Commit | Line | Data |
486c8c32 |
1 | commit 7c8a36d1711b0d18bcec53042196356cab2edf34 |
2 | Author: David Golden <dagolden@cpan.org> |
3 | Date: Fri Jun 18 18:57:51 2010 -0400 |
4 | |
5 | document missing space after regex pattern in perldelta |
6 | |
7 | commit e6897b1a5db0410e387ccbf677e89fc4a1d8c97a |
8 | Author: Karl Williamson <khw@khw-desktop.(none)> |
9 | Date: Wed Jun 9 14:40:14 2010 -0600 |
10 | |
11 | Deprecate no space between pattern, following word |
12 | |
13 | This patch raises a deprecated warning on constructs like |
14 | $result = $a =~ m/$foo/sand $bar; |
15 | which means |
16 | $result = $a =~ m/$foo/s and $bar; |
17 | |
18 | commit d9bf0e0a91525fcdd8099d78b891aa20066e9d1c |
19 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
20 | Date: Fri Jun 18 09:30:14 2010 +0200 |
21 | |
22 | Fix list of constants to import to make this test pass |
23 | |
24 | commit d12d61cff231dfdae5d1887a5c3905cbc67f0168 |
25 | Author: brian d foy <bdfoy@cpan.org> |
26 | Date: Thu Jun 17 13:41:05 2010 -0700 |
27 | |
28 | * FAQ sync |
29 | |
30 | This is commit 37550b8f812e591bcd0dd869d61677dac5bda92c from the |
31 | perlfaq repository at git@github.com:briandfoy/perlfaq.git |
32 | |
33 | commit 1e6ffe563afa06bebdef40d37cf4bdae8ac5f14d |
34 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
35 | Date: Thu Jun 17 14:19:03 2010 +0200 |
36 | |
37 | Avoid warnings with undefined hash values [perl #74280] |
38 | |
39 | commit 30fcd6c4143961133edf166c63dcc423fbcdb973 |
40 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
41 | Date: Thu Jun 17 11:40:36 2010 +0200 |
42 | |
43 | Deparse correctly "no VERSION" [perl #75482] |
44 | |
45 | commit 313b38e51ccc38cdb7b246f596287a755f8f108b |
46 | Author: Nicholas Clark <nick@ccl4.org> |
47 | Date: Thu Jun 17 17:54:25 2010 +0100 |
48 | |
49 | Avoid compiler warnings in Perl_foldEQ_utf8, spotted by Jerry D. Hedden. |
50 | |
51 | commit 2d7e78b1bd24d3af5881f95a996405c9c11b1d59 |
52 | Author: Nicholas Clark <nick@ccl4.org> |
53 | Date: Thu Jun 17 15:21:24 2010 +0100 |
54 | |
55 | Avoid creating @EXPORT_FAIL in every package using Exporter. |
56 | |
57 | Previously, if package INKLE_KLINK is an Exporter, then the Exporter code's |
58 | symbolic lookup of \@{"INKLE_KLINK::EXPORT_FAIL"} would cause both that array |
59 | and the enclosing typeglob to be instantiated. Now the typeglob and array are |
60 | only created if present. (Strictly, if there's something in INKLE_KLINK's |
61 | symbol table for EXPORT_FAIL. But likely that will only ever be the array.) |
62 | |
63 | This saves about 200 bytes per package that uses Exporter but does not need an |
64 | @EXPORT_FAIL. |
65 | |
51cf6479 |
66 | commit 79a921544258c0f3466f44738c4f0d24399782dd # DOCed |
486c8c32 |
67 | Author: Nicholas Clark <nick@ccl4.org> |
68 | Date: Thu Jun 17 10:46:37 2010 +0100 |
69 | |
70 | PERL_IMPLICIT_SYS also needs thread context for safesysfree() |
71 | |
72 | PERL_TRACK_MEMPOOL needs it to work without -DDEBUGGING. |
73 | |
74 | Fixes for 0cb20dae370512c6 not addressed by 1f4d2d4e2e4bb7bb. |
75 | |
51cf6479 |
76 | commit d9f81b50694a810f1b920a55481bdf183181048c # DOCed |
486c8c32 |
77 | Author: Nicholas Clark <nick@ccl4.org> |
78 | Date: Wed Jun 16 20:47:22 2010 +0100 |
79 | |
80 | Stop using WITH_THR and WITH_THX, as they were never necessary here. |
81 | |
51cf6479 |
82 | commit 1f4d2d4e2e4bb7bbbd526ea0b9d5592742433f9a # DOCed |
486c8c32 |
83 | Author: Nicholas Clark <nick@ccl4.org> |
84 | Date: Wed Jun 16 09:56:59 2010 +0100 |
85 | |
86 | PERL_IMPLICIT_SYS also needs thread context for the *alloc success paths. |
87 | |
88 | 0cb20dae370512c6 was a bit to aggressive in its deferral of dTHX. |
89 | |
51cf6479 |
90 | commit 0cb20dae370512c655eb665a7d5089db2819e862 # DOCed |
486c8c32 |
91 | Author: Nicholas Clark <nick@ccl4.org> |
92 | Date: Tue Jun 15 21:26:14 2010 +0100 |
93 | |
94 | In Perl_safesys{c,m,re}alloc(), defer dTHX for non-DEBUGING builds. |
95 | |
96 | Under ithreads, dTHX expands to pthread_getspecific() [or something similarly |
97 | expensive], which the compiler can't optimise away. However, its return value |
98 | isn't needed unless the allocation fails. So defer the call, hence avoiding |
99 | it entirely on a successful allocation. |
100 | |
101 | DEBUGING builds require the value of dTHX for debugging purposes, so we can't |
102 | postpone it for them. Unthreaded builds were never affected as they don't use |
103 | thread local storage for the interpreter context. |
104 | |
a59e2586 |
105 | commit 0a18a49b281a5a76e75de77e45ee27ad1b807bb2 # DOCed |
91c261ec |
106 | Author: Maik Hentsche <maik@mm-double.de> |
107 | Date: Tue Jun 15 17:19:41 2010 +0200 |
108 | |
109 | Attempt at improving the perlipc docs |
110 | |
e34085e9 |
111 | commit ee84a7925f473937ece8b2db88bffc4400a8cabe # NODOC |
91c261ec |
112 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
113 | Date: Tue Jun 15 14:48:34 2010 +0200 |
114 | |
115 | Revert "It may not be obvious to a new Perl programmer that $#ary is explained in perldata.pod (see http://stackoverflow.com/questions/2997042/ )." |
116 | |
117 | This reverts commit db2ed3fb308b082577e9315ba52b2cb266a2bd0a. |
118 | |
5062a22f |
119 | commit 4c8a458a2f2be4d98cd2b8873f15b800e9382262 # NODOC |
91c261ec |
120 | Author: Karl Williamson <khw@khw-desktop.(none)> |
121 | Date: Wed Jun 2 15:07:29 2010 -0600 |
122 | |
123 | uc(): Handle Greek YPOGEGRAMMENI |
124 | |
125 | Unicode contains two context-sensitive case-changing rules. This patch |
126 | enables one of them, dealing with the Greek YPOGEGRAMMENI. The code had |
127 | been #ifdef'd out, so the changes here are more than what the diff |
128 | shows. The reason it was #ifdef'd out was because more research was |
129 | needed to verify that it was correct, which I have now done, and think |
130 | it is. |
131 | |
132 | The issue is we may just be uppercasing a portion of the context, so |
133 | don't have complete knowledge of what should be done. This patch causes |
134 | us to move the ypogegrammeni to as far right as it should go, or to the |
135 | end of the context we know about, whichever comes first. That's the |
136 | best we can do. If it really should be moved further to the right, |
137 | there's no way we can do it, because the user has not called uc() |
138 | with the full information needed. So, in all cases, this is better than |
139 | just leaving it where it was in the input. Also, the applicable context |
140 | is limited to a logical character, that matched by /\X/, so if the user |
141 | is calling uc() on a subset of a logical character, it really is their |
142 | mistake. |
143 | |
5062a22f |
144 | commit cc70200b5f52ae10e2985e4d11670b3b76984b4d # NODOC |
91c261ec |
145 | Author: Karl Williamson <khw@khw-desktop.(none)> |
