From: Alan Mackenzie Date: Thu, 4 Feb 2010 21:15:37 +0000 (+0000) Subject: Change strategy for marking < and > as template delimiters: mark them X-Git-Tag: emacs-pretest-23.1.93~110 X-Git-Url: http://git.eshelyaron.com/gitweb/?a=commitdiff_plain;h=dd969a56a57ce94dca1b61b25c8cc4ad102112f7;p=emacs.git Change strategy for marking < and > as template delimiters: mark them strictly in matching pairs. --- diff --git a/lisp/ChangeLog b/lisp/ChangeLog index baf62e0b013..0e831a4be72 100644 --- a/lisp/ChangeLog +++ b/lisp/ChangeLog @@ -1,3 +1,30 @@ +2010-02-04 Alan Mackenzie + + Change strategy for marking < and > as template delimiters: mark + them strictly in matching pairs. + + * cc-mode.el (c-before-change): Use + c-get-state-before-change-functions. + (c-common-init): Adapt to use + c-get-state-before-change-functions (note plural). + + * cc-langs.el (c-no-parens-syntax-table): New syntax table, used + for searching syntactically for matching s. + (c-get-state-before-change-functions): New language variable (note + the plural) which supersedes c-get-state-before-change-function. + + * cc-engine.el (c-clear-<-pair-props, c-clear->-pair-props) + (c-clear-<>-pair-props, c-clear-<-pair-props-if-match-after) + (c-clear->-pair-props-if-match-before) + (c-before-change-check-<>-operators): new functions. + (c-after-change-check-<>-operators): Use macro + c-unmark-<->-as-paren. + + * cc-defs.el (c-search-backward-char-property): New macro. + + * cc-cmds.el (c-electric-lt-gt): Do not set text properties on < + and > any more. (These will be handled by font locking.) + 2010-02-04 Michael Albinus * dired.el (dired-revert): If DIRED-DIRECTORY is a cons cell, call diff --git a/lisp/progmodes/cc-cmds.el b/lisp/progmodes/cc-cmds.el index 4eade6edf58..56fc8032541 100644 --- a/lisp/progmodes/cc-cmds.el +++ b/lisp/progmodes/cc-cmds.el @@ -1086,104 +1086,76 @@ numeric argument is supplied, or the point is inside a literal." (interactive "*P") (let ((c-echo-syntactic-information-p nil) - final-pos close-paren-inserted) + final-pos close-paren-inserted found-delim) (self-insert-command (prefix-numeric-value arg)) (setq final-pos (point)) - (c-save-buffer-state (c-parse-and-markup-<>-arglists - c-restricted-<>-arglists - <-pos) +;;;; 2010-01-31: There used to be code here to put a syntax-table text +;;;; property on the new < or > and its mate (if any) when they are template +;;;; parens. This is now done in an after-change function. - (when c-recognize-<>-arglists - (if (eq last-command-event ?<) - (when (and (progn - (backward-char) - (= (point) - (progn - (c-beginning-of-current-token) - (point)))) + ;; Indent the line if appropriate. + (when (and c-electric-flag c-syntactic-indentation c-recognize-<>-arglists) + (setq found-delim + (if (eq last-command-event ?<) + ;; If a <, basically see if it's got "template" before it ..... + (or (and (progn + (backward-char) + (= (point) + (progn (c-beginning-of-current-token) (point)))) + (progn + (c-backward-token-2) + (looking-at c-opt-<>-sexp-key))) + ;; ..... or is a C++ << operator. + (and (c-major-mode-is 'c++-mode) + (progn + (goto-char (1- final-pos)) + (c-beginning-of-current-token) + (looking-at "<<")) + (>= (match-end 0) final-pos))) + + ;; It's a >. Either a C++ >> operator. ...... + (or (and (c-major-mode-is 'c++-mode) (progn - (c-backward-token-2) - (looking-at c-opt-<>-sexp-key))) - (c-mark-<-as-paren (1- final-pos))) - - ;; It's a ">". Check if there's an earlier "<" which either has - ;; open paren syntax already or that can be recognized as an arglist - ;; together with this ">". Note that this won't work in cases like - ;; "template " but they ought to be rare. - - (save-restriction - ;; Narrow to avoid that `c-forward-<>-arglist' below searches past - ;; our position. - (narrow-to-region (point-min) final-pos) - - (while (and - (progn - (goto-char final-pos) - (c-syntactic-skip-backward "^<;}" nil t) - (eq (char-before) ?