence, and the group has higher precedence than the groups in all
779 following elements. The car of each element describes the type of the
780 operator group, and the cdr is a list of the operator tokens in it.
781 The operator group types are:
783 'prefix Unary prefix operators.
784 'postfix Unary postfix operators.
785 'postfix-if-paren
786 Unary postfix operators if and only if the chars have
787 parenthesis syntax.
788 'left-assoc Binary left associative operators (i.e. a+b+c means (a+b)+c).
789 'right-assoc Binary right associative operators (i.e. a=b=c means a=(b=c)).
790 'right-assoc-sequence
791 Right associative operator that constitutes of a
792 sequence of tokens that separate expressions. All the
793 tokens in the group are in this case taken as
794 describing the sequence in one such operator, and the
795 order between them is therefore significant.
797 Operators containing a character with paren syntax are taken to match
798 with a corresponding open/close paren somewhere else. A postfix
799 operator with close paren syntax is taken to end a postfix expression
800 started somewhere earlier, rather than start a new one at point. Vice
801 versa for prefix operators with open paren syntax.
804 often are described as postfix operators are considered binary here,
805 since CC Mode treats every identifier as an expression."
807 ;; There's currently no code in CC Mode that exploit all the info
808 ;; in this variable; precedence, associativity etc are present as a
809 ;; preparation for future work.
820 ;; Primary.
829 ;; Postfix.
831 ;; The following need special treatment.
841 ;; Unary.
848 ;; The following need special treatment.
859 ;; Member selection.
863 ;; Multiplicative.
866 ;; Additive.
869 ;; Shift.
874 ;; Relational.
879 ;; Equality.
883 ;; Bitwise and.
887 ;; Bitwise exclusive or.
893 ;; Bitwise or.
899 ;; Logical and.
903 ;; Logical or.
909 ;; Conditional.
912 ;; Assignment.
915 ;; Exception.
919 ;; Sequence.
922 ;; IDL got its own definition since it has a much smaller operator
923 ;; set than the other languages.
927 ;; Primary.
929 ;; Unary.
931 ;; Multiplicative.
933 ;; Additive.
935 ;; Shift.
937 ;; And.
939 ;; Xor.
941 ;; Or.
945 ;; The operators as a flat list (without duplicates).
951 t nil
960 ;; These work like identifiers in Pike.
967 ;; Regexp tested after an "operator" token in C++.
968 t nil
974 "Regexp matching the token before the ones in
975 `c-overloadable-operators' when operators are specified in their
978 languages without operator functions or where the complete operator
979 identifier is listed in `c-overloadable-operators'.
981 This regexp is assumed to not match any non-operator identifier."
982 t nil
987 ;; Note: the following alias is an old name which was a mis-spelling. It has
988 ;; been corrected above and throughout cc-engine.el. It will be removed at
989 ;; some release very shortly in the future. ACM, 2006-04-14.
995 "List of the tokens made up of characters in the punctuation or
996 parenthesis syntax classes that have uses other than as expression
997 operators."
1015 ;; List of all tokens in the punctuation and parenthesis syntax
1016 ;; classes.
1022 ;; List containing all chars not in the word, symbol or
1023 ;; syntactically irrelevant syntax classes, i.e. all punctuation,
1024 ;; parenthesis and string delimiter chars.
1026 ;; Only go through the chars in the printable ASCII range. No
1027 ;; language so far has 8-bit or widestring operators.
1033 list)))
1036 ;; Regexp matching all tokens in the punctuation and parenthesis
1037 ;; syntax classes. Note that this also matches ".", which can start
1038 ;; a float.
1041 t
1047 ;; Regexp matching all assignment operators and only them. The
1048 ;; beginning of the first submatch is used to detect the end of the
1049 ;; token, along with the end of the whole match.
1052 ;; Need special case for "=" since it's a prefix of "==".
1064 ;; Regexp matching all tokens containing "<" or ">" which are longer
1065 ;; than one char.
1068 t
1074 ;; Regexp matching the second and subsequent characters of all
1075 ;; multicharacter tokens that begin with "<".
1078 t
1084 ;; Regexp matching the second and subsequent characters of all
1085 ;; multicharacter tokens that begin with ">".
1088 t
1094 ;; The characters that should be considered to bound statements. To
1095 ;; optimize `c-crosses-statement-barrier-p' somewhat, it's assumed to
1096 ;; begin with "^" to negate the set. If ? : operators should be
1097 ;; detected then the string must end with "?:".
1098 t "^;{}?:"
1103 ;; Variant of `c-stmt-delim-chars' that additionally contains ','.
1104 t "^;,{}?:"
1109 \f
1110 ;;; Syntactic whitespace.
1113 "Regexp matching an ordinary whitespace character.
1115 ;; "\\s " is not enough since it doesn't match line breaks.
1119 ;; Number of regexp grouping parens in `c-simple-ws'.
1123 "String that starts line comments, or nil if such don't exist.
1124 Line comments are always terminated by newlines. At least one of
1125 `c-block-comment-starter' and this one is assumed to be set.
1127 Note that it's currently not enough to set this to support a new
1128 comment style. Other stuff like the syntax table must also be set up
1129 properly."
1130 t "//"
1135 "String that starts block comments, or nil if such don't exist.
1136 Block comments are ended by `c-block-comment-ender', which is assumed
1137 to be set if this is. At least one of `c-line-comment-starter' and
1138 this one is assumed to be set.
1140 Note that it's currently not enough to set this to support a new
1141 comment style. Other stuff like the syntax table must also be set up
1142 properly."
1143 t "/*"
1144 awk nil)
1147 "String that ends block comments, or nil if such don't exist.
