/[pcre]/code/trunk/maint/README
ViewVC logotype

Contents of /code/trunk/maint/README

Parent Directory Parent Directory | Revision Log Revision Log


Revision 122 - (show annotations)
Mon Mar 12 15:10:25 2007 UTC (13 years, 6 months ago) by ph10
File size: 12080 byte(s)
Error occurred while calculating annotation data.
Documentation spring clean, and expansion of maint/README.
1 MAINTENANCE README FOR PCRE
2 ---------------------------
3
4 The files in the "maint" directory of the PCRE source contain data, scripts,
5 and programs that are used for the maintenance of PCRE, but which do not form
6 part of the PCRE distribution tarballs. This document describes these files and
7 also contains some notes for maintainers. Its contents are:
8
9 Files in the maint directory
10 Updating to a new Unicode release
11 Preparing for a PCRE release
12 Making a PCRE release
13 Long-term ideas (wish list)
14
15
16 Files in the maint directory
17 ----------------------------
18
19 Builducptable A Perl script that creates the contents of the ucptable.h file
20 from two Unicode data files, which themselves are downloaded
21 from the Unicode web site. Run this script in the "maint"
22 directory.
23
24 Tech.Notes Some notes about the internals of the PCRE code.
25
26 Unicode.tables The files in this directory, Scripts.txt and UnicodeData.txt,
27 were downloaded from the Unicode web site. They contain
28 information about Unicode characters and scripts.
29
30 ucptest.c A short C program for testing the Unicode property functions in
31 pcre_ucp_searchfuncs.c, mainly useful after rebuilding the
32 Unicode property table. Compile and run this in the "maint"
33 directory.
34
35 ucptestdata A directory containing two files, testinput1 and testoutput1,
36 to use in conjunction with the ucptest program.
37
38 utf8.c A short, freestanding C program for converting a Unicode code
39 point into a sequence of bytes in the UTF-8 encoding, and vice
40 versa. If its argument is a hex number such as 0x1234, it
41 outputs a list of the equivalent UTF-8 bytes. If its argument
42 is sequence of concatenated UTF-8 bytes (e.g. e188b4) it treats
43 them as a UTF-8 character and outputs the equivalent code point
44 in hex.
45
46
47 Updating to a new Unicode release
48 ---------------------------------
49
50 When there is a new release of Unicode, the files in Unicode.tables must be
51 refreshed from the web site, and the Buildupctable script can then be run to
52 generate a new version of ucptable.h. The ucptest program can be used to check
53 that the resulting table works properly, using the data files in ucptestdata to
54 check a number of test characters.
55
56
57 Preparing for a PCRE release
58 ----------------------------
59
60 This section contains a checklist of things that I consult before building a
61 distribution for a new release.
62
63 . Ensure that the version number and version date are correct in configure.ac.
64
65 . Run ./autogen.sh to ensure everything is up-to-date.
66
67 . Compile and test with many different config options, and combinations of
68 options:
69
70 * Totally standard ./configure with no options
71 * --disable-shared
72 * --disable-static
73 * --enable-utf8
74 * --enable-unicode-properties
75 * --disable-cpp
76 * --with-link-size=3 (occasionally check with 4 as well)
77 * --disable-stack-for-recursion
78 * --enable-newline-is-any
79
80 I've never automated this, but perhaps I should. The newline testing could be
81 enhanced; at present, some tests fail unless plain LF is a newline.
82
83 . Run perltest.pl on the test data for tests 1 and 4. The output should match
84 the PCRE test output, apart from the version identification at the top. The
85 other tests are not Perl-compatible (they use various special PCRE options).
86
87 . Test on a number of different operating systems. In particular, at the moment
88 I can test on Solaris, using Sun's cc compiler (as a change from gcc). Adding
89 -xarch=v9 to the cc options does a 64-bit test, but it also needs -S 64 for
90 pcretest to increase the stack size for test 2. I also test on FreeBSD and
91 Linux (where I develop).
92
93 . Test with valgrind by running "RunTest valgrind". There is also "RunGrepTest
94 valgrind", though that takes quite a long time.
