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

Diff of /code/trunk/maint/README

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 155 by ph10, Tue Apr 24 13:36:11 2007 UTC revision 1232 by ph10, Fri Dec 7 12:08:44 2012 UTC
# Line 1  Line 1 
1  MAINTENANCE README FOR PCRE  MAINTENANCE README FOR PCRE
2  ---------------------------  ===========================
3    
4  The files in the "maint" directory of the PCRE source contain data, scripts,  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  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  part of the PCRE distribution tarballs. This document describes these files and
7  also contains some notes for maintainers. Its contents are:  also contains some notes for maintainers. Its contents are:
8    
9    Files in the maint directory    Files in the maint directory
10    Updating to a new Unicode release    Updating to a new Unicode release
11    Preparing for a PCRE release    Preparing for a PCRE release
12    Making a PCRE release    Making a PCRE release
13    Long-term ideas (wish list)    Long-term ideas (wish list)
14    
15    
16  Files in the maint directory  Files in the maint directory
17  ----------------------------  ============================
18    
19    GenerateUtt.py   A Python script to generate part of the pcre_tables.c file
20                     that contains Unicode script names in a long string with
21                     offsets, which is tedious to maintain by hand.
22    
 Builducptable    A Perl script that creates the contents of the ucptable.h file  
                  from two Unicode data files, which themselves are downloaded  
                  from the Unicode web site. Run this script in the "maint"  
                  directory.  
   