1149 Note that it's currently not enough to set this to support a new
1150 comment style. Other stuff like the syntax table must also be set up
1151 properly."
1152 t "*/"
1153 awk nil)
1156 ;; Regexp to match the start of any type of comment.
1162 re)))
1165 ;;;; Added by ACM, 2003/9/18.
1167 ;; Regexp which matches the start of a block comment (if such exists in the
1168 ;; language)
1176 ;; Regexp to match the start of comments and string literals.
1185 "Regexp to match the start of documentation comments."
1187 ;; From font-lock.el: `doxygen' uses /*! while others use /**.
1195 "String that starts comments inserted with M-; etc.
1196 `comment-start' is initialized from this."
1197 ;; Default: Prefer line comments to block comments, and pad with a space.
1201 ;; In C we still default to the block comment style since line
1202 ;; comments aren't entirely portable.
1207 "String that ends comments inserted with M-; etc.
1208 `comment-end' is initialized from this."
1209 ;; Default: Use block comment style if comment-start uses block
1210 ;; comments, and pad with a space in that case.
1220 "Regexp to match the start of a comment plus everything up to its body.
1221 `comment-start-skip' is initialized from this."
1222 ;; Default: Allow the last char of the comment starter(s) to be
1223 ;; repeated, then allow any amount of horizontal whitespace.
1236 ;; Regexp matching any sequence that can start syntactic whitespace.
1237 ;; The only uncertain case is '#' when there are cpp directives.
1251 ;; Regexp matching any single character that might end syntactic whitespace.
1266 ;; Regexp matching an unterminated block comment that doesn't
1267 ;; contain line breaks, or nil in languages without block comments.
1268 ;; Does not contain a \| operator at the top level.
1272 ;; It's messy to cook together a regexp that matches anything
1273 ;; but c-block-comment-ender.
1281 "[^"
1282 ;; The quoting rules inside char classes are silly. :P
1291 "]"
1299 ;; Regexp matching a block comment that doesn't contain line breaks,
1300 ;; or nil in languages without block comments. The reason we don't
1301 ;; allow line breaks is to avoid going very far and risk running out
1302 ;; of regexp stack; this regexp is intended to handle only short
1303 ;; comments that might be put in the middle of limited constructs
1304 ;; like declarations. Does not contain a \| operator at the top
1305 ;; level.
1320 ;; Regexp matching a piece of syntactic whitespace that isn't a
1321 ;; sequence of simple whitespace characters. As opposed to
1322 ;; `c-(forward|backward)-syntactic-ws', this doesn't regard cpp
1323 ;; directives as syntactic whitespace.
1335 ;; Regexp matching syntactic whitespace, including possibly the
1336 ;; empty string. As opposed to `c-(forward|backward)-syntactic-ws',
1337 ;; this doesn't regard cpp directives as syntactic whitespace. Does
1338 ;; not contain a \| operator at the top level.
1346 ;; Number of regexp grouping parens in `c-syntactic-ws'.
1350 ;; Regexp matching syntactic whitespace, which is at least one
1351 ;; character long. As opposed to `c-(forward|backward)-syntactic-ws',
1352 ;; this doesn't regard cpp directives as syntactic whitespace. Does
1353 ;; not contain a \| operator at the top level.
1361 ;; Number of regexp grouping parens in `c-nonempty-syntactic-ws'.
1365 ;; Regexp matching syntactic whitespace without any line breaks. As
1366 ;; opposed to `c-(forward|backward)-syntactic-ws', this doesn't
1367 ;; regard cpp directives as syntactic whitespace. Does not contain
1368 ;; a \| operator at the top level.
1376 ;; Number of regexp grouping parens in `c-single-line-syntactic-ws'.
1380 ;; Regexp that matches when there is no syntactically significant
1381 ;; text before eol. Macros are regarded as syntactically
1382 ;; significant text here.
1384 ;; Match eol (possibly inside a block comment or preceded
1385 ;; by a line continuation backslash), or the beginning of a
1386 ;; line comment. Note: This has to be modified for awk
1387 ;; where line comments start with '#'.
1401 \f
1402 ;;; Syntactic analysis ("virtual semicolons") for line-oriented languages (AWK).
1405 Such a function takes one optional parameter, a buffer position (defaults to
1406 point), and returns nil or t. This variable contains nil for languages which
1407 don't have EOL terminated statements. "
1408 t nil
1414 The (admittedly kludgey) purpose of such a function is to prevent an infinite
1415 recursion in c-beginning-of-statement-1 when point starts at a `while' token.
1416 The function MUST NOT UNDER ANY CIRCUMSTANCES call c-beginning-of-statement-1,
1417 even indirectly. This variable contains nil for languages which don't have
1418 EOL terminated statements."
1419 t nil
1424 \f
1425 ;;; Defun functions
1427 ;; The Emacs variables beginning-of-defun-function and
1428 ;; end-of-defun-function will be set so that commands like
1429 ;; `mark-defun' and `narrow-to-defun' work right. The key sequences
1430 ;; C-M-a and C-M-e are, however, bound directly to the CC Mode
1431 ;; functions, allowing optimisation for large n.
1433 "Function to which beginning-of-defun-function will be set."
1434 t 'c-beginning-of-defun
1440 "Function to which end-of-defun-function will be set."
1441 t 'c-end-of-defun
1444 \f
1445 ;;; In-comment text handling.
1448 "Regexp to append to `paragraph-start'."
1449 t "$"
1455 "Regexp to append to `paragraph-separate'."
1456 t "$"
1460 \f
1461 ;;; Keyword lists.