146 | Date: Wed Jun 2 15:04:16 2010 -0600 |
147 | |
148 | t/uni/case.pl: allow extra tests of different form |
149 | |
150 | Add an option to case.pl so that the callers can run extra tests that |
151 | don't fit in with it's format. They just pass the number run to it, and |
152 | it adjusts the plan accordingly. |
153 | |
5062a22f |
154 | commit f686127b43eb3c08a5cb1fccce5ef34212234df5 # NODOC |
91c261ec |
155 | Author: Salvador Fandino <sfandino@yahoo.com> |
156 | Date: Tue Jun 15 11:31:04 2010 +0200 |
157 | |
158 | Add support for more TCP_ constants to Socket |
159 | |
160 | (plus bump version to 1.89) |
161 | |
5062a22f |
162 | commit 5621738aceb2535b67ff880f9ef107b9a0f8389e # NODOC |
91c261ec |
163 | Author: Karl Williamson <khw@khw-desktop.(none)> |
164 | Date: Sun May 30 21:34:53 2010 -0600 |
165 | |
166 | Remove debugging statement from mktables. |
167 | |
168 | A debugging statement was inadvertently left in which causes an |
169 | extraneous warning. |
170 | |
759d4a71 |
171 | commit db2ed3fb308b082577e9315ba52b2cb266a2bd0a # ALREADY |
91c261ec |
172 | Author: Sinan Unur <sinan@unur.com> |
173 | Date: Tue Jun 8 18:53:27 2010 -0400 |
174 | |
175 | It may not be obvious to a new Perl programmer that $#ary is explained in perldata.pod (see http://stackoverflow.com/questions/2997042/ ). |
176 | |
177 | A possible solution is to place an entry in perlvar.pod pointing out |
178 | that $# was removed and pointing the user to perldata.pod for the |
179 | explanation of $#ary usage. |
180 | |
f6374283 |
181 | commit 9b58b542ddb7ea24d774e7f780c8a19127be1f5a # DOCed |
91c261ec |
182 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
183 | Date: Tue Jun 15 10:16:35 2010 +0200 |
184 | |
185 | usemymalloc slows down newer openbsd [perl #75742] |
186 | |
291e5469 |
187 | commit c35588eb09d732b01769aad7cf303e81b9417302 # NODOC |
91c261ec |
188 | Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk> |
189 | Date: Mon Jun 14 10:07:27 2010 +0100 |
190 | |
191 | Rename copyright section of Module::CoreList to LICENSE, fixes for my kwalitee score. |
192 | |
fb2ff3cb |
193 | commit 742adbff4f079f203897faa086d954cc3a26460a # DOCed |
91c261ec |
194 | Author: Andreas J. Koenig <andk@cpan.org> |
195 | Date: Mon Jun 14 09:46:15 2010 +0100 |
196 | |
197 | Update CPAN.pm to 1.94_57 |
198 | |
199 | 2010-05-24 Andreas J. Koenig <andk@cpan.org> |
200 | |
201 | * release 1.94_57 |
202 | |
203 | * bugfix: treat modules correctly that are deprecated in perl 5.12. |
204 | |
205 | * bugfix: RT #57482 and #57788 revealed that configure_requires |
206 | implicitly assumed build_requires instead of normal requires. (Reported |
207 | by Andrew Whatson and Father Chrysostomos respectively) |
208 | |
209 | * testfix: solaris should run the tests without expect because (some?) |
210 | solaris have a broken expect |
211 | |
212 | * testfix: run tests with cache_metadata off to prevent spill over |
213 | effects from previous test runs |
214 | |
215 | Signed-off-by: Chris 'BinGOs' Williams <chris@bingosnet.co.uk> |
216 | |
46584c29 |
217 | commit 086d291379a28ceb3cd7cc6416747be8c426476b # NODOC |
91c261ec |
218 | Author: Nicholas Clark <nick@ccl4.org> |
219 | Date: Mon Jun 14 09:44:49 2010 +0200 |
220 | |
221 | Note why pp_tie can't use call_method() for a package name. |
222 | |
223 | Plus a test that would fail if it did. |
224 | |
46584c29 |
225 | commit f4adce6b773b2b6071e0d1627f4dce6ec0c3caa7 # NODOC |
91c261ec |
226 | Author: Nicholas Clark <nick@ccl4.org> |
227 | Date: Sun Jun 13 16:25:53 2010 +0200 |
228 | |
229 | Merge flags and argc parameters to S_tied_handle_method(). |
230 | |
231 | This generates slightly smaller object code overall, which means that the "hot" |
232 | code (the non-overloaded paths through the ops) will be smaller, and hence more |
233 | likely to stay in the CPU cache. |
234 | |
46584c29 |
235 | commit 74f0b5509afd269c59f1396fde551295bbeec7d5 # NODOC |
91c261ec |
236 | Author: Nicholas Clark <nick@ccl4.org> |
237 | Date: Sun Jun 13 15:02:14 2010 +0200 |
238 | |
239 | In S_tied_handle_method() default to mortalizing extra arguments. |
240 | |
241 | Convert the gimme argument to a flags argument, and add a flag bit to signal |
242 | that mortalization is not required. Only "BINMODE" needs this. |
243 | |
46584c29 |
244 | commit 0240605ecd7a040cc6058409c65a3ebc6b7d489b # NODOC |
91c261ec |
245 | Author: Nicholas Clark <nick@ccl4.org> |
246 | Date: Sun Jun 13 12:38:16 2010 +0200 |
247 | |
248 | Add a gimme parameter to S_tied_handle_method(). |
249 | |
250 | This allows "GETC" to use it. |
251 | |
46584c29 |
252 | commit bc0c81caab3813b2d61b70f94e5075bbf3a3ef69 # NODOC |
91c261ec |
253 | Author: Nicholas Clark <nick@ccl4.org> |
254 | Date: Sun Jun 13 12:12:43 2010 +0200 |
255 | |
256 | Change S_tied_handle_method() to varargs to allow extra SV parameters. |
257 | |
258 | This enables "BINMODE", "EOF" and "SYSSEEK" to use it. |
259 | |
46584c29 |
260 | commit ebc1fde647268c1d49a6096baf6ca8a708363f5b # NODOC |
91c261ec |
261 | Author: Nicholas Clark <nick@ccl4.org> |
262 | Date: Sun Jun 13 10:34:56 2010 +0200 |
263 | |
264 | Merge simple tied handle method calls into S_tied_handle_method(). |
265 | |
46584c29 |
266 | commit c33ef3ac654cbe35caea1d36f34c68f0e4a134ba # NODOC |
91c261ec |
267 | Author: Nicholas Clark <nick@ccl4.org> |
268 | Date: Sat Jun 12 22:29:03 2010 +0200 |
269 | |
270 | For tied file handle calls, use PUSH* when we know that the stack has space. |
271 | |
272 | Previously the code was using XPUSH* in many cases, often to push back values |
273 | recently popped off the stack. |
274 | |
46584c29 |
275 | commit 2addaaf3b439e1adb64bd833a257ff963c73f45a # NODOC |
91c261ec |
276 | Author: Nicholas Clark <nick@ccl4.org> |
277 | Date: Sat Jun 12 20:51:19 2010 +0200 |
278 | |
279 | Remove unecessary EXTEND()s from pp_sys.c |
280 | |
281 | If we've just POP*ed something from the stack, it will already have space. |
282 | If we conditionally POP* something from the stack, avoid the EXTEND() in that |
283 | path. Also, use PUSH() rather than XPUSH() where we know that the stack is large |
284 | enough. Changes to pp_close, pp_binmode, pp_enterwrite, pp_readlink and pp_alarm. |
285 | |
46584c29 |
286 | commit 2116934e2bf4f3c1445ea039e29ff377c3de2648 # NODOC |
91c261ec |
287 | Author: Nicholas Clark <nick@ccl4.org> |
288 | Date: Sun Jun 13 08:50:00 2010 +0100 |
289 | |
290 | In pp_eof, ensure stack space for the second argument to the tied EOF method. |
291 | |
292 | Commit 32e653230c7ccc7f added this second argument, but didn't verify that the |
293 | stack always had sufficient space for it. |
294 | |
46584c29 |
295 | commit b0c1862166fbc7bf02b768bf81c9e1e77b5220a8 # NODOC |
91c261ec |