95
96 . It can also useful to test with Electric Fence, though the fact that it
97 grumbles for missing free() calls can be a nuisance. (A missing free() in
98 pcretest is hardly a big problem.) To build with EF, use:
99
100 LIBS='/usr/lib/libefence.a -lpthread' with ./configure.
101
102 Then all normal runs use it to check for buffer overflow. Also run everything
103 with:
104
105 EF_PROTECT_BELOW=1 <whatever>
106
107 because there have been problems with lookbehinds that looked too far.
108
109 . Test with the emulated memmove() function by undefining HAVE_MEMMOVE and
110 HAVE_BCOPY in config.h.
111
112 . Documentation: check AUTHORS, COPYING, ChangeLog (check date), INSTALL,
113 LICENCE, NEWS (check date), NON-UNIX-USE, and README. Many of these won't
114 need changing, but over the long term things do change.
115
116 . Man pages: Check all man pages for \ not followed by e or f or " because
117 that indicates a markup error.
118
119
120 Making a PCRE release
121 ---------------------
122
123 Run PrepareRelease and commit the files that it changes (by removing trailing
124 spaces). Then run "make dist" to create the tarballs and the zipball.
125
126 Don't forget to update Freshmeat when the new release is out, and to tell
127 webmaster@pcre.org and the mailing list.
128
129
130 Future ideas (wish list)
131 ------------------------
132
133 This section records a list of ideas so that they do not get forgotten. They
134 vary enormously in their usefulness and potential for implementation. Some are
135 very sensible; some are rather wacky. Some have been on this list for years;
136 others are relatively new.
137
138 . Optimization
139
140 There are always ideas for new optimizations so as to speed up pattern
141 matching. Most of them try to save work by recognizing a non-match without
142 having to scan all the possibilities. These are some that I've recorded:
143
144 * /((A{0,5}){0,5}){0,5}(something complex)/ on a non-matching string is very
145 slow, though Perl is fast. Can we speed up somehow? Convert to {0,125}?
146 OTOH, this is pathological - the user could easily fix it.
147
148 * Turn ={4} into ==== ? (for speed). I once did an experiment, and it seems
149 to have little effect, and maybe makes things worse.
150
151 * "Ends with literal string" - note that a single character doesn't gain much
152 over the existing "required byte" (reqbyte) feature that just saves one
153 byte.
154
155 * These probably need to go in study():
156
157 o Remember an initial string rather than just 1 char?
158
159 o A required byte from alternatives - not just the last char, but an
160 earlier one if common to all alternatives.
161
162 o Minimum length of subject needed.
163
164 o Friedl contains other ideas.
165
166 . If Perl gets to a consistent state over the settings of capturing sub-
167 patterns inside repeats, see if we can match it. One example of the
168 difference is the matching of /(main(O)?)+/ against mainOmain, where PCRE
169 leaves $2 set. In Perl, it's unset. Changing this in PCRE will be very hard
170 because I think it needs much more state to be remembered.
171
172 . Perl 6 will be a revolution. Is it a revolution too far for PCRE?
173
174 . Unicode
175
176 * Note that in Perl, \s matches \pZ and similarly for \d, \w and the POSIX
177 character classes. For the moment, I've chosen not to support this for
178 backward compatibility, for speed, and because it would be messy to
179 implement.
180
181 * A different approach to Unicode might be to use a typedef to do everything
182 in unsigned shorts instead of unsigned chars. Actually, we'd have to have a
183 new typedef to distinguish data from bits of compiled pattern that are in
184 bytes, I think. There would need to be conversion functions in and out. I
185 don't think this is particularly trivial - and anyway, Unicode now has
186 characters that need more than 16 bits, so is this at all sensible?
187
188 * There has been a request for direct support of 16-bit characters and
189 UTF-16. However, since Unicode is moving beyond purely 16-bit characters,
190 is this worth it at all? One possible way of handling 16-bit characters
191 would be to "load" them in the same way that UTF-8 characters are loaded.