1463 ;; Note: All and only all language constants containing keyword lists
1464 ;; should end with "-kwds"; they're automatically collected into the
1465 ;; `c-kwds-lang-consts' list below and used to build `c-keywords' etc.
1468 "Primitive type keywords. As opposed to the other keyword lists, the
1469 keywords listed here are fontified with the type face instead of the
1470 keyword face.
1472 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1473 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1474 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1475 will be handled.
1477 Do not try to modify this list for end user customizations; the
1478 `*-font-lock-extra-types' variable, where `*' is the mode prefix, is
1479 the appropriate place for that."
1488 ;; Objective-C extends C, but probably not the new stuff in C99.
1495 ;; In CORBA PSDL:
1496 "ref"
1497 ;; The following can't really end a type, but we have to specify them
1498 ;; here due to the assumption in `c-primitive-type-prefix-kwds'. It
1499 ;; doesn't matter that much.
1502 ;; used as a builtin type.
1507 ;; An adorned regexp that matches `c-primitive-type-kwds'.
1512 "Keywords that might act as prefixes for primitive types. Assumed to
1513 be a subset of `c-primitive-type-kwds'."
1514 t nil
1517 ;; In CORBA PSDL:
1521 "Keywords where the following name - if any - is a type name, and
1522 where the keyword together with the symbol works as a type in
1523 declarations.
1525 Note that an alternative if the second part doesn't hold is
1526 `c-type-list-kwds'. Keywords on this list are typically also present
1527 on one of the `*-decl-kwds' lists."
1528 t nil
1534 ;; Adorned regexp matching `c-type-prefix-kwds'.
1539 "Type modifier keywords. These can occur almost anywhere in types
1540 but they don't build a type of themselves. Unlike the keywords on
1541 `c-primitive-type-kwds', they are fontified with the keyword face and
1542 not the type face."
1543 t nil
1549 ;; Adorned regexp matching `c-type-modifier-kwds', or nil in
1550 ;; languages without such keywords.
1556 ;; An adorned regexp that matches `c-primitive-type-prefix-kwds' and
1557 ;; `c-type-modifier-kwds', or nil in languages without any of them.
1567 ;; All keywords that can start a type (i.e. are either a type prefix
1568 ;; or a complete type).
1575 "Keywords introducing declarations where the following block (if any)
1576 contains another declaration level that should be considered a class.
1578 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1579 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1580 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1581 will be handled.
1583 Note that presence on this list does not automatically treat the
1584 following identifier as a type; the keyword must also be present on
1585 `c-type-prefix-kwds' or `c-type-list-kwds' to accomplish that."
1586 t nil
1594 ;; In CORBA PSDL:
1596 ;; In CORBA CIDL:
1601 ;; Regexp matching the start of a class.
1606 "Keywords introducing declarations where the following block (if
1607 any) is a brace list.
1609 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1610 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1611 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1612 will be handled."
1617 ;; Regexp matching the start of declarations where the following
1618 ;; block is a brace list.
1623 "Keywords where the following block (if any) contains another
1624 declaration level that should not be considered a class. For every
1625 keyword here, CC Mode will add a set of special syntactic symbols for
1627 `foo-open', `foo-close', and `infoo' symbols.
1629 The intention is that this category should be used for block
1630 constructs that aren't related to object orientation concepts like
1631 classes (which thus also include e.g. interfaces, templates,
1632 contracts, structs, etc). The more pragmatic distinction is that
1633 while most want some indentation inside classes, it's fairly common
1634 that they don't want it in some of these constructs, so it should be
1635 simple to configure that differently from classes. See also
1636 `c-class-decl-kwds'.
1638 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1639 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1640 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1641 will be handled."
1642 t nil
1646 ;; In CORBA CIDL:
1650 ;; Regexp matching the start of blocks besides classes that contain
1651 ;; another declaration level.
1660 'inextern-lang
1663 "Alist associating keywords in c-other-decl-block-decl-kwds with
1667 "Keywords introducing declarations where the identifier(s) being
1668 declared are types.
1670 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1671 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1672 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1673 will be handled."
1674 ;; Default to `c-class-decl-kwds' and `c-brace-list-decl-kwds'
1675 ;; (since e.g. "Foo" is a type that's being defined in "class Foo
1676 ;; {...}").
1679 ;; Languages that have a "typedef" construct.
1682 ;; Unlike most other languages, exception names are not handled as
1683 ;; types in IDL since they only can occur in "raises" specs.
1687 "Keywords introducing declarations where the \(first) identifier
1688 \(declarator) follows directly after the keyword, without any type.
1690 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1691 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1692 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1693 will be handled."
1694 ;; Default to `c-class-decl-kwds' and `c-brace-list-decl-kwds'
1695 ;; (since e.g. "Foo" is the identifier being defined in "class Foo
1696 ;; {...}").
1699 ;; Note: "manages" for CORBA CIDL clashes with its presence on
1700 ;; `c-type-list-kwds' for IDL.
1703 ;; In CORBA PSDL:
1705 ;; In CORBA CIDL:
1711 "Keywords that can prefix normal declarations of identifiers
1712 \(and typically act as flags). Things like argument declarations
1713 inside function headers are also considered declarations in this
1714 sense.
1716 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1717 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1718 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1719 will be handled."
1720 t nil
1725 ;; FIXME: Some of those below ought to be on `c-other-decl-kwds' instead.
1729 ;; In CORBA PSDL:
1731 ;; In CORBA CIDL:
1733 ;; Note: "const" is not used in Java, but it's still a reserved keyword.
1740 "Keywords that can start or prefix any declaration level construct,
1741 besides those on `c-class-decl-kwds', `c-brace-list-decl-kwds',
1742 `c-other-block-decl-kwds', `c-typedef-decl-kwds',
1743 `c-typeless-decl-kwds' and `c-modifier-kwds'.