296 | Author: David Golden <dagolden@cpan.org> |
297 | Date: Wed Jun 9 14:53:15 2010 -0400 |
298 | |
299 | remove trailing spaces in perlvar.pod |
300 | |
46584c29 |
301 | commit 9bf2270250326fb85445d6849ed84a94434dd12c # NODOC |
91c261ec |
302 | Author: Ævar Arnfjörð Bjarmason <avar@cpan.org> |
303 | Date: Thu May 13 22:02:25 2010 +0000 |
304 | |
305 | Link to perlport/PLATFORMS from the $^O docs |
306 | |
307 | I was looking for some actual values of $^O the other day and started |
308 | ack-ing t/. Turns out that this was documented in perlport. Link to |
309 | the relevant section to make it easy to find. |
310 | |
46584c29 |
311 | commit b5f55170e8ef2a91497f68ff0af6ff6cded9f433 NODOC |
91c261ec |
312 | Author: Nicholas Clark <nick@ccl4.org> |
313 | Date: Sat Jun 12 21:13:14 2010 +0100 |
314 | |
315 | Fix edge case in pp_eof where the stack extent was not checked. |
316 | |
317 | Analogous to pp_getc and pp_tell in ac3697cd90b00fae, pp_eof has a conditional |
318 | POP from the stack, but an unconditional PUSH to the stack, but no check that |
319 | the stack had space for the PUSH. This bug has been present since perl 5.000. |
320 | |
46584c29 |
321 | commit ac3697cd90b00fae88e4f19931af920bc552e2b8 NODOC |
91c261ec |
322 | Author: Nicholas Clark <nick@ccl4.org> |
323 | Date: Sat Jun 12 20:43:37 2010 +0100 |
324 | |
325 | Fix edge cases in pp_getc and pp_tell where the stack extent was not checked. |
326 | |
327 | Both conditionally POP a GV from the stack, but always PUSH a return value to |
328 | it. For the case where they did not POP the GV, they made no check that the |
329 | stack had space for the PUSH. This bug has been present since perl 5.000. |
330 | |
3d83d887 |
331 | commit 0f7807cda10b798e3d16820699b3a52f13a8de16 NODOC |
91c261ec |
332 | Author: Nicholas Clark <nick@ccl4.org> |
333 | Date: Sat Jun 12 19:05:51 2010 +0100 |
334 | |
335 | Better diagnostics on detecting case sensitive file name clashes. |
336 | |
2f787fc8 |
337 | commit f06daabb40d6dee8a7eda2c79bd7ee3e0b6275ec DOCed |
91c261ec |
338 | Author: Chris 'BinGOs' Williams <chris@bingosnet.co.uk> |
339 | Date: Fri Jun 11 07:59:48 2010 +0100 |
340 | |
341 | Added new files I forgot to add for the Thread-Semaphore update |
342 | |
2f787fc8 |
343 | commit 04febe174c342d1018822ab2c67fb381bb88c55f DOCed |
91c261ec |
344 | Author: Jerry D. Hedden <jdhedden@cpan.org> |
345 | Date: Fri Jun 11 07:57:34 2010 +0100 |
346 | |
347 | [PATCH-revised] Upgrade to Thread-Semaphore 2.11 |
348 | |
349 | Added new methods ->down_nb() and ->down_force() at the suggestion of |
350 | Rick Garlick. |
351 | |
352 | Refactored methods to skip argument validation when no argument is supplied. |
353 | |
354 | Signed-off-by: Chris 'BinGOs' Williams <chris@bingosnet.co.uk> |
355 | |
12c7294b |
356 | commit b00fc8d41de1694e3f43042249a46dd72bd0163a # NODOC |
91c261ec |
357 | Author: Nicholas Clark <nick@ccl4.org> |
358 | Date: Thu Jun 10 10:01:18 2010 +0200 |
359 | |
360 | Parameters for * in *printf must be int - add a cast to ensure this. |
361 | |
362 | Fixes a (legitimate) compiler warning present since 6e1bad6cc227c8e8. |
363 | |
12c7294b |
364 | commit c2485e0c886f5e7d6fd27b65f085b0ba231ca557 # NODOC |
91c261ec |
365 | Author: Nicholas Clark <nick@ccl4.org> |
366 | Date: Wed Jun 9 19:53:20 2010 +0100 |
367 | |
368 | Reorder CVf_* flags to be numerically contiguous again. |
369 | |
370 | The removal of CVf_ASSERTION in 584420f022db5722 and CVf_LOCKED in |
371 | e95ab0c0d2aa1b35 left two gaps in the sequence of bits in use. |
372 | |
f573233a |
373 | commit ede3a6167fdc98bf29305444db6ff93c33d8155d # NODOC |
91c261ec |
374 | Author: David Golden <dagolden@cpan.org> |
375 | Date: Wed Jun 9 11:47:41 2010 -0400 |
376 | |
377 | added Chas. Owens to AUTHORS |
378 | |
b63e0575 |
379 | commit 7f315d2e5ecf9c2453606c348568dc73e5e4d390 # DOCed |
91c261ec |
380 | Author: Chas. Owens <chas.owens@gmail.com> |
381 | Date: Wed Jun 9 11:43:46 2010 -0400 |
382 | |
383 | Document $# and $* as removed |
384 | |
385 | See email thread for reference: |
386 | |
387 | http://www.nntp.perl.org/group/perl.perl5.porters/2010/06/msg160812.html |
388 | |
b08c8d49 |
389 | commit 1f3ebc3b353ab41acc3013ef04412dda481b4282 # NODOC |
91c261ec |
390 | Author: Abigail <abigail@abigail.be> |
391 | Date: Tue Jun 8 16:35:52 2010 +0200 |
392 | |
393 | Resolve issue #74974: document File::Copy::copy(FILE, DIR) |
394 | |
e9f29d30 |
395 | commit c1bd236350b343c4ef921b81c45eabb4ecc82569 # DOCed |
91c261ec |
396 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
397 | Date: Tue Jun 8 15:50:05 2010 +0200 |
398 | |
399 | Fix long-standing typos in Policy_sh.SH |
400 | |
401 | This patch fixes several long-standing typos and naming confusions in |
402 | Policy_sh.SH, standardizing on the variable names used in config.sh. |
403 | This will change the behavior of Policy.sh if you happen to have been |
404 | accidentally relying on the Policy.sh incorrect behavior. I'd appreciate |
405 | feedback from anyone using Policy.sh to be sure nothing is broken by |
406 | this change. Thanks to Stoned Elipot for pointing this out. |
407 | |
408 | Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl> |
409 | |
3ed88686 |
410 | commit e6eb90205a8cd8f42cd4d1226ea369f3ff1a14ae # DOCed |
91c261ec |
411 | Author: David Mitchell <davem@iabyn.com> |
412 | Date: Tue Jun 8 00:08:19 2010 +0100 |
413 | |
414 | perlbug: always allow change of sender address |
415 | |
416 | The code that guessed a default sender address would under some |
417 | circumstances not display this guess to the user with the option to change |
418 | it. Specifically, if $Config{cf_me} eq {login id of the user running |
419 | perlbug}. |
420 | |
421 | In my case, cf_email got set at build time (Configure default) of |
422 | 'davem@pigeon.(none)', which was then silently used. Not good! |
423 | |
3ed88686 |
424 | commit 4ea8010ab3e95796683d1a69375c8df8c4bea3fb # NODOC |
91c261ec |
425 | Author: David Mitchell <davem@iabyn.com> |
426 | Date: Mon Jun 7 23:28:26 2010 +0100 |
427 | |
428 | perlbug -d,-v: fix uninit value warnings |
429 | |
6afcecf3 |
430 | commit 5f3e44b689cab8bd7cec81aa6fc21c2ec5b3489f # ALREADY |
91c261ec |
431 | Author: David Mitchell <davem@iabyn.com> |
432 | Date: Mon Jun 7 22:18:35 2010 +0100 |
433 | |
434 | document revert of 'local $tied' change |
435 | |
6afcecf3 |
436 | commit 89adf4e8e0de17fedf78328b32920f47635850d0 # ALREADY |
91c261ec |
437 | Author: David Mitchell <davem@iabyn.com> |
438 | Date: Mon Jun 7 22:11:42 2010 +0100 |
439 | |
440 | Revert "make 'local $tied' untied" |
441 | |
442 | This reverts commit 191ad7eff570fc96c93993e4358f83e2033365d6. |
443 | |
444 | Some modules (e.g. File::chdir) relied on the current behaviour of |