23  ManyConfigTests  A shell script that runs "configure, make, test" a number of  ManyConfigTests  A shell script that runs "configure, make, test" a number of
24                   times with different configuration settings.                   times with different configuration settings.
25    
26  Unicode.tables   The files in this directory, Scripts.txt and UnicodeData.txt,  MultiStage2.py   A Python script that generates the file pcre_ucd.c from three
27                   were downloaded from the Unicode web site. They contain                   Unicode data tables, which are themselves downloaded from the
28                   information about Unicode characters and scripts.                   Unicode web site. Run this script in the "maint" directory.
29                     The generated file contains the tables for a 2-stage lookup
30  ucptest.c        A short C program for testing the Unicode property functions                   of Unicode properties.
31                   in pcre_ucp_searchfuncs.c, mainly useful after rebuilding the  
32                   Unicode property table. Compile and run this in the "maint"  pcre_chartables.c.non-standard
33                   directory.                   This is a set of character tables that came from a Windows
34                     system. It has characters greater than 128 that are set as
35                     spaces, amongst other things. I kept it so that it can be
36                     used for testing from time to time.
37    
38    README           This file.
39    
40    Unicode.tables   The files in this directory (CaseFolding.txt,
41                     DerivedGeneralCategory.txt, GraphemeBreakProperty.txt,
42                     Scripts.txt and UnicodeData.txt) were downloaded from the
43                     Unicode web site. They contain information about Unicode
44                     characters and scripts.
45    
46    ucptest.c        A short C program for testing the Unicode property macros
47                     that do lookups in the pcre_ucd.c data, mainly useful after
48                     rebuilding the Unicode property table. Compile and run this in
49                     the "maint" directory (see comments at its head).
50    
51  ucptestdata      A directory containing two files, testinput1 and testoutput1,  ucptestdata      A directory containing two files, testinput1 and testoutput1,
52                   to use in conjunction with the ucptest program.                   to use in conjunction with the ucptest program.
53    
54  utf8.c           A short, freestanding C program for converting a Unicode code  utf8.c           A short, freestanding C program for converting a Unicode code
55                   point into a sequence of bytes in the UTF-8 encoding, and vice                   point into a sequence of bytes in the UTF-8 encoding, and vice
56                   versa. If its argument is a hex number such as 0x1234, it                   versa. If its argument is a hex number such as 0x1234, it
# Line 43  utf8.c           A short, freestanding C Line 58  utf8.c           A short, freestanding C
58                   is sequence of concatenated UTF-8 bytes (e.g. e188b4) it                   is sequence of concatenated UTF-8 bytes (e.g. e188b4) it
59                   treats them as a UTF-8 character and outputs the equivalent                   treats them as a UTF-8 character and outputs the equivalent
60                   code point in hex.                   code point in hex.
61    
62    
63  Updating to a new Unicode release  Updating to a new Unicode release
64  ---------------------------------  =================================
65    
66    When there is a new release of Unicode, the files in Unicode.tables must be
67    refreshed from the web site. If the new version of Unicode adds new character
68    scripts, the source file ucp.h and both the MultiStage2.py and the
69    GenerateUtt.py scripts must be edited to add the new names. Then MultiStage2.py
70    can be run to generate a new version of pcre_ucd.c, and GenerateUtt.py can be
71    run to generate the tricky tables for inclusion in pcre_tables.c.
72    
73    If MultiStage2.py gives the error "ValueError: list.index(x): x not in list",
74    the cause is usually a missing (or misspelt) name in the list of scripts. I
75    couldn't find a straightforward list of scripts on the Unicode site, but
76    there's a useful Wikipedia page that list them, and notes the Unicode version
77    in which they were introduced:
78    
79    http://en.wikipedia.org/wiki/Unicode_scripts#Table_of_Unicode_scripts
80    
81    The ucptest program can be compiled and used to check that the new tables in
82    pcre_ucd.c work properly, using the data files in ucptestdata to check a number
83    of test characters. The source file ucptest.c must be updated whenever new
84    Unicode script names are added.
85    
86  When there is a new release of Unicode, the files in Unicode.tables must be  Note also that both the pcresyntax.3 and pcrepattern.3 man pages contain lists
87  refreshed from the web site, and the Buildupctable script can then be run to  of Unicode script names.
 generate a new version of ucptable.h. The ucptest program can be used to check  
 that the resulting table works properly, using the data files in ucptestdata to  
 check a number of test characters.  
88    
89    
90  Preparing for a PCRE release  Preparing for a PCRE release
91  ----------------------------  ============================
92    
93  This section contains a checklist of things that I consult before building a  This section contains a checklist of things that I consult before building a
94  distribution for a new release.  distribution for a new release.
95    
96  . Ensure that the version number and version date are correct in configure.ac,  . Ensure that the version number and version date are correct in configure.ac.
   ChangeLog, and NEWS.  
97    
98  . Run ./autogen.sh to ensure everything is up-to-date.  . Update the library version numbers in configure.ac according to the rules
99      given below.
100    
101  . Compile and test with many different config options, and combinations of  . If new build options have been added, ensure that they are added to the CMake
102    options. The maint/ManyConfigTests script now encapsulates this testing.    files as well as to the autoconf files. The relevant files are CMakeLists.txt
103      and config-cmake.h.in. After making a release tarball, test it out with CMake
104  . Run perltest.pl on the test data for tests 1 and 4. The output should match    if there have been changes here.
   the PCRE test output, apart from the version identification at the top. The  
   other tests are not Perl-compatible (they use various special PCRE options).  
105    
106  . Test with valgrind by running "RunTest valgrind". There is also "RunGrepTest  . Run ./autogen.sh to ensure everything is up-to-date.
   valgrind", though that takes quite a long time.  
   
 . It may also useful to test with Electric Fence, though the fact that it  
   grumbles for missing free() calls can be a nuisance. (A missing free() in  
   pcretest is hardly a big problem.) To build with EF, use:  
   