1745 If any of these also are on `c-type-list-kwds', `c-ref-list-kwds',
1746 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1747 `c-<>-type-kwds', or `c-<>-arglist-kwds' then the associated clauses
1748 will be handled."
1749 t nil
1755 "Keywords that always start declarations, wherever they occur.
1756 This can be used for declarations that aren't recognized by the normal
1757 combination of `c-decl-prefix-re' and `c-decl-start-re'."
1758 t nil
1759 ;; Classes can be declared anywhere in a Pike expression.
1763 "Keywords that can occur anywhere in a declaration level construct.
1764 This is used for self-contained things that can be tacked on anywhere
1765 on a declaration and that should be ignored to be able to recognize it
1766 correctly. Typical cases are compiler extensions like
1769 __declspec(noreturn) void foo();
1770 class __declspec(dllexport) classname {...};
1771 void foo() __attribute__((noreturn));
1773 Note that unrecognized plain symbols are skipped anyway if they occur
1774 before the type, so such things are not necessary to mention here.
1775 Mentioning them here is necessary only if they can occur in other
1776 places, or if they are followed by a construct that must be skipped
1778 examples above). In the last case, they alse need to be present on
1779 one of `c-type-list-kwds', `c-ref-list-kwds',
1780 `c-colon-type-list-kwds', `c-paren-nontype-kwds', `c-paren-type-kwds',
1781 `c-<>-type-kwds', or `c-<>-arglist-kwds'."
1782 ;; NB: These are currently not recognized in all parts of a
1783 ;; declaration. Specifically, they aren't recognized in the middle
1784 ;; of multi-token types, inside declarators, and between the
1785 ;; identifier and the arglist paren of a function declaration.
1786 ;;
1787 ;; FIXME: This ought to be user customizable since compiler stuff
1788 ;; like this usually is wrapped in project specific macros. (It'd
1789 ;; of course be even better if we could cope without knowing this.)
1790 t nil
1792 "__attribute__"
1793 ;; MSVC extension.
1797 ;; Adorned regexp matching `c-decl-hangon-kwds'.
1802 ;; All keywords that can occur in the preamble of a declaration.
1803 ;; They typically occur before the type, but they are also matched
1804 ;; after presumptive types since we often can't be sure that
1805 ;; something is a type or just some sort of macro in front of the
1806 ;; declaration. They might be ambiguous with types or type
1807 ;; prefixes.
1820 ;; Adorned regexp of `c-prefix-spec-kwds'.
1825 ;; Adorned regexp of the keywords in `c-prefix-spec-kwds' that aren't
1826 ;; ambiguous with types or type prefixes. These are the keywords (like
1827 ;; extern, namespace, but NOT template) that can modify a declaration.
1836 ;; Keywords that can occur after argument list of a function header
1837 ;; declaration, i.e. in the "K&R region".
1842 ;; Adorned regexp matching all keywords that can't appear at the
1843 ;; start of a declaration.
1853 "Access protection label keywords in classes."
1854 t nil
1859 "Keywords introducing declarations that can contain a block which
1862 like that, it ends with the following ';' and not right away.
1864 The keywords on list are assumed to also be present on one of the
1865 `*-decl-kwds' lists."
1866 t nil
1871 ;; Regexp matching the `c-block-decls-with-vars' keywords, or nil in
1872 ;; languages without such constructs.
1879 "Keywords introducing extra declaration specifiers in the region
1881 declarations."
1882 t nil
1885 "supports"
1886 ;; In CORBA PSDL:
1890 "Keywords that may be followed by a nonsymbol sexp before whatever
1891 construct it's part of continues."
1892 t nil
1896 "Keywords that may be followed by a comma separated list of type
1897 identifiers, where each optionally can be prefixed by keywords. (Can
1898 also be used for the special case when the list can contain only one
1899 element.)
1901 Assumed to be mutually exclusive with `c-ref-list-kwds'. There's no
1902 reason to put keywords on this list if they are on `c-type-prefix-kwds'.
1903 There's also no reason to add keywords that prefixes a normal
1904 declaration consisting of a type followed by a declarator (list), so
1905 the keywords on `c-modifier-kwds' should normally not be listed here
1906 either.
1908 Note: Use `c-typeless-decl-kwds' for keywords followed by a function
1909 or variable identifier (that's being defined)."
1910 t nil
1915 ;; In CORBA PSDL:
1920 "Keywords that may be followed by a comma separated list of
1921 reference (i.e. namespace/scope/module) identifiers, where each
1922 optionally can be prefixed by keywords. (Can also be used for the
1923 special case when the list can contain only one element.) Assumed to
1924 be mutually exclusive with `c-type-list-kwds'.
1926 Note: Use `c-typeless-decl-kwds' for keywords followed by a function
1927 or variable identifier (that's being defined)."
1928 t nil
1932 ;; In CORBA CIDL:
1937 "Keywords that may be followed (not necessarily directly) by a colon
1938 and then a comma separated list of type identifiers, where each
1939 optionally can be prefixed by keywords. (Can also be used for the
1940 special case when the list can contain only one element.)"
1941 t nil
1944 ;; In CORBA PSDL:
1948 "Regexp matched after the keywords in `c-colon-type-list-kwds' to skip
1949 forward to the colon. The end of the match is assumed to be directly
1951 regexp if `c-colon-type-list-kwds' isn't nil."
1953 ;; Disallow various common punctuation chars that can't come
1954 ;; before the ":" that starts the inherit list after "class"
1955 ;; or "struct" in C++. (Also used as default for other
1956 ;; languages.)
