Contents of /code/trunk/maint/README

MAINTENANCE README FOR PCRE
---------------------------

The files in the "maint" directory of the PCRE source contain data, scripts,
and programs that are used for the maintenance of PCRE, but which do not form
part of the PCRE distribution tarballs. This document describes these files and
also contains some notes for maintainers. Its contents are:

  Files in the maint directory
  Updating to a new Unicode release
  Preparing for a PCRE release
  Making a PCRE release
  Long-term ideas (wish list)


Files in the maint directory
----------------------------

----------------- This file is now OBSOLETE and no longer used ----------------
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.
----------------- This file is now OBSOLETE and no longer used ----------------

GenerateUtt.py   A Python script to generate part of the pcre_tables.c file
                 that contains Unicode script names in a long string with
                 offsets, which is tedious to maintain by hand. 

ManyConfigTests  A shell script that runs "configure, make, test" a number of
                 times with different configuration settings.
                 
MultiStage2.py   A Python script that generates the file pcre_ucd.c from three
                 Unicode data tables, which are themselves downloaded from the
                 Unicode web site. Run this script in the "maint" directory. 
                 The generated file contains the tables for a 2-stage lookup
                 of Unicode properties.  

README           This file.

Unicode.tables   The files in this directory, DerivedGeneralCategory.txt, 
                 Scripts.txt and UnicodeData.txt, were downloaded from the
                 Unicode web site. They contain information about Unicode
                 characters and scripts.

ucptest.c        A short C program for testing the Unicode property macros
                 that do lookups in the pcre_ucd.c data, mainly useful after
                 rebuilding the Unicode property table. Compile and run this in
                 the "maint" directory (see comments at its head).

ucptestdata      A directory containing two files, testinput1 and testoutput1,
                 to use in conjunction with the ucptest program.

utf8.c           A short, freestanding C program for converting a Unicode code
                 point into a sequence of bytes in the UTF-8 encoding, and vice
                 versa. If its argument is a hex number such as 0x1234, it
                 outputs a list of the equivalent UTF-8 bytes. If its argument
                 is sequence of concatenated UTF-8 bytes (e.g. e188b4) it
                 treats them as a UTF-8 character and outputs the equivalent
                 code point in hex.


Updating to a new Unicode release
---------------------------------

When there is a new release of Unicode, the files in Unicode.tables must be
refreshed from the web site. If the new version of Unicode adds new character
scripts, the source file ucp.h and both the MultiStage2.py and the
GenerateUtt.py scripts must be edited to add the new names. Then MultiStage2.py
can be run to generate a new version of pcre_ucd.c, and GenerateUtt.py can be
run to generate the tricky tables for inclusion in pcre_tables.c.

If MultiStage2.py gives the error "ValueError: list.index(x): x not in list",
the cause is usually a missing (or misspelt) name in the list of scripts. I 
couldn't find a straightforward list of scripts on the Unicode site, but 
there's a useful Wikipedia page that list them, and notes the Unicode version 
in which they were introduced:

http://en.wikipedia.org/wiki/Unicode_scripts#Table_of_Unicode_scripts

The ucptest program can be compiled and used to check that the new tables in
pcre_ucd.c work properly, using the data files in ucptestdata to check a number
of test characters. The source file ucptest.c must be updated whenever new
Unicode script names are added.

Note also that both the pcresyntax.3 and pcrepattern.3 man pages contain lists 
of Unicode script names.


Preparing for a PCRE release
----------------------------

This section contains a checklist of things that I consult before building a
distribution for a new release.

. Ensure that the version number and version date are correct in configure.ac.
  
. If new build options have been added, ensure that they are added to the CMake
  files as well as to the autoconf files. 

. Run ./autogen.sh to ensure everything is up-to-date.

. Compile and test with many different config options, and combinations of
  options. The maint/ManyConfigTests script now encapsulates this testing.

. Run perltest.pl on the test data for tests 1, 4, 6, and 11. The first two can 
  be run with Perl 5.8 or 5.10; the last two require Perl 5.10. The output
  should match the PCRE test output, apart from the version identification at
  the start of each test. The other tests are not Perl-compatible (they use
  various PCRE-specific features or options).

. Test with valgrind by running "RunTest valgrind". There is also "RunGrepTest
  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.

  Then all normal runs use it to check for buffer overflow. Also run everything
  with:

    EF_PROTECT_BELOW=1 <whatever>

  because there have been problems with lookbehinds that looked too far.