     LIBS='/usr/lib/libefence.a -lpthread' with ./configure.  
107    
108    Then all normal runs use it to check for buffer overflow. Also run everything  . Compile and test with many different config options, and combinations of
109    with:    options. Also, test with valgrind by running "RunTest valgrind" and
110      "RunGrepTest valgrind" (which takes quite a long time). The script
111      EF_PROTECT_BELOW=1 <whatever>    maint/ManyConfigTests now encapsulates this testing. It runs tests with
112      different configurations, and it also runs some of them with valgrind, all of
113    because there have been problems with lookbehinds that looked too far.    which can take quite some time.
114    
115  . Test with the emulated memmove() function by undefining HAVE_MEMMOVE and  . Run perltest.pl on the test data for tests 1, 4, and 6. The output
116    HAVE_BCOPY in config.h. You may see a number of "pcre_memmove defined but not    should match the PCRE test output, apart from the version identification at
117    used" warnings for the modules in which there is no call to memmove(). These    the start of each test. The other tests are not Perl-compatible (they use
118    can be ignored.    various PCRE-specific features or options).
119    
120  . Documentation: check AUTHORS, COPYING, ChangeLog (check date), INSTALL,  . It is possible to test with the emulated memmove() function by undefining
121    LICENCE, NEWS (check date), NON-UNIX-USE, and README. Many of these won't    HAVE_MEMMOVE and HAVE_BCOPY in config.h, though I do not do this often. You
122    need changing, but over the long term things do change.    may see a number of "pcre_memmove defined but not used" warnings for the
123      modules in which there is no call to memmove(). These can be ignored.
124  . Man pages: Check all man pages for \ not followed by e or f or " because  
125    that indicates a markup error.  . Documentation: check AUTHORS, ChangeLog (check version and date), LICENCE,
126      NEWS (check version and date), NON-AUTOTOOLS-BUILD, and README. Many of these
127      won't need changing, but over the long term things do change.
128    
129    . I used to test new releases myself on a number of different operating
130      systems, using different compilers as well. For example, on Solaris it is
131      helpful to test using Sun's cc compiler as a change from gcc. Adding
132      -xarch=v9 to the cc options does a 64-bit test, but it also needs -S 64 for
133      pcretest to increase the stack size for test 2. Since I retired I can no
134      longer do this, but instead I rely on putting out release candidates for
135      folks on the pcre-dev list to test.
136    
137    
138    Updating version info for libtool
139    =================================
140    
141    This set of rules for updating library version information came from a web page
142    whose URL I have forgotten. The version information consists of three parts:
143    (current, revision, age).
144    
145    1. Start with version information of 0:0:0 for each libtool library.
146    
147    2. Update the version information only immediately before a public release of
148       your software. More frequent updates are unnecessary, and only guarantee
149       that the current interface number gets larger faster.
150    
151    3. If the library source code has changed at all since the last update, then
152       increment revision; c:r:a becomes c:r+1:a.
153    
154    4. If any interfaces have been added, removed, or changed since the last
155       update, increment current, and set revision to 0.
156    
157    5. If any interfaces have been added since the last public release, then
158       increment age.
159    
160    6. If any interfaces have been removed or changed since the last public
161       release, then set age to 0.
162    
163    The following explanation may help in understanding the above rules a bit
164    better. Consider that there are three possible kinds of reaction from users to
165    changes in a shared library:
166    
167    1. Programs using the previous version may use the new version as a drop-in
168       replacement, and programs using the new version can also work with the
169       previous one. In other words, no recompiling nor relinking is needed. In
170       this case, increment revision only, don't touch current or age.
171    
172    2. Programs using the previous version may use the new version as a drop-in
173       replacement, but programs using the new version may use APIs not present in
174       the previous one. In other words, a program linking against the new version
175       may fail if linked against the old version at run time. In this case, set
176       revision to 0, increment current and age.
177    
178  . When the release is built, test it on a number of different operating  3. Programs may need to be changed, recompiled, relinked in order to use the
179    systems if possible, and using different compilers as well. For example,     new version. Increment current, set revision and age to 0.
   on Solaris it is helpful to test using Sun's cc compiler as a change from  
   gcc. Adding -xarch=v9 to the cc options does a 64-bit test, but it also  
   needs -S 64 for pcretest to increase the stack size for test 2.  
180    
181    
182  Making a PCRE release  Making a PCRE release
183  ---------------------  =====================
184    
185  Run PrepareRelease and commit the files that it changes (by removing trailing  Run PrepareRelease and commit the files that it changes (by removing trailing
186  spaces). Then run "make distcheck" to create the tarballs and the zipball.  spaces). The first thing this script does is to run CheckMan on the man pages;
187    if it finds any markup errors, it reports them and then aborts.
188    
189    Once PrepareRelease has run clean, run "make distcheck" to create the tarballs
190    and the zipball. Double-check with "svn status", then create an SVN tagged
191    copy:
192    
193      svn copy svn://vcs.exim.org/pcre/code/trunk \
194               svn://vcs.exim.org/pcre/code/tags/pcre-8.xx
195    
196  Don't forget to update Freshmeat when the new release is out, and to tell  Don't forget to update Freshmeat when the new release is out, and to tell
197  webmaster@pcre.org and the mailing list.  webmaster@pcre.org and the mailing list. Also, update the list of version
198    numbers in Bugzilla (edit products).
199    
200    
201  Future ideas (wish list)  Future ideas (wish list)
202  ------------------------  ========================
203    
204  This section records a list of ideas so that they do not get forgotten. They  This section records a list of ideas so that they do not get forgotten. They
205  vary enormously in their usefulness and potential for implementation. Some are  vary enormously in their usefulness and potential for implementation. Some are
206  very sensible; some are rather wacky. Some have been on this list for years;  very sensible; some are rather wacky. Some have been on this list for years;
207  others are relatively new.  others are relatively new.
208    
209  . Optimization  . Optimization
210    
211    There are always ideas for new optimizations so as to speed up pattern    There are always ideas for new optimizations so as to speed up pattern
212    matching. Most of them try to save work by recognizing a non-match without    matching. Most of them try to save work by recognizing a non-match without
213    having to scan all the possibilities. These are some that I've recorded:    having to scan all the possibilities. These are some that I've recorded:
214    
215    * /((A{0,5}){0,5}){0,5}(something complex)/ on a non-matching string is very    * /((A{0,5}){0,5}){0,5}(something complex)/ on a non-matching string is very
216      slow, though Perl is fast. Can we speed up somehow? Convert to {0,125}?      slow, though Perl is fast. Can we speed up somehow? Convert to {0,125}?
217      OTOH, this is pathological - the user could easily fix it.      OTOH, this is pathological - the user could easily fix it.
218    
219    * Turn ={4} into ==== ? (for speed). I once did an experiment, and it seems    * Turn ={4} into ==== ? (for speed). I once did an experiment, and it seems
220      to have little effect, and maybe makes things worse.      to have little effect, and maybe makes things worse.
221    
222    * "Ends with literal string" - note that a single character doesn't gain much    * "Ends with literal string" - note that a single character doesn't gain much
223      over the existing "required byte" (reqbyte) feature that just saves one      over the existing "required byte" (reqbyte) feature that just remembers one
224      byte.      data unit.
225    
226    * These probably need to go in study():    * These probably need to go in pcre_study():
227    
228      o Remember an initial string rather than just 1 char?      o Remember an initial string rather than just 1 char?
229    
230      o A required byte from alternatives - not just the last char, but an      o A required data unit from alternatives - not just the last unit, but an
231        earlier one if common to all alternatives.        earlier one if common to all alternatives.
232    
     o Minimum length of subject needed.  
   