1961 "Keywords that may be followed by a parenthesis expression that doesn't
1962 contain type identifiers."
1963 t nil
1965 "__attribute__"
1966 ;; MSVC extension.
1970 "Keywords that may be followed by a parenthesis expression containing
1971 type identifiers separated by arbitrary tokens."
1972 t nil
1984 "Keywords that may be followed by an angle bracket expression
1986 `c-<>-arglist-kwds' is that unknown names are taken to be types and
1987 not other identifiers. `c-recognize-<>-arglists' is assumed to be set
1988 if this isn't nil."
1989 t nil
1992 ;; In CORBA PSDL:
1996 "Keywords that can be followed by a C++ style template arglist; see
1997 `c-recognize-<>-arglists' for details. That language constant is
1998 assumed to be set if this isn't nil."
1999 t nil
2004 ;; All keywords that can be followed by an angle bracket sexp.
2010 ;; Adorned regexp matching keywords that can be followed by an angle
2011 ;; bracket sexp. Always set when `c-recognize-<>-arglists' is.
2017 "Keywords that may be followed by a brace block containing a comma
2018 separated list of identifier definitions, i.e. like the list of
2019 identifiers that follows the type in a normal declaration."
2023 "Statement keywords followed directly by a substatement."
2028 idl nil)
2031 ;; Regexp matching the start of any statement followed directly by a
2032 ;; substatement (doesn't match a bare block, however).
2037 "Statement keywords followed by a paren sexp and then by a substatement."
2042 idl nil
2047 ;; Regexp matching the start of any statement followed by a paren sexp
2048 ;; and then by a substatement.
2053 ;; Union of `c-block-stmt-1-kwds' and `c-block-stmt-2-kwds'.
2059 ;; Regexp matching the start of any statement that has a
2060 ;; substatement (except a bare block). Nil in languages that
2061 ;; don't have such constructs.
2070 "Statement keywords followed by an expression or nothing."
2073 ;; Note: `goto' is not valid in Java, but the keyword is still reserved.
2075 idl nil
2082 ;; Adorned regexp matching `c-simple-stmt-kwds'.
2087 "Statement keywords followed by a parenthesis expression that
2088 nevertheless contains a list separated with ';' and not ','."
2090 idl nil)
2093 ;; Adorned regexp matching `c-paren-stmt-kwds'.
2098 "Statement keywords followed by an assembler expression."
2099 t nil
2103 ;; Regexp matching the start of an assembler statement. Nil in
2104 ;; languages that don't support that.
2110 "Keywords introducing colon terminated labels in blocks."
2112 awk nil)
2115 ;; Adorned regexp matching any keyword that introduces a label.
2120 "Keywords that might be followed by a label identifier."
2124 idl nil
2125 awk nil)
2128 "Keywords for constants."
2129 t nil
2138 "Keywords besides constants and operators that start primary expressions."
2139 t nil
2146 ;; Keywords that can occur anywhere in expressions. Built from
2147 ;; `c-primary-expr-kwds' and all keyword operators in `c-operators'.
2151 t
2156 "Keywords that start lambda constructs, i.e. function definitions in
2157 expressions."
2158 t nil
2162 "Keywords that start constructs followed by statement blocks which can
2163 be used in expressions \(the gcc extension for this in C and C++ is
2164 handled separately by `c-recognize-paren-inexpr-blocks')."
2165 t nil
2169 "Keywords that can start classes inside expressions."
2170 t nil
2175 "Keywords that can start brace list blocks inside expressions.
2176 Note that Java specific rules are currently applied to tell this from
2177 `c-inexpr-class-kwds'."
2178 t nil
2182 ;; Regexp matching the start of a brace list in an expression, or
2183 ;; nil in languages that don't have such things. This should not
2184 ;; match brace lists recognized through `c-special-brace-lists'.
2191 ;; Regexp matching keywords in any construct that contain another
2192 ;; declaration level, i.e. that isn't followed by a function block
2193 ;; or brace list. When the first submatch matches, it's an
2194 ;; unambiguous construct, otherwise it's an ambiguous match that
2195 ;; might also be the return type of a function declaration.
2213 "Keywords that can introduce bitfields."
2214 t nil
2218 ;; Regexp matching the start of a bitfield (not uniquely), or nil in
2219 ;; languages without bitfield support.
2220 t nil
2225 "Keywords not accounted for by any other `*-kwds' language constant."
2226 t nil
2228 ;; In CORBA CIDL: (These are declaration keywords that never
2229 ;; can start a declaration.)
2232 \f
2233 ;;; Constants built from keywords.
2235 ;; Note: No `*-kwds' language constants may be defined below this point.
2239 ;; List of all the language constants that contain keyword lists.
2244 ;; Make the list of globally interned symbols
2245 ;; instead of ones interned in `c-lang-constants'.
2247 c-lang-constants)
2248 list)))
2251 ;; All keywords as a list.
2256 c-kwds-lang-consts)
2257 nil))
2261 ;; All keywords as an adorned regexp.
2266 ;; An alist with all the keywords in the cars. The cdr for each
2267 ;; keyword is a list of the symbols for the `*-kwds' lists that
2268 ;; contains it.
2274 c-kwds-lang-consts))))
2275 lang-const kwd-list kwd
2276 result-alist elem)
2288 result-alist))
2291 ;; An obarray containing all keywords as symbols. The property list
2292 ;; of each symbol has a non-nil entry for the specific `*-kwds'
2293 ;; lists it's a member of.
