Mercurial > emacs
annotate src/region-cache.h @ 106395:f2b36fb84bf7
Enhance `c-parse-state' to run efficiently in "brace desserts".
* progmodes/cc-mode.el (c-basic-common-init): Call
c-state-cache-init.
(c-neutralize-syntax-in-and-mark-CPP): Renamed from
c-extend-and-neutralize-syntax-in-CPP. Mark each CPP construct by
placing `category' properties value 'c-cpp-delimiter at its
boundaries.
* progmodes/cc-langs.el (c-before-font-lock-function):
c-extend-and-neutralize-syntax-in-CPP has been renamed
c-neutralize-syntax-in-and-mark-CPP.
* progmodes/cc-fonts.el (c-cpp-matchers): Mark template brackets
with `category' properties now, not `syntax-table' ones.
* progmodes/cc-engine.el (c-syntactic-end-of-macro): A new
enhanced (but slower) version of c-end-of-macro that won't land
inside a literal or on another awkward character.
(c-state-cache-too-far, c-state-cache-start)
(c-state-nonlit-pos-interval, c-state-nonlit-pos-cache)
(c-state-nonlit-pos-cache-limit, c-state-point-min)
(c-state-point-min-lit-type, c-state-point-min-lit-start)
(c-state-min-scan-pos, c-state-brace-pair-desert)
(c-state-old-cpp-beg, c-state-old-cpp-end): New constants and
buffer local variables.
(c-state-literal-at, c-state-lit-beg)
(c-state-cache-non-literal-place, c-state-get-min-scan-pos)
(c-state-mark-point-min-literal, c-state-cache-top-lparen)
(c-state-cache-top-paren, c-state-cache-after-top-paren)
(c-get-cache-scan-pos, c-get-fallback-scan-pos)
(c-state-balance-parens-backwards, c-parse-state-get-strategy)
(c-renarrow-state-cache)
(c-append-lower-brace-pair-to-state-cache)
(c-state-push-any-brace-pair, c-append-to-state-cache)
(c-remove-stale-state-cache)
(c-remove-stale-state-cache-backwards, c-state-cache-init)
(c-invalidate-state-cache-1, c-parse-state-1)
(c-invalidate-state-cache): New defuns/defmacros/defsubsts.
(c-parse-state): Enhanced and refactored.
(c-debug-parse-state): Amended to deal with all the new variables.
* progmodes/cc-defs.el (c-<-as-paren-syntax, c-mark-<-as-paren)
(c->-as-paren-syntax, c-mark->-as-paren, c-unmark-<->-as-paren):
modify to use category text properties rather than syntax-table
ones.
(c-suppress-<->-as-parens, c-restore-<->-as-parens): new defsubsts
to switch off/on the syntactic paren property of C++ template
delimiters using the category property.
(c-with-<->-as-parens-suppressed): Macro to invoke code with
template delims suppressed.
(c-cpp-delimiter, c-set-cpp-delimiters, c-clear-cpp-delimiters):
New constant/macros which apply category properties to the start
and end of preprocessor constructs.
(c-comment-out-cpps, c-uncomment-out-cpps): defsubsts which
"comment out" the syntactic value of characters in preprocessor
constructs.
(c-with-cpps-commented-out)
(c-with-all-but-one-cpps-commented-out): Macros to invoke code
with characters in all or all but one preprocessor constructs
"commented out".
author | Alan Mackenzie <acm@muc.de> |
---|---|
date | Thu, 03 Dec 2009 16:02:10 +0000 |
parents | e038c1a8307c |
children | 1d1d5d9bd884 |
rev | line source |
---|---|
11047 | 1 /* Header file: Caching facts about regions of the buffer, for optimization. |
75227
e90d04cd455a
Update copyright for years from Emacs 21 to present (mainly adding
Glenn Morris <rgm@gnu.org>
parents:
68651
diff
changeset
|
2 Copyright (C) 1985, 1986, 1993, 1995, 2001, 2002, 2003, 2004, |
100951 | 3 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc. |
11047 | 4 |
5 This file is part of GNU Emacs. | |
6 | |
94994
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
79759
diff
changeset
|
7 GNU Emacs is free software: you can redistribute it and/or modify |
11047 | 8 it under the terms of the GNU General Public License as published by |
94994
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
79759
diff
changeset
|
9 the Free Software Foundation, either version 3 of the License, or |
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
79759
diff
changeset
|
10 (at your option) any later version. |
11047 | 11 |
12 GNU Emacs is distributed in the hope that it will be useful, | |
13 but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 GNU General Public License for more details. | |
16 | |
17 You should have received a copy of the GNU General Public License | |
94994
29adfc9354e7
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
79759
diff
changeset
|
18 along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */ |
11047 | 19 |
20 | |
21 /* This code was written by Jim Blandy <jimb@cs.oberlin.edu> to help | |
22 GNU Emacs better support the gene editor written for the University | |
23 of Illinois at Urbana-Champagne's Ribosome Database Project (RDP). | |
24 | |
25 Emacs implements line operations (finding the beginning/end of the | |
26 line, vertical motion, all the redisplay stuff) by searching for | |