233      o Friedl contains other ideas.      o Friedl contains other ideas.
234    
235      * pcre_study() does not set initial byte flags for Unicode property types
236        such as \p; I don't know how much benefit there would be for, for example,
237        setting the bits for 0-9 and all bytes >= xC0 when a pattern starts with
238        \p{N}.
239    
240      * There is scope for more "auto-possessifying" in connection with \p and \P.
241    
242  . If Perl gets to a consistent state over the settings of capturing sub-  . If Perl gets to a consistent state over the settings of capturing sub-
243    patterns inside repeats, see if we can match it. One example of the    patterns inside repeats, see if we can match it. One example of the
244    difference is the matching of /(main(O)?)+/ against mainOmain, where PCRE    difference is the matching of /(main(O)?)+/ against mainOmain, where PCRE
245    leaves $2 set. In Perl, it's unset. Changing this in PCRE will be very hard    leaves $2 set. In Perl, it's unset. Changing this in PCRE will be very hard
246    because I think it needs much more state to be remembered.    because I think it needs much more state to be remembered.
   
 . Perl 6 will be a revolution. Is it a revolution too far for PCRE?  
247    
248  . Unicode  . Perl 6 will be a revolution. Is it a revolution too far for PCRE?
249    
   * Note that in Perl, \s matches \pZ and similarly for \d, \w and the POSIX  
     character classes. For the moment, I've chosen not to support this for  
     backward compatibility, for speed, and because it would be messy to  
     implement.  
   
   * A different approach to Unicode might be to use a typedef to do everything  
     in unsigned shorts instead of unsigned chars. Actually, we'd have to have a  
     new typedef to distinguish data from bits of compiled pattern that are in  
     bytes, I think. There would need to be conversion functions in and out. I  
     don't think this is particularly trivial - and anyway, Unicode now has  
     characters that need more than 16 bits, so is this at all sensible?  
   