2294 ;;
2295 ;; E.g. to see whether the string str contains a keyword on
2296 ;; `c-class-decl-kwds', one can do like this:
2297 ;; (get (intern-soft str c-keyword-obarray) 'c-class-decl-kwds)
2298 ;; Which preferably is written using the associated functions in
2299 ;; cc-engine:
2300 ;; (c-keyword-member (c-keyword-sym str) 'c-class-decl-kwds)
2302 ;; The obarray is not stored directly as a language constant since
2303 ;; the printed representation for obarrays used in .elc files isn't
2304 ;; complete.
2307 kwd lang-const-list
2314 ;; Emacs has an odd bug that causes `mapcan' to fail
2315 ;; with unintelligible errors. (XEmacs works.)
2316 ;;(mapcan (lambda (lang-const)
2317 ;; (list lang-const t))
2318 ;; lang-const-list)
2321 lang-const-list))))
2322 obarray))
2325 ;; Adorned regexp matching all keywords that should be fontified
2326 ;; with the keywords face. I.e. that aren't types or constants.
2336 ;; Regexp matching the start of any primary expression, i.e. any
2337 ;; literal, symbol, prefix operator, and '('. It doesn't need to
2338 ;; exclude keywords; they are excluded afterwards unless the second
2339 ;; submatch matches. If the first but not the second submatch
2340 ;; matches then it is an ambiguous primary expression; it could also
2341 ;; be a match of e.g. an infix operator. (The case with ambiguous
2342 ;; keyword operators isn't handled.)
2348 ;; Filter out the special case prefix
2349 ;; operators that are close parens.
2354 t
2360 postfix-if-paren
2361 left-assoc
2362 right-assoc
2363 right-assoc-sequence)
2364 t))
2367 in-or-postfix-ops
2370 in-or-postfix-ops
2375 ;; Take out all symbol class operators from `prefix-ops' and make the
2376 ;; first submatch from them together with `c-primary-expr-kwds'.
2383 ;; Match all ambiguous operators.
2390 ;; Now match all other symbols.
2394 ;; The chars that can start integer and floating point
2395 ;; constants.
2399 ;; The nonambiguous operators from `prefix-ops'.
2405 ;; Match string and character literals.
2412 \f
2413 ;;; Additional constants for parser-level constructs.
2416 "Regexp matching something that might precede a declaration, cast or
2417 label, such as the last token of a preceding statement or declaration.
2418 This is used in the common situation where a declaration or cast
2419 doesn't start with any specific token that can be searched for.
2421 The regexp should not match bob; that is done implicitly. It can't
2422 require a match longer than one token. The end of the token is taken
2423 to be at the end of the first submatch, which is assumed to always
2424 match. It's undefined whether identifier syntax (see
2425 `c-identifier-syntax-table') is in effect or not. This regexp is
2426 assumed to be a superset of `c-label-prefix-re' if
2427 `c-recognize-colon-labels' is set.
2429 Besides this, `c-decl-start-kwds' is used to find declarations.
2431 Note: This variable together with `c-decl-start-re' and
2433 declaration/cast/label starts. I.e. they might produce more matches
2434 but should not miss anything (or else it's necessary to use text
2435 properties - see the next note). Wherever they match, the following
2436 construct is analyzed to see if it indeed is a declaration, cast or
2437 label. That analysis is not cheap, so it's important that not too
2438 many false matches are triggered.
2440 Note: If a declaration/cast/label start can't be detected with this
2441 variable, it's necessary to use the `c-type' text property with the
2442 value `c-decl-end' on the last char of the last token preceding the
2443 declaration. See the comment blurb at the start of cc-engine.el for
2444 more info."
2446 ;; We match a sequence of characters to skip over things like \"};\"
2447 ;; more quickly. We match ")" in C for K&R region declarations, and
2448 ;; in all languages except Java for when a cpp macro definition
2449 ;; begins with a declaration.
2452 ;; Match "<" in C++ to get the first argument in a template arglist.
2453 ;; In that case there's an additional check in `c-find-decl-spots'
2454 ;; that it got open paren syntax.
2456 ;; Additionally match the protection directives in Objective-C.
2457 ;; Note that this doesn't cope with the longer directives, which we
2458 ;; would have to match from start to end since they don't end with
2459 ;; any easily recognized characters.
2463 ;; Pike is like C but we also match "[" for multiple value
2464 ;; assignments and type casts.
2470 "Regexp matching the start of any declaration, cast or label.
2471 It's used on the token after the one `c-decl-prefix-re' matched. This
2472 regexp should not try to match those constructs accurately as it's
2473 only used as a sieve to avoid spending more time checking other
2474 constructs."
2479 ;; Regexp matching something that might precede or start a
2480 ;; declaration, cast or label.
2481 ;;
2482 ;; If the first submatch matches, it's taken to match the end of a
2483 ;; token that might precede such a construct, e.g. ';', '}' or '{'.
2484 ;; It's built from `c-decl-prefix-re'.
2485 ;;
2486 ;; If the first submatch did not match, the match of the whole
2487 ;; regexp is taken to be at the first token in the declaration.
2488 ;; `c-decl-start-re' is not checked in this case.
2489 ;;
2490 ;; Design note: The reason the same regexp is used to match both
2491 ;; tokens that precede declarations and start them is to avoid an
2492 ;; extra regexp search from the previous declaration spot in
2493 ;; `c-find-decl-spots'. Users of `c-find-decl-spots' also count on
2494 ;; that it finds all declaration/cast/label starts in approximately
2495 ;; linear order, so we can't do the searches in two separate passes.
2506 ;; List containing the paren characters that can open a cast, or nil in
2507 ;; languages without casts.