<)) - (progn - (backward-char) - ;; If the "<" already got open paren syntax we know we - ;; have the matching closer. Handle it and exit the - ;; loop. - (if (looking-at "\\s\(") - (progn - (c-mark->-as-paren (1- final-pos)) - (setq close-paren-inserted t) - nil) - t)) + (goto-char (1- final-pos)) + (c-beginning-of-current-token) + (looking-at ">>")) + (>= (match-end 0) final-pos)) + ;; ...., or search back for a < which isn't already marked as an + ;; opening template delimiter. + (save-restriction + (widen) + ;; Narrow to avoid `c-forward-<>-arglist' below searching past + ;; our position. + (narrow-to-region (point-min) final-pos) + (goto-char final-pos) + (while + (and + (progn + (c-syntactic-skip-backward "^<;}" nil t) + (eq (char-before) ?<)) + (progn + (backward-char) + (looking-at "\\s\(")))) + (and (eq (char-after) ?<) + (not (looking-at "\\s\(")) + (progn (c-backward-syntactic-ws) + (c-simple-skip-symbol-backward)) + (or (looking-at c-opt-<>-sexp-key) + (not (looking-at c-keywords-regexp))))))))) - (progn - (setq <-pos (point)) - (c-backward-syntactic-ws) - (c-simple-skip-symbol-backward)) - (or (looking-at c-opt-<>-sexp-key) - (not (looking-at c-keywords-regexp))) - - (let ((c-parse-and-markup-<>-arglists t) - c-restricted-<>-arglists - (containing-sexp - (c-most-enclosing-brace (c-parse-state)))) - (when (and containing-sexp - (progn (goto-char containing-sexp) - (eq (char-after) ?\()) - (not (eq (get-text-property (point) 'c-type) - 'c-decl-arg-start))) - (setq c-restricted-<>-arglists t)) - (goto-char <-pos) - (c-forward-<>-arglist nil)) - - ;; Loop here if the "<" we found above belongs to a nested - ;; angle bracket sexp. When we start over we'll find the - ;; previous or surrounding sexp. - (if (< (point) final-pos) - t - (setq close-paren-inserted t) - nil))))))) (goto-char final-pos) - - ;; Indent the line if appropriate. - (when (and c-electric-flag c-syntactic-indentation) - (backward-char) - (when (prog1 (or (looking-at "\\s\(\\|\\s\)") - (and (c-major-mode-is 'c++-mode) - (progn - (c-beginning-of-current-token) - (looking-at "<<\\|>>")) - (= (match-end 0) final-pos))) - (goto-char final-pos)) - (indent-according-to-mode))) - - (when (and close-paren-inserted - (not executing-kbd-macro) - blink-paren-function) - ;; Note: Most paren blink functions, such as the standard - ;; `blink-matching-open', currently doesn't handle paren chars - ;; marked with text properties very well. Maybe we should avoid - ;; this call for the time being? - (funcall blink-paren-function)))) + (when found-delim + (indent-according-to-mode) + (when (and (eq (char-before) ?>) + (not executing-kbd-macro) + blink-paren-function) + ;; Note: Most paren blink functions, such as the standard + ;; `blink-matching-open', currently doesn't handle paren chars + ;; marked with text properties very well. Maybe we should avoid + ;; this call for the time being? + (funcall blink-paren-function))))) (defun c-electric-paren (arg) "Insert a parenthesis. diff --git a/lisp/progmodes/cc-defs.el b/lisp/progmodes/cc-defs.el index bb91dee6ce8..7eb0016ff43 100644 --- a/lisp/progmodes/cc-defs.el +++ b/lisp/progmodes/cc-defs.el @@ -1029,6 +1029,44 @@ MODE is either a mode symbol or a list of mode symbols." ;; Emacs. `(remove-text-properties ,from ,to '(,property nil)))) +(defmacro c-search-forward-char-property (property value &optional limit) + "Search forward for a text-property PROPERTY having value VALUE. +LIMIT bounds the search. The comparison is done with `equal'. + +Leave point just after the character, and set the match data on +this character, and return point. If VALUE isn't found, Return +nil; point is then left undefined." + `(let ((place (point))) + (while + (and + (< place ,(or limit '(point-max))) + (not (equal (get-text-property place ,property) ,value))) + (setq place (next-single-property-change + place ,property nil ,(or limit '(point-max))))) + (when (< place ,(or limit '(point-max))) + (goto-char place) + (search-forward-regexp ".") ; to set the match-data. + (point)))) + +(defmacro c-search-backward-char-property (property value &optional limit) + "Search backward for a text-property PROPERTY having value VALUE. +LIMIT bounds the search. The comparison is done with `equal'. + +Leave point just before the character, set the match data on this +character, and return point. If VALUE isn't found, Return nil; +point is then left undefined." + `(let ((place (point))) + (while + (and + (> place ,(or limit '(point-min))) + (not (equal (get-text-property (1- place) ,property) ,value))) + (setq place (previous-single-property-change + place ,property nil ,(or limit '(point-min))))) + (when (> place ,(or limit '(point-max))) + (goto-char place) + (search-backward-regexp ".") ; to set the match-data. + (point)))) + (defun c-clear-char-property-with-value-function (from to property value) "Remove all text-properties PROPERTY from the region (FROM, TO) which have the value VALUE, as tested by `equal'. These diff --git a/lisp/progmodes/cc-engine.el b/lisp/progmodes/cc-engine.el index 9d0af1d53ce..3b7f6e4ea0a 100644 --- a/lisp/progmodes/cc-engine.el +++ b/lisp/progmodes/cc-engine.el @@ -4880,7 +4880,168 @@ comment at the start of cc-engine.el for more info." ))) -;; Handling of small scale constructs like types and names. +;; Setting and removing syntax properties on < and > in languages (C++ +;; and Java) where they can be template/generic delimiters as well as +;; their normal meaning of "less/greater than". + +;; Normally, < and > have syntax 'punctuation'. When they are found to +;; be delimiters, they are marked as such with the category properties +;; c-<-as-paren-syntax, c->-as-paren-syntax respectively. + +;; STRATEGY: +;; +;; It is impossible to determine with certainty whether a <..> pair in +;; C++ is two comparison operators or is template delimiters, unless +;; one duplicates a lot of a C++ compiler. For example, the following +;; code fragment: +;; +;; foo (a < b, c > d) ; +;; +;; could be a function call with two integer parameters (each a +;; relational expression), or it could be a constructor for class foo +;; taking one parameter d of templated type "a < b, c >". They are +;; somewhat easier to distinguish in Java. +;; +;; The strategy now (2010-01) adopted is to mark and unmark < and +;; > IN MATCHING PAIRS ONLY. [Previously, they were marked +;; individually when their context so indicated. This gave rise to +;; intractible problems when one of a matching pair was deleted, or +;; pulled into a literal.] +;; +;; At each buffer change, the syntax-table properties are removed in a +;; before-change function and reapplied, when needed, in an +;; after-change function. It is far more important that the +;; properties get removed when they they are spurious than that they +;; be present when wanted. +;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; +(defun c-clear-<-pair-props (&optional pos) + ;; POS (default point) is at a < character. If it is marked with + ;; open paren syntax-table text property, remove the property, + ;; together with the close paren property on the matching > (if + ;; any). + (save-excursion + (if pos + (goto-char pos) + (setq pos (point))) + (when (equal (c-get-char-property (point) 'syntax-table) + c-<-as-paren-syntax) + (with-syntax-table c-no-parens-syntax-table ; ignore unbalanced [,{,(,.. + (c-go-list-forward)) + (when (equal (c-get-char-property (1- (point)) 'syntax-table) + c->-as-paren-syntax) ; should always be true. + (c-clear-char-property (1- (point)) 'syntax-table)) + (c-clear-char-property pos 'syntax-table)))) + +(defun c-clear->-pair-props (&optional pos) + ;; POS (default point) is at a > character. If it is marked with + ;; close paren