. Test with the emulated memmove() function by undefining HAVE_MEMMOVE and
  HAVE_BCOPY in config.h. You may see a number of "pcre_memmove defined but not
  used" warnings for the modules in which there is no call to memmove(). These
  can be ignored.

. Documentation: check AUTHORS, COPYING, ChangeLog (check version and date), 
  INSTALL, LICENCE, NEWS (check version and date), NON-UNIX-USE, and README.
  Many of these won't need changing, but over the long term things do change.

. Man pages: Check all man pages for \ not followed by e or f or " because
  that indicates a markup error.

. When the release is built, test it on a number of different operating
  systems if possible, and using different compilers as well. For example,
  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.


Making a PCRE release
---------------------

Run PrepareRelease and commit the files that it changes (by removing trailing
spaces). Then run "make distcheck" to create the tarballs and the zipball.
Double-check with "svn status", then create an SVN tagged copy:

  svn copy svn://vcs.exim.org/pcre/code/trunk \
           svn://vcs.exim.org/pcre/code/tags/pcre-8.xx 

Don't forget to update Freshmeat when the new release is out, and to tell
webmaster@pcre.org and the mailing list.


Future ideas (wish list)
------------------------

This section records a list of ideas so that they do not get forgotten. They
vary enormously in their usefulness and potential for implementation. Some are
very sensible; some are rather wacky. Some have been on this list for years;
others are relatively new.

. Optimization

  There are always ideas for new optimizations so as to speed up pattern
  matching. Most of them try to save work by recognizing a non-match without
  having to scan all the possibilities. These are some that I've recorded:

  * /((A{0,5}){0,5}){0,5}(something complex)/ on a non-matching string is very
    slow, though Perl is fast. Can we speed up somehow? Convert to {0,125}?
    OTOH, this is pathological - the user could easily fix it.

  * Turn ={4} into ==== ? (for speed). I once did an experiment, and it seems
    to have little effect, and maybe makes things worse.

  * "Ends with literal string" - note that a single character doesn't gain much
    over the existing "required byte" (reqbyte) feature that just remembers one
    byte.

  * These probably need to go in study():

    o Remember an initial string rather than just 1 char?

    o A required byte from alternatives - not just the last char, but an
      earlier one if common to all alternatives.

    o Friedl contains other ideas.
    
. If Perl gets to a consistent state over the settings of capturing sub-
  patterns inside repeats, see if we can match it. One example of the
  difference is the matching of /(main(O)?)+/ against mainOmain, where PCRE
  leaves $2 set. In Perl, it's unset. Changing this in PCRE will be very hard
  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?

. Unicode

  * 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.

. Allow errorptr and erroroffset to be NULL. I don't like this idea.

. Line endings:

  * Option to use NUL as a line terminator in subject strings. This could now
    be done relatively easily since the extension to support LF, CR, and CRLF.
    If it is done, a suitable option for pcregrep is also required.

. Option to provide the pattern with a length instead of with a NUL terminator.
  This affects quite a few places in the code and is not trivial.

. Catch SIGSEGV for stack overflows?

. A feature to suspend a match via a callout was once requested.

. Option to convert results into character offsets and character lengths.

. Option for pcregrep to scan only the start of a file. I am not keen - this is
  the job of "head".

. A (non-Unix) user wanted pcregrep options to (a) list a file name just once,
  preceded by a blank line, instead of adding it to every matched line, and (b)
  support --outputfile=name.

. Consider making UTF-8 and UCP the default for PCRE n.0 for some n > 8.

. Add a user pointer to pcre_malloc/free functions -- some option would be
  needed to retain backward compatibility.

. Define a union for the results from pcre_fullinfo().

. Provide a "random access to the subject" facility so that the way in which it
  is stored is independent of PCRE. For efficiency, it probably isn't possible
  to switch this dynamically. It would have to be specified when PCRE was
  compiled. PCRE would then call a function every time it wanted a character.

. Wild thought: the ability to compile from PCRE's internal byte code to a real
  FSM and a very fast (third) matcher to process the result. There would be
  even more restrictions than for pcre_dfa_exec(), however. This is not easy.

. Should pcretest have some private locale data, to avoid relying on the
  available locales for the test data, since different OS have different ideas?
  This won't be as thorough a test, but perhaps that doesn't really matter.

. pcregrep: add -rs for a sorted recurse? Having to store file names and sort
  them will of course slow it down.