27 newlines in the buffer. Usually, this is a good design; it's very | |
28 clean to just represent the buffer as an unstructured string of | |
29 characters, and the lines in most files are very short (less than | |
30 eighty characters), meaning that scanning usually costs about the | |
31 same as the overhead of maintaining some more complicated data | |
32 structure. | |
33 | |
34 However, some applications, like gene editing, make use of very | |
35 long lines --- on the order of tens of kilobytes. In such cases, | |
36 it may well be worthwhile to try to avoid scanning, because the | |
37 scans have become two orders of magnitude more expensive. It would | |
38 be nice if this speedup could preserve the simplicity of the | |
39 existing data structure, and disturb as little of the existing code | |
40 as possible. | |
41 | |
42 So here's the tack. We add some caching to the scan_buffer | |
43 function, so that when it searches for a newline, it notes that the | |
44 region between the start and end of the search contained no | |
45 newlines; then, the next time around, it consults this cache to see | |
46 if there are regions of text it can skip over completely. The | |
47 buffer modification primitives invalidate this cache. | |
48 | |
49 (Note: Since the redisplay code needs similar information on | |
50 modified regions of the buffer, we can use the code that helps out | |
51 redisplay as a guide to where we need to add our own code to | |
52 invalidate our cache. prepare_to_modify_buffer seems to be the | |
53 central spot.) | |
54 | |
55 Note that the cache code itself never mentions newlines | |
56 specifically, so if you wanted to cache other properties of regions | |
57 of the buffer, you could use this code pretty much unchanged. So | |
58 this cache really holds "known/unknown" information --- "I know | |
59 this region has property P" vs. "I don't know if this region has | |
60 property P or not." */ | |
61 | |
62 | |
63 /* Allocate, initialize and return a new, empty region cache. */ | |
20349 | 64 struct region_cache *new_region_cache P_ ((void)); |
11047 | 65 |
66 /* Free a region cache. */ | |
20349 | 67 void free_region_cache P_ ((struct region_cache *)); |
11047 | 68 |
69 /* Assert that the region of BUF between START and END (absolute | |
70 buffer positions) is "known," for the purposes of CACHE (e.g. "has | |
71 no newlines", in the case of the line cache). */ | |
20349 | 72 extern void know_region_cache P_ ((struct buffer *BUF, |
11047 | 73 struct region_cache *CACHE, |
20349 | 74 int START, int END)); |
11047 | 75 |
76 /* Indicate that a section of BUF has changed, to invalidate CACHE. | |
77 HEAD is the number of chars unchanged at the beginning of the buffer. | |
78 TAIL is the number of chars unchanged at the end of the buffer. | |
79 NOTE: this is *not* the same as the ending position of modified | |
80 region. | |
81 (This way of specifying regions makes more sense than absolute | |
82 buffer positions in the presence of insertions and deletions; the | |
83 args to pass are the same before and after such an operation.) */ | |
20349 | 84 extern void invalidate_region_cache P_ ((struct buffer *BUF, |
85 struct region_cache *CACHE, | |
86 int HEAD, int TAIL)); | |
11047 | 87 |
49600
23a1cea22d13
Trailing whitespace deleted.
Juanma Barranquero <lekktu@gmail.com>
parents:
20349
diff
changeset
|
88 /* The scanning functions. |
11047 | 89 |
90 Basically, if you're scanning forward/backward from position POS, | |
91 and region_cache_forward/backward returns true, you can skip all | |
92 the text between POS and *NEXT. And if the function returns false, | |
93 you should examine all the text from POS to *NEXT, and call | |
94 know_region_cache depending on what you find there; this way, you | |
95 might be able to avoid scanning it again. */ | |
96 | |
97 /* Return true if the text immediately after POS in BUF is known, for | |
49600
23a1cea22d13
Trailing whitespace deleted.
Juanma Barranquero <lekktu@gmail.com>
parents:
20349
diff
changeset
|
98 the purposes of CACHE. If NEXT is non-zero, set *NEXT to the nearest |
11047 | 99 position after POS where the knownness changes. */ |
20349 | 100 extern int region_cache_forward P_ ((struct buffer *BUF, |
11047 | 101 struct region_cache *CACHE, |
102 int POS, | |
20349 | 103 int *NEXT)); |
11047 | 104 |
105 /* Return true if the text immediately before POS in BUF is known, for | |
106 the purposes of CACHE. If NEXT is non-zero, set *NEXT to the nearest | |
107 position before POS where the knownness changes. */ | |
20349 | 108 extern int region_cache_backward P_ ((struct buffer *BUF, |
11047 | 109 struct region_cache *CACHE, |
110 int POS, | |
20349 | 111 int *NEXT)); |
52401 | 112 |
113 /* arch-tag: 70f79125-ef22-4f58-9aec-a48ca2791435 | |
114 (do not change this comment) */ |