2515 "List of syntactically relevant characters that never can occur before
2516 the open brace in any construct that contains a brace block, e.g. in
2519 class Foo: public Bar {int x();} a, *b;
2521 If parens can occur, the chars inside those aren't filtered with this
2522 list.
2524 '<' and '>' should be disallowed even if angle bracket arglists can
2525 occur. That since the search function needs to stop at them anyway to
2526 ensure they are given paren syntax.
2528 This is used to skip backward from the open brace to find the region
2531 possible for good performance."
2533 ;; Default to all chars that only occurs in nonsymbol tokens outside
2534 ;; identifiers.
2540 t
2541 t
2545 op pos)
2548 res))))
2550 ;; Allow cpp operatios (where applicable).
2556 ;; Allow ':' for inherit list starters.
2560 ;; Allow ',' for multiple inherits.
2564 ;; Allow parentheses for anonymous inner classes in Java and class
2565 ;; initializer lists in Pike.
2569 ;; Allow '"' for extern clauses (e.g. extern "C" {...}).
2574 ;; `c-block-prefix-disallowed-chars' as an inverted charset suitable
2575 ;; for `c-syntactic-skip-backward'.
2580 "Regexp matching the declarator operators that might precede the
2583 The end of the first submatch is taken as the end of the operator.
2584 Identifier syntax is in effect when this is matched \(see
2585 `c-identifier-syntax-table')."
2588 ;; Default to a regexp that never matches.
2590 ;; Check that there's no "=" afterwards to avoid matching tokens
2591 ;; like "*=".
2593 "[*\(]"
2599 "[*\(&]"
2602 ;; If this matches there's special treatment in
2603 ;; `c-font-lock-declarators' and
2604 ;; `c-font-lock-declarations' that check for a
2605 ;; complete name followed by ":: *".
2617 "Regexp matching the declarator operators that might follow after the
2620 it matches an open paren of some kind, the type declaration check
2621 continues at the corresponding close paren, otherwise the end of the
2622 first submatch is taken as the end of the operator. Identifier syntax
2623 is in effect when this is matched (see `c-identifier-syntax-table')."
2624 ;; Default to a regexp that matches `c-type-modifier-kwds' and a
2625 ;; function argument list parenthesis.
2633 "[\)\[\(]"
2637 ;; "throw" in `c-type-modifier-kwds' is followed
2638 ;; by a parenthesis list, but no extra measures
2639 ;; are necessary to handle that.
2649 "This regexp is matched after a declarator expression where
2650 `c-type-decl-suffix-key' has matched. If it matches then the
2651 construct is taken as a declaration. It's typically used to match the
2652 beginning of a function body or whatever might occur after the
2653 function header in a function declaration or definition. It's
2654 undefined whether identifier syntax (see `c-identifier-syntax-table')
2655 is in effect or not.
2660 t "{"
2661 ;; If K&R style declarations should be recognized then one could
2662 ;; consider to match the start of any symbol since we want to match
2663 ;; the start of the first declaration in the "K&R region". That
2664 ;; could however produce false matches on code like "FOO(bar) x"
2665 ;; where FOO is a cpp macro, so it's better to leave it out and rely
2666 ;; on the other heuristics in that case.
2668 ;; Add on the keywords in `c-postfix-spec-kwds'.
2673 ;; Also match the colon that starts a base class initializer list in
2674 ;; C++. That can be confused with a function call before the colon
2675 ;; in a ? : operator, but we count on that `c-decl-prefix-re' won't
2676 ;; match before such a thing (as a declaration-level construct;
2677 ;; matches inside arglist contexts are already excluded).
2684 ;; Regexp that in addition to `c-after-suffixed-type-decl-key'
2685 ;; matches ";" and ",".
2694 end of the operator. nil in languages without such operators. It's
2695 undefined whether identifier syntax (see `c-identifier-syntax-table')
2696 is in effect or not."
2697 t nil
2703 "Regexp matching operators that might follow after a type, or nil in
2704 languages that don't have such operators. The end of the first
2705 submatch is taken as the end of the operator. This should not match
2706 things like C++ template arglists if `c-recognize-<>-arglists' is set.
2707 It's undefined whether identifier syntax (see `c-identifier-syntax-table')
2708 is in effect or not."
2709 t nil
2715 ;; Regexp matching the known type identifiers. This is initialized
2716 ;; from the type keywords and `*-font-lock-extra-types'. The first
2717 ;; submatch is the one that matches the type. Note that this regexp
2718 ;; assumes that symbol constituents like '_' and '$' have word
2719 ;; syntax.
2728 extra-types)))
2734 extra-types))))
2739 plain-strings)))
2740 regexp-strings)
2745 "List of open- and close-chars that makes up a pike-style brace list,
2747 list."
2748 t nil
2753 "Non-nil means K&R style argument declarations are valid."
2754 t nil
2755 c t)
2759 "Non-nil means function declarations without return type should be
2760 recognized. That can introduce an ambiguity with parenthesized macro
2761 calls before a brace block. This setting does not affect declarations
2762 that are preceded by a declaration starting keyword, so
2763 e.g. `c-typeless-decl-kwds' may still be used when it's set to nil."
2764 t nil
2770 "Non-nil means C++ style template arglists should be handled. More
2771 specifically, this means a comma separated list of types or
2773 identifier or one of the keywords on `c-<>-type-kwds' or
2774 `c-<>-arglist-kwds'. If there's an identifier before then the whole
2775 expression is considered to be a type."
2781 "Non-nil means that parenthesis style initializers exist,
2782 i.e. constructs like
2784 Foo bar (gnu);
2786 in addition to the more classic
2788 Foo bar = gnu;"
2789 t nil
2794 "Non-nil to recognize gcc style in-expression blocks,
2795 i.e. compound statements surrounded by parentheses inside expressions."