. Someone suggested --disable-callout to save code space when callouts are
  never wanted. This seems rather marginal.
  
. Check names that consist entirely of digits: PCRE allows, but do Perl and 
  Python, etc? 
  
. A user suggested a parameter to limit the length of string matched, for 
  example if the parameter is N, the current match should fail if the matched 
  substring exceeds N. This could apply to both match functions. The value 
  could be a new field in the extra block.
  
. Callouts with arguments: (?Cn:ARG) for instance.

. A user is going to supply a patch to generalize the API for user-specific 
  memory allocation so that it is more flexible in threaded environments. Thiw 
  was promised a long time ago, and never appeared...
  
. Write a function that generates random matching strings for a compiled regex.

. Write a wrapper to maintain a structure with specified runtime parameters, 
  such as recurse limit, and pass these to PCRE each time it is called. Also 
  maybe malloc and free. A user sent a prototype.
  
. Pcregrep: an option to specify the output line separator, either as a string 
  or select from a fixed list. This is not dead easy, because at the moment it 
  outputs whatever is in the input file.
  
. Improve the code for duplicate checking in pcre_dfa_exec(). An incomplete, 
  non-thread-safe patch showed that this can help performance for patterns 
  where there are many alternatives. However, a simple thread-safe 
  implementation that I tried made things worse in many simple cases, so this 
  is not an obviously good thing.
  
. Make the longest lookbehind available via pcre_fullinfo(). This is not 
  straightforward because lookbehinds can be nested inside lookbehinds. This 
  case will have to be identified, and the amounts added. This should then give 
  the maximum possible lookbehind length. The reason for wanting this is to 
  help when implementing multi-segment matching using pcre_exec() with partial
  matching and overlapping segments.
  
. PCRE cannot at present distinguish between subpatterns with different names,
  but the same number (created by the use of ?|). In order to do so, a way of 
  remembering *which* subpattern numbered n matched is needed. Bugzilla #760.