192
193 . Allow errorptr and erroroffset to be NULL. I don't like this idea.
194
195 . Line endings:
196
197 * Option to use NUL as a line terminator in subject strings. This could now
198 be done relatively easily since the extension to support LF, CR, and CRLF.
199 If this is done, a suitable option for pcregrep is also required.
200
201 . Option to provide the pattern with a length instead of with a NUL terminator.
202 This probably affects quite a few places in the code.
203
204 . Catch SIGSEGV for stack overflows?
205
206 . "Cut" as described in Jeffrey Friedl's book, p364: \v and \V. The definitions
207 aren't yet clear enough for me. \v flushes saved states so that no
208 backtracking to anything earlier can happen; \V says "no more bumpalong", but
209 does it fail the current match? As described in the book, these aren't really
210 "cut" as in Prolog, are they? NOTE: (a) PCRE once had "cut", but it was
211 removed when atomic groups were introduced. (b) Perl 5.10 has some (*PRUNE)
212 features -- see below.
213
214 . A feature to suspend a match via a callout was once requested.
215
216 . Option to convert results into character offsets and character lengths.
217
218 . Option for pcregrep to scan only the start of a file. I am not keen - this is
219 the job of "head".
220
221 . A (non-Unix) user wanted pcregrep options to (a) list a file name just once,
222 preceded by a blank line, instead of adding it to every matched line, and (b)
223 support --outputfile=name.
224
225 . Consider making UTF-8 and UCP the default for PCRE n.0 for some n > 7.
226
227 . Add a user pointer to pcre_malloc/free functions -- some option would be
228 needed to retain backward compatibility.
229
230 . Define a union for the results from pcre_fullinfo().
231
232 . Provide a "random access to the subject" facility so that the way in which it
233 is stored is independent of PCRE. For efficiency, it probably isn't possible
234 to switch this dynamically. It would have to be specified when PCRE was
235 compiled. PCRE would then call a function every time it wanted a character.
236
237 . There are new (*PRUNE) facilities in Perl 5.10, some of which it might be
238 relatively easy to implement.
239
240 . Also in Perl 5.10 are relative subroutine references (?&-1) and (?&+1) which
241 I didn't know about when I added some 5.10 features for PCRE 7.0. What about
242 (?(-1)... as a condition? That's an obvious extension, even if Perl 5.10
243 doesn't have it.
244
245 . Wild thought: the ability to compile from PCRE's internal byte code to a real
246 FSM and a very fast (third) matcher to process the result. There would be
247 even more restrictions than for pcre_dfa_exec(), however. This is not easy.
248
249 . Should pcretest have some private locale data, to avoid relying on the
250 available locales for the test data, since different OS have different ideas?
251 This won't be as thorough a test, but perhaps that doesn't really matter.
252
253 . pcregrep: add -rs for a sorted recurse? Having to store file names and sort
254 them will of course slow it down.
255
256 . Re-arrange test 2: take out the link-size dependent stuff for a separate test
257 that is run only when the link size *is* 2; leave in some non-numbered
258 debugging tests using the new /Z feature.
259
260 . Stan Switzer's goto replacement for longjmp, which is apparently very slow on
261 OS-X. This is used when stack recursion is disabled. It would be worth doing
262 some timing tests on other OS.
263
264 . Someone suggested --disable-callout to save code space when callouts are
265 never wanted. This seems rather marginal.
266
267 . Automate some of the testing before release into a script that compiles with
268 different options and runs the tests in each case.
269
270 . How about distributing a fixed pcre_chartables.c file and abandoning the
271 on-the-fly generation using dftables. This will make cross-compiling easier,
272 and in any case, locales are going out of fashion.
273
274 Philip Hazel
275 Email local part: ph10
276 Email domain: cam.ac.uk
277 Last updated: 12 March 2007

  ViewVC Help
Powered by ViewVC 1.1.5