445 | local $tied_scalar, so lets leave things as-is for now. See |
446 | |
447 | http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2010-05/msg00627.html |
448 | |
e31fcbd3 |
449 | commit e1137bc782e702f287ec7909f121d66bfb4707ba # DOCed |
91c261ec |
450 | Author: Sullivan Beck <sbeck@cpan.org> |
451 | Date: Mon Jun 7 21:58:34 2010 +0100 |
452 | |
453 | PATCH: Bump Locale-Codes from 3.12 to 3.13 |
454 | |
455 | Attached is a patch to upgrade the Locale-Codes distribution (containing |
456 | the core modules Locale::Country, Locale::Language, and |
457 | Locale::Currency) to the most recent version. |
458 | |
459 | ==== |
460 | Background: |
461 | |
462 | The core modules Locale::Country, Locale::Language, and Locale::Currency |
463 | (all part of the Locale-Codes distribution) should be updated on a |
464 | regular basis. They contain "codes" from various internet standards |
465 | which change over time. |
466 | |
467 | I plan on releasing new versions twice a year to keep the codes |
468 | up-to-date. At this point, I'm not planning on any significant code |
469 | changes (other than bug fixes), so the only significant changes |
470 | between releases should be to update the codes. |
471 | |
019b5979 |
472 | commit 16d811d41019c65e82b767b91041c1bebb9b4407 # NODOC |
91c261ec |
473 | Author: David Mitchell <davem@iabyn.com> |
474 | Date: Mon Jun 7 18:23:43 2010 +0100 |
475 | |
476 | identify a few vars in intrpvar.h |
477 | |
019b5979 |
478 | commit f2ed9b323720c96381a3407a26ac73176799e8ee # NODOC |
91c261ec |
479 | Author: Karl Williamson <khw@khw-desktop.(none)> |
480 | Date: Sun Jun 6 08:50:24 2010 -0600 |
481 | |
482 | regexec.c: change names of two vars for clarity |
483 | |
484 | do_utf8 is changed to utf8_target |
485 | UTF is changed to UTF_PATTERN |
486 | |
487 | This will help me keep track of the fact that there are four possible |
488 | combinations of these, and that ! do_utf8 doesn't necessarily mean don't |
489 | do utf8. |
490 | |
79fd2aac |
491 | commit 60eaec425bbc5e93d5dab2c98aa44af5d0baeb52 # DOCed |
91c261ec |
492 | Author: David Golden <dagolden@cpan.org> |
493 | Date: Mon Jun 7 07:07:10 2010 -0400 |
494 | |
495 | Remove -w recommendation in perl -h |
496 | |
497 | Abigail already said it best: |
498 | |
499 | I do not think 'perl -h' is the place to give recommendations on how |
500 | code should be written. 'perl -h' gives a list and a brief explanation |
501 | of the command line switches. IMO, it should do just that, and nothing |
502 | else. |
503 | |
15136b62 |
504 | commit 98ccfbbf7f3673c554b54933d94b68014231a32e # NODOC |
91c261ec |
505 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
506 | Date: Mon Jun 7 12:19:32 2010 +0200 |
507 | |
508 | Spell-check fixes in ext |
509 | |
15136b62 |
510 | commit fb78fdcd12d30a6a41c12c697ef31593f4bc5d94 # NODOC |
91c261ec |
511 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
512 | Date: Mon Jun 7 12:17:22 2010 +0200 |
513 | |
514 | Spell-check fixes in lib |
515 | |
15136b62 |
516 | commit 2108160ef633eaf83b7335df0336ca1f5e0e0462 # NODOC |
91c261ec |
517 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
518 | Date: Mon Jun 7 12:17:09 2010 +0200 |
519 | |
520 | Spell check fixes in CPAN |
521 | |
522 | already fixed upstream |
523 | |
2002ccfc |
524 | commit 4984aa345cae6d92b88b5afcf09ce72cb121191a # DOCed |
91c261ec |
525 | Author: Father Chrysostomos <sprout@cpan.org> |
526 | Date: Mon Jun 7 12:07:11 2010 +0200 |
527 | |
528 | glob crashes when %File::Glob:: is empty |
529 | |
530 | In 5.6.2, a failure to find a CORE::GLOBAL::glob after loading |
531 | File::Glob would result in a fallback to external glob via pp_glob. |
532 | Now it crashes. |
533 | |
534 | The attached patch should fix this. |
535 | |
343f0970 |
536 | commit b3fe9f3f93e825d0b858b198dad1d2da4b7b4a95 # NODOC |
91c261ec |
537 | Author: Karl Williamson <khw@khw-desktop.(none)> |
538 | Date: Wed Jun 2 12:33:56 2010 -0600 |
539 | |
540 | perlhack: Add reference to TAP for test protocol |
541 | |
343f0970 |
542 | commit 195c30cea8a0d6f651c1a4c8bcc30506f46cbbaa # NODOC |
91c261ec |
543 | Author: Karl Williamson <khw@khw-desktop.(none)> |
544 | Date: Wed Jun 2 12:32:19 2010 -0600 |
545 | |
546 | perlhack: fix formatting issues |
547 | |
548 | Change some lines so won't overflow 80 column width; make a link. |
549 | |
bba1ab09 |
550 | commit 66fd7fd0f5a20fad5e7cda931bc1dd21789dc9b2 # NODOC |
91c261ec |
551 | Author: Karl Williamson <khw@khw-desktop.(none)> |
552 | Date: Sun May 30 14:11:52 2010 -0600 |
553 | |
554 | PATCH: uniprops.t take advantage of EBCDIC test.pl |
555 | |
556 | This patch removes the partial solution to testing on EBCDIC platforms |
557 | that was in uniprops.t (generated by mktables), and replaces it with the |
558 | simple complete solution now in test.pl |
559 | |
f3e45b20 |
560 | commit d02d6d97d5eefad4e164003699595f59abb06506 # NODOC |
91c261ec |
561 | Author: David Mitchell <davem@iabyn.com> |
562 | Date: Sun Jun 6 21:09:22 2010 +0100 |
563 | |
564 | reduce size of regmatch_state.u.curlyx by 2 words |
565 | |
f3e45b20 |
566 | commit 6dd2be570d715119e05672f6f0266d924022b65a # NODOC |
91c261ec |
567 | Author: David Mitchell <davem@iabyn.com> |
568 | Date: Sun Jun 6 18:48:49 2010 +0100 |
569 | |
570 | micro-optimise a bit of trie code |
571 | |
bd03357c |
572 | commit 0bf6a637d180d2ff237212513f8b816d40ead86a # NODOC |
91c261ec |
573 | Author: Nicholas Clark <nick@ccl4.org> |
574 | Date: Sun Jun 6 11:50:56 2010 +0200 |
575 | |
576 | Test the return value of push and unshift. |
577 | |
bd03357c |
578 | commit af5c7f63b8ce18483dfeb72829f70ee387ece366 # NODOC |
91c261ec |
579 | Author: Nicholas Clark <nick@ccl4.org> |
580 | Date: Sun Jun 6 11:46:16 2010 +0200 |
581 | |
582 | Convert tiearray.t to test.pl. |
583 | |
0b09c0bc |
584 | commit 4c1b470c26738a632b119c2932798183ae51b3ec # NODOC |
91c261ec |
585 | Author: Karl Williamson <khw@khw-desktop.(none)> |
586 | Date: Sat Jun 5 11:25:58 2010 -0600 |
587 | |
588 | Change regexec.c to use new foldEQ functions |
589 | |
bfedc2a3 |
590 | commit e6226b18246ce7d24213c41123114ac7967ed04f # DOCed |
91c261ec |
591 | Author: Karl Williamson <khw@khw-desktop.(none)> |
592 | Date: Sat Jun 5 11:12:47 2010 -0600 |
593 | |
594 | Change name of ibcmp to foldEQ |
595 | |
596 | As discussed on p5p, ibcmp has different semantics from other cmp |
597 | functions in that it is a binary instead of ternary function. It is |
598 | less confusing then to have a name that implies true/false. |
599 | |
600 | There are three functions affected: ibcmp, ibcmp_locale and ibcmp_utf8. |
601 | ibcmp is actually equivalent to foldNE, but for the same reason that things |
602 | like 'unless' and 'until' are cautioned against, I changed the functions |