Philip Hazel
Email local part: ph10
Email domain: cam.ac.uk
Last updated: 26 September 2009
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	----------------- This file is now OBSOLETE and no longer used ----------------
20	Builducptable A Perl script that creates the contents of the ucptable.h file
21	from two Unicode data files, which themselves are downloaded
22	from the Unicode web site. Run this script in the "maint"
23	directory.
24	----------------- This file is now OBSOLETE and no longer used ----------------
25
26	GenerateUtt.py A Python script to generate part of the pcre_tables.c file
27	that contains Unicode script names in a long string with
28	offsets, which is tedious to maintain by hand.
29
30	ManyConfigTests A shell script that runs "configure, make, test" a number of
31	times with different configuration settings.
32
33	MultiStage2.py A Python script that generates the file pcre_ucd.c from three
34	Unicode data tables, which are themselves downloaded from the
35	Unicode web site. Run this script in the "maint" directory.
36	The generated file contains the tables for a 2-stage lookup
37	of Unicode properties.
38
39	README This file.
40
41	Unicode.tables The files in this directory, DerivedGeneralCategory.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,
52	to use in conjunction with the ucptest program.
53
54	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
56	versa. If its argument is a hex number such as 0x1234, it
57	outputs a list of the equivalent UTF-8 bytes. If its argument
58	is sequence of concatenated UTF-8 bytes (e.g. e188b4) it
59	treats them as a UTF-8 character and outputs the equivalent
60	code point in hex.
61
62
63	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	Note also that both the pcresyntax.3 and pcrepattern.3 man pages contain lists
87	of Unicode script names.
88
89
90	Preparing for a PCRE release
91	----------------------------
92
93	This section contains a checklist of things that I consult before building a
94	distribution for a new release.
95
96	. Ensure that the version number and version date are correct in configure.ac.
97
98	. If new build options have been added, ensure that they are added to the CMake
99	files as well as to the autoconf files.
100
101	. Run ./autogen.sh to ensure everything is up-to-date.
102
103	. Compile and test with many different config options, and combinations of
104	options. The maint/ManyConfigTests script now encapsulates this testing.
105
106	. Run perltest.pl on the test data for tests 1, 4, 6, and 11. The first two can
107	be run with Perl 5.8 or 5.10; the last two require Perl 5.10. The output
108	should match the PCRE test output, apart from the version identification at
109	the start of each test. The other tests are not Perl-compatible (they use
110	various PCRE-specific features or options).
111
112	. Test with valgrind by running "RunTest valgrind". There is also "RunGrepTest
113	valgrind", though that takes quite a long time.
114
115	. It may also useful to test with Electric Fence, though the fact that it
116	grumbles for missing free() calls can be a nuisance. (A missing free() in
117	pcretest is hardly a big problem.) To build with EF, use:
118
119	LIBS='/usr/lib/libefence.a -lpthread' with ./configure.
120
121	Then all normal runs use it to check for buffer overflow. Also run everything
122	with:
123
124	EF_PROTECT_BELOW=1 <whatever>
125
126	because there have been problems with lookbehinds that looked too far.
127
128	. Test with the emulated memmove() function by undefining HAVE_MEMMOVE and
129	HAVE_BCOPY in config.h. You may see a number of "pcre_memmove defined but not
130	used" warnings for the modules in which there is no call to memmove(). These
131	can be ignored.
132
133	. Documentation: check AUTHORS, COPYING, ChangeLog (check version and date),
134	INSTALL, LICENCE, NEWS (check version and date), NON-UNIX-USE, and README.
135	Many of these won't need changing, but over the long term things do change.
136
137	. Man pages: Check all man pages for \ not followed by e or f or " because
138	that indicates a markup error.
139
140	. When the release is built, test it on a number of different operating
141	systems if possible, and using different compilers as well. For example,
142	on Solaris it is helpful to test using Sun's cc compiler as a change from
143	gcc. Adding -xarch=v9 to the cc options does a 64-bit test, but it also
144	needs -S 64 for pcretest to increase the stack size for test 2.
145
146
147	Making a PCRE release
148	---------------------
149
150	Run PrepareRelease and commit the files that it changes (by removing trailing
151	spaces). Then run "make distcheck" to create the tarballs and the zipball.
152	Double-check with "svn status", then create an SVN tagged copy:
153
154	svn copy svn://vcs.exim.org/pcre/code/trunk \
155	svn://vcs.exim.org/pcre/code/tags/pcre-8.xx
156
157	Don't forget to update Freshmeat when the new release is out, and to tell
158	webmaster@pcre.org and the mailing list.
159
160
161	Future ideas (wish list)
162	------------------------
163
164	This section records a list of ideas so that they do not get forgotten. They
165	vary enormously in their usefulness and potential for implementation. Some are
166	very sensible; some are rather wacky. Some have been on this list for years;
167	others are relatively new.
168
169	. Optimization
170
171	There are always ideas for new optimizations so as to speed up pattern
172	matching. Most of them try to save work by recognizing a non-match without
173	having to scan all the possibilities. These are some that I've recorded:
174
175	* /((A{0,5}){0,5}){0,5}(something complex)/ on a non-matching string is very
176	slow, though Perl is fast. Can we speed up somehow? Convert to {0,125}?
177	OTOH, this is pathological - the user could easily fix it.
178
179	* Turn ={4} into ==== ? (for speed). I once did an experiment, and it seems
180	to have little effect, and maybe makes things worse.
181
182	* "Ends with literal string" - note that a single character doesn't gain much
183	over the existing "required byte" (reqbyte) feature that just remembers one
184	byte.
185
186	* These probably need to go in study():
187
188	o Remember an initial string rather than just 1 char?
189