syntax-table property, remove the property, together + ;; with the open paren property on the matching < (if any). + (save-excursion + (if pos + (goto-char pos) + (setq pos (point))) + (when (equal (c-get-char-property (point) 'syntax-table) + c->-as-paren-syntax) + (with-syntax-table c-no-parens-syntax-table ; ignore unbalanced [,{,(,.. + (c-go-up-list-backward)) + (when (equal (c-get-char-property (point) 'syntax-table) + c-<-as-paren-syntax) ; should always be true. + (c-clear-char-property (point) 'syntax-table)) + (c-clear-char-property pos 'syntax-table)))) + +(defun c-clear-<>-pair-props (&optional pos) + ;; POS (default point) is at a < or > character. If it has an + ;; open/close paren syntax-table property, remove this property both + ;; from the current character and its partner (which will also be + ;; thusly marked). + (cond + ((eq (char-after) ?\<) + (c-clear-<-pair-props pos)) + ((eq (char-after) ?\>) + (c-clear->-pair-props pos)) + (t (c-benign-error + "c-clear-<>-pair-props called from wrong position")))) + +(defun c-clear-<-pair-props-if-match-after (lim &optional pos) + ;; POS (default point) is at a < character. If it is both marked + ;; with open/close paren syntax-table property, and has a matching > + ;; (also marked) which is after LIM, remove the property both from + ;; the current > and its partner. + (save-excursion + (if pos + (goto-char pos) + (setq pos (point))) + (when (equal (c-get-char-property (point) 'syntax-table) + c-<-as-paren-syntax) + (with-syntax-table c-no-parens-syntax-table ; ignore unbalanced [,{,(,.. + (c-go-list-forward)) + (when (and (>= (point) lim) + (equal (c-get-char-property (1- (point)) 'syntax-table) + c->-as-paren-syntax)) ; should always be true. + (c-unmark-<->-as-paren (1- (point))) + (c-unmark-<->-as-paren pos))))) + +(defun c-clear->-pair-props-if-match-before (lim &optional pos) + ;; POS (default point) is at a > character. If it is both marked + ;; with open/close paren syntax-table property, and has a matching < + ;; (also marked) which is before LIM, remove the property both from + ;; the current < and its partner. + (save-excursion + (if pos + (goto-char pos) + (setq pos (point))) + (when (equal (c-get-char-property (point) 'syntax-table) + c->-as-paren-syntax) + (with-syntax-table c-no-parens-syntax-table ; ignore unbalanced [,{,(,.. + (c-go-up-list-backward)) + (when (and (<= (point) lim) + (equal (c-get-char-property (point) 'syntax-table) + c-<-as-paren-syntax)) ; should always be true. + (c-unmark-<->-as-paren (point)) + (c-unmark-<->-as-paren pos))))) + +(defun c-before-change-check-<>-operators (beg end) + ;; Unmark certain pairs of "< .... >" which are currently marked as + ;; template/generic delimiters. (This marking is via syntax-table + ;; text properties). + ;; + ;; These pairs are those which are in the current "statement" (i.e., + ;; the region between the {, }, or ; before BEG and the one after + ;; END), and which enclose any part of the interval (BEG END). + ;; + ;; Note that in C++ (?and Java), template/generic parens cannot + ;; enclose a brace or semicolon, so we use these as bounds on the + ;; region we must work on. + ;; + ;; This function is called from before-change-functions (via + ;; c-get-state-before-change-functions). Thus the buffer is widened, + ;; and point is undefined, both at entry and exit. + ;; + ;; FIXME!!! This routine ignores the possibility of macros entirely. + ;; 2010-01-29. + (save-excursion + (let ((beg-lit-limits (progn (goto-char beg) (c-literal-limits))) + (end-lit-limits (progn (goto-char end) (c-literal-limits)))) + ;; Locate the barrier before the changed region + (goto-char (if beg-lit-limits (car beg-lit-limits) beg)) + (c-syntactic-skip-backward "^;{}" (max (- beg 2048) (point-min))) + + ;; Remove the syntax-table properties from each pertinent <...