603 | to foldEQ, so that the existing names, like ibcmp_utf8 are defined as |
604 | macros as being the complement of foldEQ. |
605 | |
606 | This patch also changes the one file where turning ibcmp into a macro |
607 | causes problems. It changes it to use the new name. It also documents |
608 | for the first time ibcmp, ibcmp_locale and their new names. |
609 | |
43b8024e |
610 | commit d51c1b21fa08933083b4723794b68ac09a7a248b # NODOC |
91c261ec |
611 | Author: Karl Williamson <khw@khw-desktop.(none)> |
612 | Date: Sat Jun 5 11:08:25 2010 -0600 |
613 | |
614 | utf8.c: further doc tweaks |
615 | |
43b8024e |
616 | commit 48ef279ea70605b40a74c3e2c4a5c4ca2cf48054 # NODOC |
91c261ec |
617 | Author: Karl Williamson <khw@khw-desktop.(none)> |
618 | Date: Fri Jun 4 12:04:45 2010 -0600 |
619 | |
620 | utf8.c: Modify doc comment; change whitespace |
621 | |
622 | This removes the comment about the function name, and converts tabs to |
623 | blanks throughout the function, as so much of it is changing already. |
624 | |
625 | It also removes trailing whitespace in other lines of the file. |
626 | |
4e8a1019 |
627 | commit 8b35872c947d4c76532f1e4874411afa9125575d # DOCed |
91c261ec |
628 | Author: Karl Williamson <khw@khw-desktop.(none)> |
629 | Date: Tue May 25 11:18:42 2010 -0600 |
630 | |
631 | Revamp ibcmp_utf8 for performance and clarity |
632 | |
633 | I had a hard time understanding how this routine worked; there were no |
634 | comments. In figuring it out, I discovered it could be made more |
635 | efficient. This routine is called over and over in the innermost loops |
636 | in regex matching, so efficiency is a concern. |
637 | |
638 | Setup is done once before the main while loop so that it now has two |
639 | conditions instead of eight. The loop was rearranged slightly to be |
640 | smaller and a couple of unneeded assignments to temporaries were |
641 | removed, and recomputation of some values was avoided. Several other |
642 | small efficiency changes were made. |
643 | |
644 | Several asserts had been commented out, saying that they make tests |
645 | fail. But they no longer do, at least on my platform. There was a |
646 | reason that they were asserts to begin with, and that is they denoted an |
647 | insane or trivial condition. Apparently there have been fixes to the |
648 | other code calling this, so I re-enabled them. |
649 | |
650 | The names of several variables were changed to be less confusing; hence |
651 | f1 means the fold buffer for string 1 whereas it used to mean its goal, |
652 | which is now g1. |
653 | |
654 | The leading indent was changed from 5 to 4 blanks. I made enough |
655 | other changes that I didn't submit this as a separate commit |
656 | |
c99930b9 |
657 | commit 970ea3cb860f5a38a7f7582cc02c1b88c0bf4b0f # NODOC |
91c261ec |
658 | Author: Karl Williamson <khw@khw-desktop.(none)> |
659 | Date: Tue May 25 11:17:22 2010 -0600 |
660 | |
661 | Clarify some documentation |
662 | |
cee1191e |
663 | commit 5ac3629776cb773f868455ce02860eafa022c003 # NODOC |
91c261ec |
664 | Author: David Mitchell <davem@iabyn.com> |
665 | Date: Sat Jun 5 13:44:04 2010 +0100 |
666 | |
667 | silence some warnings in hv.c |
668 | |
669 | Since de0a224a057997a65d38856f1981702fca5d7c18, xhv_keys and xhv_max |
670 | are the same type, so no casting needed |
671 | |
cee1191e |
672 | commit 4a75fa1ae0acfb6204ce925df66ec1e7f85767ce # NODOC |
91c261ec |
673 | Author: David Mitchell <davem@iabyn.com> |
674 | Date: Sat Jun 5 13:12:11 2010 +0100 |
675 | |
676 | add handy note on vtable fields |
677 | |
cee1191e |
678 | commit 1a1a5af7b042b1599e52170095e55b02673d4c53 # NODOC |
91c261ec |
679 | Author: David Mitchell <davem@iabyn.com> |
680 | Date: Sat Jun 5 12:46:09 2010 +0100 |
681 | |
682 | update Perl_magic_methcall description |
683 | |
684 | it's now a varargs function |
685 | |
eb5d070d |
686 | commit 7c75014e4b3bd5ebe368b5d6b981f310525d1389 # DOCed |
91c261ec |
687 | Author: David Mitchell <davem@iabyn.com> |
688 | Date: Fri Jun 4 23:09:21 2010 +0100 |
689 | |
690 | fix for RT #8438: $tied->() doesn't call FETCH |
691 | |
692 | pp_entersub checked for ROK *before* calling magic. If the tied scalar |
693 | already had ROK set (perhaps from a previous time), then get magic (and |
694 | hence FETCH) wasn't called. |
695 | |
16433fde |
696 | commit dfae30232f1b277231b0dee813acbeca0b958afd # NODOC |
91c261ec |
697 | Author: Craig A. Berry <craigberry@mac.com> |
698 | Date: Thu Jun 3 22:17:39 2010 -0500 |
699 | |
700 | Avoid overflow warning in chop.t. |
701 | |
702 | It overflows at compile time on 32-bit architectures, so the skip |
703 | isn't enough. |
704 | |
90f79897 |
705 | commit 613a9ec2cfb133a94ef11f3e50bdfef024885d96 # NODOC |
91c261ec |
706 | Author: David Mitchell <davem@iabyn.com> |
707 | Date: Fri Jun 4 21:30:18 2010 +0100 |
708 | |
709 | fix indendation of DM_* flag definitions |
710 | |
90f79897 |
711 | commit 6ca8229ddb2de12c1f8f877d70989420b70a8072 # NODOC |
91c261ec |
712 | Author: David Mitchell <davem@iabyn.com> |
713 | Date: Fri Jun 4 21:28:20 2010 +0100 |
714 | |
715 | express DM_[GU]ID flags in terms of components |
716 | |
90f79897 |
717 | commit 354b0578ec63c058cd73f018f484808b22cc8631 # NODOC |
91c261ec |
718 | Author: David Mitchell <davem@iabyn.com> |
719 | Date: Fri Jun 4 21:25:07 2010 +0100 |
720 | |
721 | rename DM_ARRAY flag to DM_ARRAY_ISA |
722 | |
723 | This better represents its current role as specifically delaying magic on |
724 | @ISA as opposed to a general array magic delay mechanism. |
725 | |
4ff81d7c |
726 | commit 3a19377ba2315fce9354aa3f06bafbbbab740f1b # NODOC |
91c261ec |
727 | Author: David Mitchell <davem@iabyn.com> |
728 | Date: Fri Jun 4 21:16:48 2010 +0100 |
729 | |
730 | belatedly add a test for RT #51636: |
731 | |
732 | segmentation fault with array ties |
733 | |
b1e21aac |
734 | commit 8ef242405b8c660c02e953dbc987fbc06897af10 # NODOC |
91c261ec |
735 | Author: David Mitchell <davem@iabyn.com> |
736 | Date: Fri Jun 4 21:01:43 2010 +0100 |
737 | |
738 | Revert "Re: [perl #51636] segmentation fault with array ties" |
739 | |
740 | This reverts commit 90630e3c741716305d7f1da4df5eab5c1bee42cc. |
741 | |
742 | This fix turns out to be wrong, and also made ($<,$>)=(...) fail |
743 | (RT #75212). |
744 | |
745 | The original problem was a SEGV in av_clear(). This was mis-diagnosed |
746 | as recursive PL_delaymagic issue, and the fix was to temprarily reset |
747 | PL_delaymagic to zero. This stopped the mg_set() of $> and $> being |
748 | delayed. |
749 | |
750 | The real problem was that mg_free wasn't clearing the [GSR]MG flags |
751 | after freeing xmg_magic. This was independently fixed by commit |
752 | 68f8932eb570af656553ed44c11a23f0a216a3ec. |
753 | |
e88341ca |