   * There has been a request for direct support of 16-bit characters and  
     UTF-16. However, since Unicode is moving beyond purely 16-bit characters,  
     is this worth it at all? One possible way of handling 16-bit characters  
     would be to "load" them in the same way that UTF-8 characters are loaded.  
   
250  . Allow errorptr and erroroffset to be NULL. I don't like this idea.  . Allow errorptr and erroroffset to be NULL. I don't like this idea.
251    
252  . Line endings:  . Line endings:
253    
254    * Option to use NUL as a line terminator in subject strings. This could now    * Option to use NUL as a line terminator in subject strings. This could now
255      be done relatively easily since the extension to support LF, CR, and CRLF.      be done relatively easily since the extension to support LF, CR, and CRLF.
256      If this is done, a suitable option for pcregrep is also required.      If it is done, a suitable option for pcregrep is also required.
257    
258  . Option to provide the pattern with a length instead of with a NUL terminator.  . Option to provide the pattern with a length instead of with a NUL terminator.
259    This probably affects quite a few places in the code.    This affects quite a few places in the code and is not trivial.
260    
261  . Catch SIGSEGV for stack overflows?  . Catch SIGSEGV for stack overflows?
   
 . "Cut" as described in Jeffrey Friedl's book, p364: \v and \V. The definitions  
   aren't yet clear enough for me. \v flushes saved states so that no  
   backtracking to anything earlier can happen; \V says "no more bumpalong", but  
   does it fail the current match? As described in the book, these aren't really  
   "cut" as in Prolog, are they? NOTE: (a) PCRE once had "cut", but it was  
   removed when atomic groups were introduced. (b) Perl 5.10 has some (*PRUNE)  
   features -- see below.  
262    
263  . A feature to suspend a match via a callout was once requested.  . A feature to suspend a match via a callout was once requested.
264    
265  . Option to convert results into character offsets and character lengths.  . Option to convert results into character offsets and character lengths.
266    
267  . Option for pcregrep to scan only the start of a file. I am not keen - this is  . Option for pcregrep to scan only the start of a file. I am not keen - this is
268    the job of "head".    the job of "head".
269    
270  . A (non-Unix) user wanted pcregrep options to (a) list a file name just once,  . A (non-Unix) user wanted pcregrep options to (a) list a file name just once,
271    preceded by a blank line, instead of adding it to every matched line, and (b)    preceded by a blank line, instead of adding it to every matched line, and (b)
272    support --outputfile=name.    support --outputfile=name.
   
 . Consider making UTF-8 and UCP the default for PCRE n.0 for some n > 7.  
273    
274  . Add a user pointer to pcre_malloc/free functions -- some option would be  . Consider making UTF-8 and UCP the default for PCRE n.0 for some n > 8.
275      (And now presumably UTF-16 and UCP for the 16-bit library, and UTF-32 and UCP
276      for the 32-bit library.)
277    
278    . Add a user pointer to pcre_malloc/free functions -- some option would be
279    needed to retain backward compatibility.    needed to retain backward compatibility.
280    
281  . Define a union for the results from pcre_fullinfo().  . Define a union for the results from pcre_fullinfo().
282    
283  . Provide a "random access to the subject" facility so that the way in which it  . Provide a "random access to the subject" facility so that the way in which it
284    is stored is independent of PCRE. For efficiency, it probably isn't possible    is stored is independent of PCRE. For efficiency, it probably isn't possible
285    to switch this dynamically. It would have to be specified when PCRE was    to switch this dynamically. It would have to be specified when PCRE was
286    compiled. PCRE would then call a function every time it wanted a character.    compiled. PCRE would then call a function every time it wanted a character.
287    
 . There are new (*PRUNE) facilities in Perl 5.10, some of which it might be  
   relatively easy to implement.  
   
 . Also in Perl 5.10 are relative subroutine references (?&-1) and (?&+1) which  
   I didn't know about when I added some 5.10 features for PCRE 7.0. What about  
   (?(-1)... as a condition? That's an obvious extension, even if Perl 5.10  
   doesn't have it.  
   