> pair. + ;; Firsly, the ones with the < before beg and > after beg. + (while (c-search-forward-char-property 'category 'c-<-as-paren-syntax beg) + (c-clear-<-pair-props-if-match-after beg (1- (point)))) + + ;; Locate the barrier after END. + (goto-char (if end-lit-limits (cdr end-lit-limits) end)) + (c-syntactic-re-search-forward "[;{}]" + (min (+ end 2048) (point-max)) 'end) + + ;; Remove syntax-table properties from the remaining pertinent <...> + ;; pairs, those with a > after end and < before end. + (while (c-search-backward-char-property 'category 'c->-as-paren-syntax end) + (c-clear->-pair-props-if-match-before end))))) + + (defun c-after-change-check-<>-operators (beg end) ;; This is called from `after-change-functions' when @@ -4902,7 +5063,7 @@ comment at the start of cc-engine.el for more info." (< beg (setq beg (match-end 0)))) (while (progn (skip-chars-forward "^<>" beg) (< (point) beg)) - (c-clear-char-property (point) 'syntax-table) + (c-clear-<>-pair-props) (forward-char)))) (when (< beg end) @@ -4917,9 +5078,13 @@ comment at the start of cc-engine.el for more info." (< end (setq end (match-end 0)))) (while (progn (skip-chars-forward "^<>" end) (< (point) end)) - (c-clear-char-property (point) 'syntax-table) + (c-clear-<>-pair-props) (forward-char))))))) + + +;; Handling of small scale constructs like types and names. + ;; Dynamically bound variable that instructs `c-forward-type' to also ;; treat possible types (i.e. those that it normally returns 'maybe or ;; 'found for) as actual types (and always return 'found for them). diff --git a/lisp/progmodes/cc-langs.el b/lisp/progmodes/cc-langs.el index e4c61b32a25..d3669f259de 100644 --- a/lisp/progmodes/cc-langs.el +++ b/lisp/progmodes/cc-langs.el @@ -391,6 +391,27 @@ The syntax tables aren't stored directly since they're quite large." (and (c-lang-const c++-make-template-syntax-table) (funcall (c-lang-const c++-make-template-syntax-table)))) +(c-lang-defconst c-no-parens-syntax-table + ;; A variant of the standard syntax table which is used to find matching + ;; "<"s and ">"s which have been marked as parens using syntax table + ;; properties. The other paren characters (e.g. "{", ")" "]") are given a + ;; non-paren syntax here. so that the list commands will work on "< ... >" + ;; even when there's unbalanced other parens inside them. + ;; + ;; This variable is nil for languages which don't have template stuff. + t `(lambda () + (if (c-lang-const c-recognize-<>-arglists) + (let ((table (funcall ,(c-lang-const c-make-mode-syntax-table)))) + (modify-syntax-entry ?\( "." table) + (modify-syntax-entry ?\) "." table) + (modify-syntax-entry ?\[ "." table) + (modify-syntax-entry ?\] "." table) + (modify-syntax-entry ?\{ "." table) + (modify-syntax-entry ?\} "." table) + table)))) +(c-lang-defvar c-no-parens-syntax-table + (funcall (c-lang-const c-no-parens-syntax-table))) + (c-lang-defconst c-identifier-syntax-modifications "A list that describes the modifications that should be done to the mode syntax table to get a syntax table that matches all identifiers @@ -423,26 +444,36 @@ the new syntax, as accepted by `modify-syntax-entry'." classifies symbol constituents like '_' and '$' as word constituents, so that all identifiers are recognized as words.") -(c-lang-defconst c-get-state-before-change-function - "If non-nil, a function called from c-before-change-hook. -Typically it will record enough state to allow +(c-lang-defconst c-get-state-before-change-functions + ;; For documentation see the following c-lang-defvar of the same name. + ;; The value here may be a list of functions or a single function. + t nil + c++ '(c-extend-region-for-CPP c-before-change-check-<>-operators) + (c objc) 'c-extend-region-for-CPP + ;; java 'c-before-change-check-<>-operators + awk 'c-awk-record-region-clear-NL) +(c-lang-defvar c-get-state-before-change-functions + (let ((fs (c-lang-const c-get-state-before-change-functions))) + (if (listp fs) + fs + (list fs))) + "If non-nil, a list of functions called from c-before-change-hook. +Typically these will record enough state to allow `c-before-font-lock-function' to extend the region to fontify, and may do such things as removing text-properties which must be recalculated. -It takes 2 parameters, the BEG and END supplied to every +These functions will be run in the order given. Each of them +takes 2 parameters, the BEG and END supplied to every before-change function; on entry, the buffer will have been widened and match-data will have been saved; point is undefined on both entry and exit; the return value is ignored. -When the mode is initialized, this function is called with -parameters \(point-min) and \(point-max)." - t nil - (c c++ objc) 'c-extend-region-for-CPP - awk 'c-awk-record-region-clear-NL) -(c-lang-defvar c-get-state-before-change-function - (c-lang-const c-get-state-before-change-function)) - +The functions are called even when font locking isn't enabled. + +When the mode is initialized, the functions are called with +parameters \(point-min) and \(point-max).") + (c-lang-defconst c-before-font-lock-function "If non-nil, a function called just before font locking. Typically it will extend the region about to be fontified \(see diff --git a/lisp/progmodes/cc-mode.el b/lisp/progmodes/cc-mode.el index 91fc29cdfa8..9cdc0f97b50 100644 --- a/lisp/progmodes/cc-mode.el +++ b/lisp/progmodes/cc-mode.el @@ -641,8 +641,10 @@ compatible with old code; callers should always specify it." (save-restriction (widen) (save-excursion - (if c-get-state-before-change-function - (funcall c-get-state-before-change-function (point-min) (point-max))) + (if c-get-state-before-change-functions + (mapc (lambda (fn) + (funcall fn beg end)) + c-get-state-before-change-functions)) (if c-before-font-lock-function (funcall c-before-font-lock-function (point-min) (point-max) (- (point-max) (point-min)))))) @@ -775,7 +777,7 @@ Note that the style variables are always made local to the buffer." (defmacro c-run-mode-hooks (&rest hooks) ;; Emacs 21.1 has introduced a system with delayed mode hooks that - ;; require the use of the new function `run-mode-hooks'. + ;; requires the use of the new function `run-mode-hooks'. (if (cc-bytecomp-fboundp 'run-mode-hooks) `(run-mode-hooks ,@hooks) `(progn ,@(mapcar (lambda (hook) `(run-hooks ,hook)) hooks)))) @@ -808,8 +810,8 @@ Note that the style variables are always made local to the buffer." ;; has already been widened, and match-data saved. The return value is ;; meaningless. ;; - ;; This function is the C/C++/ObjC value of - ;; `c-get-state-before-change-function' and is called exclusively as a + ;; This function is in the C/C++/ObjC values of + ;; `c-get-state-before-change-functions' and is called exclusively as a ;; before change function. (goto-char beg) (c-beginning-of-macro) @@ -922,8 +924,8 @@ Note that the style variables are always made local to the buffer." ))))) (defun c-before-change (beg end) - ;; Function to be put on `before-change-function'. Primarily, this calls - ;; the language dependent `c-get-state-before-change-function'. It is + ;; Function to be put on `before-change-functions'. Primarily, this calls + ;; the language dependent `c-get-state-before-change-functions'. It is ;; otherwise used only to remove stale entries from the `c-found-types' ;; cache, and to record entries which a `c-after-change' function might ;; confirm as stale. @@ -1001,8 +1003,10 @@ Note that the style variables are always made local to the buffer." ;; larger than (beg end). (setq c-new-BEG beg c-new-END end) - (if c-get-state-before-change-function - (funcall c-get-state-before-change-function beg end)) + (if c-get-state-before-change-functions + (mapc (lambda (fn) + (funcall fn beg end)) + c-get-state-before-change-functions)) )))) (defun c-after-change (beg end old-len)