754 | commit 65c7421c80585e0d12a20773935dc01f4ffa3e42 # DOCed (verify needed) |
91c261ec |
755 | Author: David Mitchell <davem@iabyn.com> |
756 | Date: Fri Jun 4 17:05:21 2010 +0100 |
757 | |
758 | threads::shared: veto signal despatch if locked |
759 | |
760 | This fixes RT #74868: Safe signals changes causing hangs with threads. |
761 | |
762 | The basic issue is that due to changes in where safe signals can be |
763 | despatched, (including now on leaving scope), it's possible for a |
764 | perl-level signal handler to be called while PL_sharedsv_lock is held. |
765 | If the handler does locking or manipulation of shared vars, then deadlock |
766 | can occur. |
767 | |
768 | A robust fix for this is to ensure that the signal handler isn't called |
769 | while we have the lock. This is done using the signal handler hook added |
770 | in the previous commit. |
771 | |
24e9621f |
772 | commit 92f022bbf8c129c6f2379a382f1eaaa5c7bd9f3f # NODOC |
91c261ec |
773 | Author: David Mitchell <davem@iabyn.com> |
774 | Date: Fri Jun 4 16:18:05 2010 +0100 |
775 | |
776 | add PL_signalhook to hook into signal dispatch |
777 | |
778 | This is initially intended for threads::shared and shouldn't (yet) |
779 | be considered part of the public API. |
780 | |
f5816479 |
781 | commit 83918a83a5dfae8071d366c33fa37b83aabecfc4 # DOCed |
91c261ec |
782 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
783 | Date: Thu Jun 3 16:11:26 2010 +0200 |
784 | |
785 | Do not advertise Math::TrulyRandom, which hasn't been updated since 1996 |
786 | |
787 | (Spotted by Giel Goudsmit) |
788 | |
a5ed29ae |
789 | commit f58883a1a38c36629a8af25cd2fe9b2c5c3fcfec # NODOC |
91c261ec |
790 | Author: Craig A. Berry <craigberry@mac.com> |
791 | Date: Thu Jun 3 07:44:17 2010 -0500 |
792 | |
793 | Remove extraneous semicolon from OP_PRIVATE_ONCE. |
794 | |
795 | Thus silencing compiler noise like: |
796 | |
797 | OP_PRIVATE_ONCE(op_aassign, OPpASSIGN_COMMON, ",COMMON"); |
798 | ........................................................^ |
799 | %CC-I-EXTRASEMI, Extraneous semicolon. |
800 | at line number 846 in file D0:[craig.blead]dump.c;1 |
801 | |
413d5c38 |
802 | commit 03d5bcf83cdbcccf9aefa373ec7b6a3dc11c5142 # ALREADY |
91c261ec |
803 | Author: Vincent Pit <perl@profvince.com> |
804 | Date: Thu Jun 3 12:08:14 2010 +0200 |
805 | |
806 | Deprecate find_rundefsvoffset() |
807 | |
413d5c38 |
808 | commit 483ce06a4f6c720f66285117b98747911f62226a # ALREADy |
91c261ec |
809 | Author: Vincent Pit <perl@profvince.com> |
810 | Date: Thu Jun 3 11:55:20 2010 +0200 |
811 | |
812 | The UNDERBAR macro should use find_rundefsv() as well |
813 | |
413d5c38 |
814 | commit 789bd863840ef4ff6c46f7c2ee0f3f64e0b5daa6 # ALREADY |
91c261ec |
815 | Author: Vincent Pit <perl@profvince.com> |
816 | Date: Thu Jun 3 11:44:15 2010 +0200 |
817 | |
818 | Make pp_reverse fetch the lexical $_ from the correct pad |
819 | |
820 | This is achieved by introducing a new find_rundefsv() function in pad.c |
821 | |
822 | This fixes [perl #75436]. |
823 | |
3856e817 |
824 | commit 540810e8986e170e75f4b34a7ca1f1dd5b0da3c4 # NODOC |
91c261ec |
825 | Author: Abigail <abigail@abigail.be> |
826 | Date: Tue Jun 1 18:27:43 2010 +0200 |
827 | |
828 | Skip two Japhs that use the fact that split in void context splitted |
829 | to @_. This is no longer true in 5.12.0 and upwards. |
830 | |
97845720 |
831 | commit 321499b5be1bfbade4d0e2e11136d1d4d87a5e7f # NODOC |
91c261ec |
832 | Author: Slaven Rezic <srezic@iconmobile.com> |
833 | Date: Thu Apr 29 16:04:35 2010 +0200 |
834 | |
835 | Document IO::Socket getsockopt and setsockopt |
836 | |
837 | get/setsockopt are not explicitly documented. Note that the documented |
838 | sockopt() method is a wrapper around get/setsockopt, but does not cover |
839 | all cases because of the hardcoded level (= SOL_SOCKET). |
840 | |
33879272 |
841 | commit 065144c63ee2795acbc962d225b0a3f8d7967ae3 # NODOC |
91c261ec |
842 | Author: Tony Cook <tony@develop-help.com> |
843 | Date: Tue Jun 1 09:50:08 2010 +1000 |
844 | |
845 | move the fresh_perl() tests above the environment tests |
846 | |
847 | the environment tests were resetting environment variables useful for |
848 | running fresh perls, like LD_LIBRARY_PATH. |
849 | |
33879272 |
850 | commit 1b1ee2ef87e2dcc8a1699cc870aefd1b91c5f645 # NODOC |
91c261ec |
851 | Author: Karl Williamson <khw@khw-desktop.(none)> |
852 | Date: Sun May 30 21:54:32 2010 -0600 |
853 | |
854 | PATCH: teach diag.t new warning function names |
855 | |
856 | A number of function names that do warnings have been added, but diag.t |
857 | hasn't kept up. |
858 | |
859 | This patch changes it to look for likely function names in embed.fnc, so |
860 | it will automatically keep up in the future. There's no need to worry |
861 | about it looking for inappropriate functions, as the syntax of messages |
862 | that it looks for is so restrictive, that there won't be false |
863 | positives. Instead there are still many messages it fails to catch. |
864 | |
865 | As a result of it's falling behind several issues have crept in. I |
866 | resolved the couple I thought were clear (including one in a comment; |
867 | diag.t doesn't strip comments, but mostly it doesn't matter), and added |
868 | the others to the <DATA> section to ignore. |
869 | are |
870 | |
933ee134 |
871 | commit 51eec7ec9cf1a154df61e6fc6c46acab7c69b296 # DOCed |
91c261ec |
872 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
873 | Date: Mon May 31 14:35:09 2010 +0200 |
874 | |
875 | State the requirement of a C89 compliant ANSI C-compiler |
876 | |
434477a5 |
877 | commit c1bf42f3e6ad8f1c3d821a2ae616c5703f66237c # NODOC (verify needed) |
91c261ec |
878 | Author: Nicholas Clark <nick@ccl4.org> |
879 | Date: Mon May 31 13:19:22 2010 +0100 |
880 | |
881 | In Perl_pad_add_name(), use sv_upgrade() directly rather than new[AH]V(). |
882 | |
883 | As newAV() and newHV() are now merely wrappers around sv_upgrade(), and the |
884 | existing SV is always brand new and of type SVt_NULL, call them on it, rather |
885 | than disposing of it as a side effect of storing a(nother new) SV. |
886 | |
887 | Also, no need to set SvPADMY() again, as it is already set. |
888 | |
889 | Resolves RT #73092. |
890 | |
152cfc22 |
891 | commit 4c050ad563ece4467b3b083d8efcf2b62ad0b9c5 # NODOC |
91c261ec |
892 | Author: Nicholas Clark <nick@ccl4.org> |
893 | Date: Mon May 31 11:47:04 2010 +0100 |
894 | |
895 | Reorder the entry for die, moving discussion of the exit code later. |
896 | |
897 | Change to start with "C<die> raises an exception." and phrase in terms of |
898 | handling exceptions, and then how uncaught exceptions result in process exit. |
899 | Do not give details of the exit code in the first paragraph. Move this to a |