288  . Wild thought: the ability to compile from PCRE's internal byte code to a real  . Wild thought: the ability to compile from PCRE's internal byte code to a real
289    FSM and a very fast (third) matcher to process the result. There would be    FSM and a very fast (third) matcher to process the result. There would be
290    even more restrictions than for pcre_dfa_exec(), however. This is not easy.    even more restrictions than for pcre_dfa_exec(), however. This is not easy.
291      This is probably obsolete now that we have the JIT support.
292    
293  . Should pcretest have some private locale data, to avoid relying on the  . Should pcretest have some private locale data, to avoid relying on the
294    available locales for the test data, since different OS have different ideas?    available locales for the test data, since different OS have different ideas?
295    This won't be as thorough a test, but perhaps that doesn't really matter.    This won't be as thorough a test, but perhaps that doesn't really matter.
296    
297  . pcregrep: add -rs for a sorted recurse? Having to store file names and sort  . pcregrep: add -rs for a sorted recurse? Having to store file names and sort
298    them will of course slow it down.    them will of course slow it down.
299    
300  . Re-arrange test 2: take out the link-size dependent stuff for a separate test  . Someone suggested --disable-callout to save code space when callouts are
301    that is run only when the link size *is* 2; leave in some non-numbered    never wanted. This seems rather marginal.
   debugging tests using the new /Z feature.  
   
 . Stan Switzer's goto replacement for longjmp, which is apparently very slow on  
   OS-X. This is used when stack recursion is disabled. It would be worth doing  
   some timing tests on other OS.  
302    
303  . Someone suggested --disable-callout to save code space when callouts are  . Check names that consist entirely of digits: PCRE allows, but do Perl and
304    never wanted. This seems rather marginal.    Python, etc?
305    
306  . Work needs doing so that the pcregrep tests work better with different  . A user suggested a parameter to limit the length of string matched, for
307    linebreak settings. Currently, some tests don't work when the input files    example if the parameter is N, the current match should fail if the matched
308    do not have \n line endings.    substring exceeds N. This could apply to both match functions. The value
309      could be a new field in the extra block.
310  . If the fr_FR locale isn't available for testing, try "french" instead,  
311    because this may be available on Windows. It means modifying the test data,  . Callouts with arguments: (?Cn:ARG) for instance.
312    however.  
313    . A user is going to supply a patch to generalize the API for user-specific
314      memory allocation so that it is more flexible in threaded environments. This
315      was promised a long time ago, and never appeared. However, this is a live
316      issue not only for threaded environments, but for libraries that use PCRE and
317      want not to be beholden to their caller's memory allocation.
318    
319    . Write a wrapper to maintain a structure with specified runtime parameters,
320      such as recurse limit, and pass these to PCRE each time it is called. Also
321      maybe malloc and free. A user sent a prototype. This relates the the previous
322      item.
323    
324    . Write a function that generates random matching strings for a compiled regex.
325    
326    . Pcregrep: an option to specify the output line separator, either as a string
327      or select from a fixed list. This is not dead easy, because at the moment it
328      outputs whatever is in the input file.
329    
330    . Improve the code for duplicate checking in pcre_dfa_exec(). An incomplete,
331      non-thread-safe patch showed that this can help performance for patterns
332      where there are many alternatives. However, a simple thread-safe
333      implementation that I tried made things worse in many simple cases, so this
334      is not an obviously good thing.
335    
336    . PCRE cannot at present distinguish between subpatterns with different names,
337      but the same number (created by the use of ?|). In order to do so, a way of
338      remembering *which* subpattern numbered n matched is needed. Bugzilla #760.
339      Now that (*MARK) has been implemented, it can perhaps be used as a way round
340      this problem.
341    
342    . Instead of having #ifdef HAVE_CONFIG_H in each module, put #include
343      "something" and the the #ifdef appears only in one place, in "something".
344    
345  Philip Hazel  Philip Hazel
346  Email local part: ph10  Email local part: ph10
347  Email domain: cam.ac.uk  Email domain: cam.ac.uk
348  Last updated: 24 April 2007  Last updated: 07 December 2012

Legend:
Removed from v.155  
changed lines
  Added in v.1232

  ViewVC Help
Powered by ViewVC 1.1.5