Skip to content
This repository
tag: Aquamacs-1.9
Fetching contributors…


Cannot retrieve contributors at this time

file 117 lines (95 sloc) 5.632 kb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
/* Header file: Caching facts about regions of the buffer, for optimization.
Copyright (C) 1985, 1986, 1993, 1995, 2001, 2002, 2003, 2004,
2005, 2006, 2007, 2008 Free Software Foundation, Inc.

This file is part of GNU Emacs.

GNU Emacs is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3, or (at your option)
any later version.

GNU Emacs is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with GNU Emacs; see the file COPYING. If not, write to
the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
Boston, MA 02110-1301, USA. */

/* This code was written by Jim Blandy <> to help
GNU Emacs better support the gene editor written for the University
of Illinois at Urbana-Champagne's Ribosome Database Project (RDP).

Emacs implements line operations (finding the beginning/end of the
line, vertical motion, all the redisplay stuff) by searching for
newlines in the buffer. Usually, this is a good design; it's very
clean to just represent the buffer as an unstructured string of
characters, and the lines in most files are very short (less than
eighty characters), meaning that scanning usually costs about the
same as the overhead of maintaining some more complicated data

However, some applications, like gene editing, make use of very
long lines --- on the order of tens of kilobytes. In such cases,
it may well be worthwhile to try to avoid scanning, because the
scans have become two orders of magnitude more expensive. It would
be nice if this speedup could preserve the simplicity of the
existing data structure, and disturb as little of the existing code
as possible.

So here's the tack. We add some caching to the scan_buffer
function, so that when it searches for a newline, it notes that the
region between the start and end of the search contained no
newlines; then, the next time around, it consults this cache to see
if there are regions of text it can skip over completely. The
buffer modification primitives invalidate this cache.

(Note: Since the redisplay code needs similar information on
modified regions of the buffer, we can use the code that helps out
redisplay as a guide to where we need to add our own code to
invalidate our cache. prepare_to_modify_buffer seems to be the
central spot.)

Note that the cache code itself never mentions newlines
specifically, so if you wanted to cache other properties of regions
of the buffer, you could use this code pretty much unchanged. So
this cache really holds "known/unknown" information --- "I know
this region has property P" vs. "I don't know if this region has
property P or not." */

/* Allocate, initialize and return a new, empty region cache. */
struct region_cache *new_region_cache P_ ((void));

/* Free a region cache. */
void free_region_cache P_ ((struct region_cache *));

/* Assert that the region of BUF between START and END (absolute
buffer positions) is "known," for the purposes of CACHE (e.g. "has
no newlines", in the case of the line cache). */
extern void know_region_cache P_ ((struct buffer *BUF,
                                   struct region_cache *CACHE,
                                   int START, int END));

/* Indicate that a section of BUF has changed, to invalidate CACHE.
HEAD is the number of chars unchanged at the beginning of the buffer.
TAIL is the number of chars unchanged at the end of the buffer.
NOTE: this is *not* the same as the ending position of modified
(This way of specifying regions makes more sense than absolute
buffer positions in the presence of insertions and deletions; the
args to pass are the same before and after such an operation.) */
extern void invalidate_region_cache P_ ((struct buffer *BUF,
struct region_cache *CACHE,
int HEAD, int TAIL));

/* The scanning functions.

Basically, if you're scanning forward/backward from position POS,
and region_cache_forward/backward returns true, you can skip all
the text between POS and *NEXT. And if the function returns false,
you should examine all the text from POS to *NEXT, and call
know_region_cache depending on what you find there; this way, you
might be able to avoid scanning it again. */

/* Return true if the text immediately after POS in BUF is known, for
the purposes of CACHE. If NEXT is non-zero, set *NEXT to the nearest
position after POS where the knownness changes. */
extern int region_cache_forward P_ ((struct buffer *BUF,
                                     struct region_cache *CACHE,
                                     int POS,
                                     int *NEXT));

/* Return true if the text immediately before POS in BUF is known, for
the purposes of CACHE. If NEXT is non-zero, set *NEXT to the nearest
position before POS where the knownness changes. */
extern int region_cache_backward P_ ((struct buffer *BUF,
                                      struct region_cache *CACHE,
                                      int POS,
                                      int *NEXT));

/* arch-tag: 70f79125-ef22-4f58-9aec-a48ca2791435
(do not change this comment) */
Something went wrong with that request. Please try again.