900 | later paragraph, clarify that $! is often unpredictable, and stress that 255 |
901 | is the last resort exit code. |
902 | |
fd1ad2d0 |
903 | commit 65ab9279784aa811d78b2903b57bc0e7947dec78 # DOCed |
91c261ec |
904 | Author: Tony Cook <tony@develop-help.com> |
905 | Date: Tue Mar 16 23:46:48 2010 +1100 |
906 | |
907 | handle perl extended utf8 start bytes |
908 | |
909 | perl uses UTF8_IS_START() to test if a byte is a valid start byte, |
910 | this didn't take perl's extended UTF-8 range into account. |
911 | |
a9d055ae |
912 | commit e57ed4ecd4d7de38a79a316da8d657dad656f93f # NODOC |
91c261ec |
913 | Author: Father Chrysostomos <sprout@cpan.org> |
914 | Date: Mon May 31 20:03:11 2010 +1000 |
915 | |
916 | [perl #70075] no 6; broken |
917 | |
918 | the original patch included a fix, but this problem was fixed in |
919 | faee19b5, so only apply the test. |
920 | |
765d1953 |
921 | commit 1408fb84de0c28007d91730cde177d893e427181 # NODOC |
91c261ec |
922 | Author: Karl Williamson <khw@khw-desktop.(none)> |
923 | Date: Sun May 30 13:05:48 2010 -0600 |
924 | |
925 | PATCH: [perl #75138] "\c`" -> " " |
926 | |
927 | Attached is a patch for some of this issue. I took Nicholas' advice, |
928 | and if the result of \cX isn't a word character, the output message will |
929 | precede it with a backslash, so the message in the example would be |
930 | |
931 | "\c`" more clearly written simply as "\ " at -e line 1. |
932 | |
933 | I think that message is true. |
934 | |
935 | I also added tests. |
936 | |
937 | There is a test that guarantees that we won't ship 5.14 with things as |
938 | they are now in it. I added wording to the comments next to that test |
939 | to be sure to verify with this email thread if we should remove the |
940 | deprecation, and mentioned that in the explanatory wording in the pod. |
941 | I support removing the deprecation, but for now I'm not touching that, |
942 | to see what other issues may yet arise before 5.14. |
943 | |
e47ad259 |
944 | commit 71648f9a69cff1f8ee90adfed51c64c3c2dfeaf1 # DOCed |
91c261ec |
945 | Author: Karl Williamson <khw@khw-desktop.(none)> |
946 | Date: Thu May 27 22:24:40 2010 -0600 |
947 | |
948 | Document tricks, work-arounds for user-defined casing |
949 | |
950 | And add a .t file to verify that it works. |
951 | |
e6b9fcab |
952 | commit abd889b86d676557bdda48bf0873bd04aa95a9ff # NODOC |
91c261ec |
953 | Author: Jerry D. Hedden <jdhedden@cpan.org> |
954 | Date: Wed May 26 14:50:39 2010 -0400 |
955 | |
956 | Un-TODO a threads-shared test that now passes |
957 | |
958 | Commit bb1bc619ea68d9703fbd3fe5bc65ae000f90151f has |
959 | fixed a threads-shared TODO test in t/object.t. |
960 | |
ca9af26c |
961 | commit 9ef2b94f54586a775e4e47eebd8bbd9e08ce5fdd # NODOC |
91c261ec |
962 | Author: Karl Williamson <khw@khw-desktop.(none)> |
963 | Date: Tue May 4 21:43:56 2010 -0600 |
964 | |
965 | Add mktables option for development use |
966 | |
967 | The -output_names option was added. It will cause the generated file |
968 | tables to not have ranges, and each line will have the character name. |
969 | This makes it easier to compare what characters are in given tables, |
970 | from version to version, or to compare the differences between |
971 | properties. |
972 | |
ca9af26c |
973 | commit 28093d0e3328797fc0783f9d909d7931ba57bd59 # NODOC |
91c261ec |
974 | Author: Karl Williamson <khw@khw-desktop.(none)> |
975 | Date: Tue May 4 15:14:24 2010 -0600 |
976 | |
977 | mktables -- don't create Names table unless asked |
978 | |
979 | This speeds up mktables by not creating the Names table unless asked to, |
980 | by someone adding it to the list of tables to be output. Perl uses a |
981 | different table than this one for charnames, so the one being suppressed |
982 | isn't generally used. Previously it was created but not output. Now, |
983 | we skip the useless creation step. |
984 | |
b3d32175 |
985 | commit ec11e5f48c39adae8dd35f4d50a4b1b04eb80b97 # HALPHALP |
91c261ec |
986 | Author: Karl Williamson <khw@khw-desktop.(none)> |
987 | Date: Tue May 4 15:10:18 2010 -0600 |
988 | |
989 | Fix priority of suppressed vs. explicitly output |
990 | |
991 | It's not clear this is a real bug, but it is a surprise. If a table is |
992 | in the suppressed list, it isn't output, even if it is in the |
993 | to-be-output override list. This latter list is non-empty only if the |
994 | user has hand-edited the the program to force an output. So this patch |
995 | makes that list have priority. |
996 | |
b3d32175 |
997 | commit 9d682c834ea42d040e5ba121ea8595c2f298b64f # NODOC |
91c261ec |
998 | Author: Karl Williamson <khw@khw-desktop.(none)> |
999 | Date: Tue May 4 11:18:59 2010 -0600 |
1000 | |
1001 | Remove obsolete comment |
1002 | |
1003 | Commit: 6c4b69c35161f79a5088d6c3070cc17a0e4978b2 made this comment |
1004 | obsolete; forgot to remove it then. |
1005 | |
b3d32175 |
1006 | commit 959ce5bf0ff22e4213e8987e43954483009d4a8e # NODOC |
91c261ec |
1007 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1008 | Date: Tue May 4 09:13:35 2010 -0600 |
1009 | |
1010 | Generate simple case folding tables only if asked |
1011 | |
1012 | Speed up mktables by not generating the simple case folding tables |
1013 | unless asked to. Previously the simple tables were generated, and then |
1014 | the full tables were initialized with them, and then overwritten with |
1015 | the full mappings. This is an artifact from the fact that the data |
1016 | comes to us in two files, one with the simple mappings (among other |
1017 | things), and another with the full mapping overrides. Now, the full |
1018 | tables are initialized from the first file, and the second file |
1019 | overrides the full mappings. The simple tables are not generated by |
1020 | default, so this saves, copying them. |
1021 | |
b3d32175 |
1022 | commit 6c68572bb711304d008bc1439d541dd69fcde4e0 # NODOC |
91c261ec |
1023 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1024 | Date: Sun May 30 16:49:28 2010 +0200 |
1025 | |
1026 | Add comment about objaddr in mktables |
1027 | |
1028 | Slightly modified to include the commit id by Steffen. |
1029 | |
b3d32175 |
1030 | commit f998e60c2aa1b9879e2c7c365acd05c0799ae46a # NODOC |
91c261ec |
1031 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1032 | Date: Mon May 3 10:06:30 2010 -0600 |
1033 | |
1034 | Use in-line 'no overloading' for speed |
1035 | |
1036 | An earlier performance enhancement was to change the subroutine that |
1037 | gets the address of a ref to using 'no overloading' and then numifying |
1038 | the ref, which returns its address. This patch speeds things up |
1039 | slightly by in-lining the "no overloading" so that the function call |
1040 | overhead is avoided. |
1041 | |