190	o A required byte from alternatives - not just the last char, but an
191	earlier one if common to all alternatives.
192
193	o Friedl contains other ideas.
194
195	. If Perl gets to a consistent state over the settings of capturing sub-
196	patterns inside repeats, see if we can match it. One example of the
197	difference is the matching of /(main(O)?)+/ against mainOmain, where PCRE
198	leaves $2 set. In Perl, it's unset. Changing this in PCRE will be very hard
199	because I think it needs much more state to be remembered.
200
201	. Perl 6 will be a revolution. Is it a revolution too far for PCRE?
202
203	. Unicode
204
205	* Note that in Perl, \s matches \pZ and similarly for \d, \w and the POSIX
206	character classes. For the moment, I've chosen not to support this for
207	backward compatibility, for speed, and because it would be messy to
208	implement.
209
210	* A different approach to Unicode might be to use a typedef to do everything
211	in unsigned shorts instead of unsigned chars. Actually, we'd have to have a
212	new typedef to distinguish data from bits of compiled pattern that are in
213	bytes, I think. There would need to be conversion functions in and out. I
214	don't think this is particularly trivial - and anyway, Unicode now has
215	characters that need more than 16 bits, so is this at all sensible?
216
217	* There has been a request for direct support of 16-bit characters and
218	UTF-16. However, since Unicode is moving beyond purely 16-bit characters,
219	is this worth it at all? One possible way of handling 16-bit characters
220	would be to "load" them in the same way that UTF-8 characters are loaded.
221
222	. Allow errorptr and erroroffset to be NULL. I don't like this idea.
223
224	. Line endings:
225
226	* Option to use NUL as a line terminator in subject strings. This could now
227	be done relatively easily since the extension to support LF, CR, and CRLF.
228	If it is done, a suitable option for pcregrep is also required.
229
230	. Option to provide the pattern with a length instead of with a NUL terminator.
231	This affects quite a few places in the code and is not trivial.
232
233	. Catch SIGSEGV for stack overflows?
234
235	. A feature to suspend a match via a callout was once requested.
236
237	. Option to convert results into character offsets and character lengths.
238
239	. Option for pcregrep to scan only the start of a file. I am not keen - this is
240	the job of "head".
241
242	. A (non-Unix) user wanted pcregrep options to (a) list a file name just once,
243	preceded by a blank line, instead of adding it to every matched line, and (b)
244	support --outputfile=name.
245
246	. Consider making UTF-8 and UCP the default for PCRE n.0 for some n > 8.
247
248	. Add a user pointer to pcre_malloc/free functions -- some option would be
249	needed to retain backward compatibility.
250
251	. Define a union for the results from pcre_fullinfo().
252
253	. Provide a "random access to the subject" facility so that the way in which it
254	is stored is independent of PCRE. For efficiency, it probably isn't possible
255	to switch this dynamically. It would have to be specified when PCRE was
256	compiled. PCRE would then call a function every time it wanted a character.
257
258	. Wild thought: the ability to compile from PCRE's internal byte code to a real
259	FSM and a very fast (third) matcher to process the result. There would be
260	even more restrictions than for pcre_dfa_exec(), however. This is not easy.
261
262	. Should pcretest have some private locale data, to avoid relying on the
263	available locales for the test data, since different OS have different ideas?
264	This won't be as thorough a test, but perhaps that doesn't really matter.
265
266	. pcregrep: add -rs for a sorted recurse? Having to store file names and sort
267	them will of course slow it down.
268
269	. Someone suggested --disable-callout to save code space when callouts are
270	never wanted. This seems rather marginal.
271
272	. Check names that consist entirely of digits: PCRE allows, but do Perl and
273	Python, etc?
274
275	. A user suggested a parameter to limit the length of string matched, for
276	example if the parameter is N, the current match should fail if the matched
277	substring exceeds N. This could apply to both match functions. The value
278	could be a new field in the extra block.
279
280	. Callouts with arguments: (?Cn:ARG) for instance.
281
282	. A user is going to supply a patch to generalize the API for user-specific
283	memory allocation so that it is more flexible in threaded environments. Thiw
284	was promised a long time ago, and never appeared...
285
286	. Write a function that generates random matching strings for a compiled regex.
287
288	. Write a wrapper to maintain a structure with specified runtime parameters,
289	such as recurse limit, and pass these to PCRE each time it is called. Also
290	maybe malloc and free. A user sent a prototype.
291
292	. Pcregrep: an option to specify the output line separator, either as a string
293	or select from a fixed list. This is not dead easy, because at the moment it
294	outputs whatever is in the input file.
295
296	. Improve the code for duplicate checking in pcre_dfa_exec(). An incomplete,
297	non-thread-safe patch showed that this can help performance for patterns
298	where there are many alternatives. However, a simple thread-safe
299	implementation that I tried made things worse in many simple cases, so this
300	is not an obviously good thing.
301
302	. Make the longest lookbehind available via pcre_fullinfo(). This is not
303	straightforward because lookbehinds can be nested inside lookbehinds. This
304	case will have to be identified, and the amounts added. This should then give
305	the maximum possible lookbehind length. The reason for wanting this is to
306	help when implementing multi-segment matching using pcre_exec() with partial
307	matching and overlapping segments.
308
309	. PCRE cannot at present distinguish between subpatterns with different names,
310	but the same number (created by the use of ?\|). In order to do so, a way of
311	remembering which subpattern numbered n matched is needed. Bugzilla #760.
312
313	Philip Hazel
314	Email local part: ph10
315	Email domain: cam.ac.uk
316	Last updated: 26 September 2009