2796 t nil
2802 ;; Regexp matching the start of angle bracket arglists in languages
2803 ;; where `c-recognize-<>-arglists' is set. Does not exclude
2804 ;; keywords outside `c-<>-arglist-kwds'. The first submatch is
2805 ;; assumed to surround the preceding symbol. The whole match is
2806 ;; assumed to end directly after the opening "<".
2816 ;; Regexp that in addition to `c-opt-<>-arglist-start' matches close
2817 ;; parens. The first submatch is assumed to surround
2818 ;; `c-opt-<>-arglist-start'.
2827 ;; Regexp matching the beginning of a declaration specifier in the
2828 ;; region between the header and the body of a declaration.
2829 ;;
2830 ;; TODO: This is currently not used uniformly; c++-mode and
2831 ;; java-mode each have their own ways of using it.
2832 t nil
2845 That includes labels in code and access keys in classes. This does
2846 not apply to labels recognized by `c-label-kwds' and
2847 `c-opt-extra-label-key'."
2848 t nil
2854 "Regexp like `c-decl-prefix-re' that matches any token that can precede
2855 a generic colon label. Not used if `c-recognize-colon-labels' is
2856 nil."
2862 "Regexp matching things that can't occur in generic colon labels,
2863 neither in a statement nor in a declaration context. The regexp is
2864 tested at the beginning of every sexp in a suspected label,
2867 ;; Don't allow string literals.
2869 ;; All keywords except `c-label-kwds' and `c-protection-kwds'.
2875 ;; Also check for open parens in C++, to catch member init lists in
2876 ;; constructors. We normally allow it so that macros with arguments
2877 ;; work in labels.
2882 "Optional regexp matching labels.
2883 Normally, labels are detected according to `c-nonlabel-token-key',
2884 `c-decl-prefix-re' and `c-nonlabel-decl-prefix-re'. This regexp can
2885 be used if there are additional labels that aren't recognized that
2886 way."
2887 t nil
2892 ;; Regexp describing friend declarations classes, or nil in
2893 ;; languages that don't have such things.
2894 ;;
2895 ;; TODO: Ought to use `c-prefix-spec-kwds-re' or similar, and the
2896 ;; template skipping isn't done properly. This will disappear soon.
2897 t nil
2902 "<.+>"
2904 "friend"
2909 ;; Special regexp to match the start of Objective-C methods. The
2910 ;; first submatch is assumed to end after the + or - key.
2911 t nil
2913 ;; TODO: Ought to use a better method than anchoring on bol.
2918 "([^\)]*)"
2925 ;; Must be set in buffers where the `c-type' text property might be
2926 ;; used with the value `c-decl-end'.
2927 ;;
2928 ;; `c-decl-end' is used to mark the ends of labels and access keys
2929 ;; to make interactive refontification work better.
2932 ;; `c-decl-end' is used to mark the end of the @-style directives in
2933 ;; Objective-C.
2934 objc t)
2937 \f
2938 ;;; Wrap up the `c-lang-defvar' system.
2940 ;; Compile in the list of language variables that has been collected
2941 ;; with the `c-lang-defvar' and `c-lang-setvar' macros. Note that the
2942 ;; first element of each is nil.
2946 ;; Make the `c-lang-setvar' variables buffer local in the current buffer.
2947 ;; These are typically standard emacs variables such as `comment-start'.
2955 "Create a function that initializes all the language dependent variables
2956 for the given mode.
2958 This function should be evaluated at compile time, so that the
2959 function it returns is byte compiled with all the evaluated results
2960 from the language constants. Use the `c-init-language-vars' macro to
2961 accomplish that conveniently."
2967 ;; No need to byte compile this lambda since the byte compiler is
2968 ;; smart enough to detect the `funcall' construct in the
2969 ;; `c-init-language-vars' macro below and compile it all straight
2970 ;; into the function that contains `c-init-language-vars'.
2973 ;; This let sets up the context for `c-mode-var' and similar
2974 ;; that could be in the result from `cl-macroexpand-all'.
2976 current-var source-eval)
2983 ;; `c-lang-const' will expand to the evaluated
2984 ;; constant immediately in `cl-macroexpand-all'
2985 ;; below.
2991 ;; Note: The following `append' copies the
2992 ;; first argument. That list is small, so
2993 ;; this doesn't matter too much.
2997 ;; This diagnostic message isn't useful for end
2998 ;; users, so it's disabled.
2999 ;;(unless (get ',mode 'c-has-warned-lang-consts)
3000 ;; (message ,(concat "%s compiled with CC Mode %s "
3001 ;; "but loaded with %s - evaluating "
3002 ;; "language constants from source")
3003 ;; ',mode ,c-version c-version)
3004 ;; (put ',mode 'c-has-warned-lang-consts t))
3018 current-var
3021 (fallback source eval - %s compiled with CC Mode %s but loaded with %s)"
3024 err)
3027 ;; Being evaluated from source. Always use the dynamic method to
3028 ;; work well when `c-lang-defvar's in this file are reevaluated
3029 ;; interactively.
3035 current-var)
3047 "Eval error in the `c-lang-defvar' or `c-lang-setver' for `%s' (source eval): %S"
3048 current-var err)
3050 ))
3053 "Initialize all the language dependent variables for the given mode.
3054 This macro is expanded at compile time to a form tailored for the mode
3055 in question, so MODE must be a constant. Therefore MODE is not
3056 evaluated and should not be quoted."
3059 \f
3062 ;; arch-tag: 1ab57482-cfc2-4c5b-b628-3539c3098822
3063 ;;; cc-langs.el ends here