1042 | It also gets rid of the kludge that was done before the original |
1043 | speed-up that created a local in the call stack of one of the classes so |
1044 | that the address would only have to be executed once per call stack; |
1045 | This was subject to failure if maintenance of the code perturbed things |
1046 | so it didn't work; now the overhead is minimal, so the address is gotten |
1047 | in each call. |
1048 | |
d4451e15 |
1049 | commit 4a68bf9d410f9046c46490523fe5d75c93632a5d # NODOC |
91c261ec |
1050 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1051 | Date: Sun May 16 06:56:01 2010 -0600 |
1052 | |
1053 | Don't require doubled backslashes in perldiag.pod |
1054 | |
1055 | Prior to this patch, messages in perldiag.pod had to have \\ instead of |
1056 | the correct single backslash in order for diag.t to not complain. |
1057 | |
d4451e15 |
1058 | commit b081dd7eaaec2b6ee43335645ab40cff0ca3f91a # NODOC |
91c261ec |
1059 | Author: Nicholas Clark <nick@ccl4.org> |
1060 | Date: Sun May 30 12:52:01 2010 +0100 |
1061 | |
1062 | Eliminate some newSV(0)s by merging the SV allocation with first modification. |
1063 | |
86b722f7 |
1064 | commit 77bac227771c643a8a6e305b2bac4665a8f772d1 # DOCed |
91c261ec |
1065 | Author: Nicholas Clark <nick@ccl4.org> |
1066 | Date: Sun May 30 11:51:15 2010 +0100 |
1067 | |
1068 | Only allocate entries for @_ when the subroutine is first called. |
1069 | |
1070 | Previously, @_ was allocated for every subroutine at compile time, with a |
1071 | default sized AV, hence space for 4 entries. We don't actually need to |
1072 | allocate the space for entries, as there is already a check at call time as to |
1073 | whether @_ is long enough. |
1074 | |
1075 | valgrind suggests that this saves allocating 23K (on a 64 bit platform) when |
1076 | running pod2man on perlfunc.pod. As well as the absolute saving, there is also |
1077 | benefit in deferring allocation for subroutines actually called - for a program |
1078 | which forks, @_ is less likely to be in pages shared COW with the parent. |
1079 | |
1080 | Resolves RT #72416. |
1081 | |
bb97bb8f |
1082 | commit c85ae797ecb755d3bcfabd44aa268e3e6e2e7c13 # NODOC |
91c261ec |
1083 | Author: Nicholas Clark <nick@ccl4.org> |
1084 | Date: Sat May 29 22:22:36 2010 +0100 |
1085 | |
1086 | Fix the regexp in t/porting/args_assert.t, and add 3 missing macros. |
1087 | |
1088 | Resolves RT #72800. |
1089 | |
bb97bb8f |
1090 | commit af534a0431cd837092de9abe4bff29feeeafebb3 # NODOC |
91c261ec |
1091 | Author: Nicholas Clark <nick@ccl4.org> |
1092 | Date: Sat May 29 22:07:04 2010 +0100 |
1093 | |
1094 | Properly free paren_name_list with its regexp. |
1095 | |
1096 | Previously the AV paren_name_list would "leak" until global destruction. |
1097 | This was only an issue under -DDEBUGGING. Fixes RT #73438. |
1098 | |
bb97bb8f |
1099 | commit 1c8f8eb1d500716bc7368dab00869ae3be841e0c # NODOC |
91c261ec |
1100 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1101 | Date: Thu Jan 14 21:21:37 2010 -0700 |
1102 | |
1103 | Clarify that count is bytes not unicode characters |
1104 | |
bb97bb8f |
1105 | commit c2dc4c7d5e51c6211637044820c7a560da7e6268 # NODOC |
91c261ec |
1106 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1107 | Date: Thu Jan 14 19:19:22 2010 -0700 |
1108 | |
1109 | Display characters as Unicode for clarity |
1110 | |
bb97bb8f |
1111 | commit 35db910f236b637e47a514ee62ca3e5a98169ddd # NODOC |
91c261ec |
1112 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1113 | Date: Thu Jan 14 17:36:46 2010 -0700 |
1114 | |
1115 | Add tested for corrupted regnode |
1116 | |
bb97bb8f |
1117 | commit 525aa6214fa98872c8e7604a6b63bd6c7914f3a4 # NODOC |
91c261ec |
1118 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1119 | Date: Thu Jan 14 16:02:14 2010 -0700 |
1120 | |
1121 | Use sizeof instead of hard-coded array size |
1122 | |
1123 | The array should be declared with its actual size. |
1124 | |
50220c3d |
1125 | commit 22c985d500f518cdb601db921ea5b3edaa6c4ae6 # NODOC |
91c261ec |
1126 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1127 | Date: Thu Jan 14 16:01:13 2010 -0700 |
1128 | |
1129 | Typo |
1130 | |
a6e352d6 |
1131 | commit 4c0f30d62c000e3bbbd4f45eb9fd4bd4b4015cf9 # DOCed |
91c261ec |
1132 | Author: Nicholas Clark <nick@ccl4.org> |
1133 | Date: Sat May 29 16:23:53 2010 +0100 |
1134 | |
1135 | When assigning to $^P, don't zero $DB::single, $DB::trace and $DB::signal. |
1136 | |
1137 | Previously, whenever a true value was assigned to $^P, all 3 were set to 0. |
1138 | Now only set them to 0 if they aren't already SvIOK(). |
1139 | |
1140 | Resolves RT #72422. |
1141 | |
216cbae1 |
1142 | commit 261c990ef860580c9092985d3f4629073a671b2f # NODOC |
91c261ec |
1143 | Author: Nicholas Clark <nick@ccl4.org> |
1144 | Date: Fri May 28 16:02:12 2010 +0100 |
1145 | |
1146 | Migrate most other op_private to name conversion into S_op_private_to_names(). |
1147 | |
216cbae1 |
1148 | commit 1fe3abee1ef8693b6241cf4cb1670e7e5e2d4a51 # NODOC |
91c261ec |
1149 | Author: Nicholas Clark <nick@ccl4.org> |
1150 | Date: Fri May 28 14:33:34 2010 +0100 |
1151 | |
1152 | In Perl_do_op_dump(), move calls to append_flags() into S_op_private_to_names() |
1153 | |
216cbae1 |
1154 | commit cd431fdedf8dec69c65efa7edc1d60005800919a # NODOC |
91c261ec |
1155 | Author: Nicholas Clark <nick@ccl4.org> |
1156 | Date: Fri May 28 14:53:04 2010 +0100 |
1157 | |
1158 | Add C_ARRAY_END(), returning a pointer to after the last element of an array. |
1159 | |
1160 | Refactor the macro append_flags() in dump.c to use it. |
1161 | |
216cbae1 |
1162 | commit 5c135d48c0468f552840e5ac9811f70aebcac766 # NODOC |
91c261ec |
1163 | Author: Nicholas Clark <nick@ccl4.org> |
1164 | Date: Fri May 28 13:53:02 2010 +0100 |
1165 | |
1166 | In Perl_do_op_dump(), reorder the ops within the if (o->op_private) clause. |
1167 | |
216cbae1 |
1168 | commit ae1f06a125e806e4a0c111878fb9da530a3df3c6 # NODOC |
91c261ec |
1169 | Author: Nicholas Clark <nick@ccl4.org> |
1170 | Date: Fri May 28 13:43:43 2010 +0100 |
1171 | |
1172 | In Perl_do_sv_dump(), use append_flags() for PVCV, PVFM and PVGP flags. |
1173 | |
216cbae1 |
1174 | commit 16cf2f8dc20a01967bab359ae99891d1259c6812 # NODOC |
91c261ec |
1175 | Author: Nicholas Clark <nick@ccl4.org> |
1176 | Date: Fri May 28 11:50:54 2010 +0100 |
1177 | |
1178 | In Perl_do_sv_dump(), for PVCV and PVFM, test for SvCOMPILED(sv) last. |
1179 | |
216cbae1 |
1180 | commit ea9ad1f28e506daa41693d2ef7ab39b8182101b1 # NODOC |
91c261ec |
1181 | Author: Nicholas Clark <nick@ccl4.org> |
1182 | Date: Fri May 28 11:26:16 2010 +0100 |
1183 | |
1184 | In Perl_do_op_dump(), move runs of op_private name tests to S_append_flags(). |
1185 | |
216cbae1 |
1186 | commit a0c2f4dde00175193ad56350e2a8982e360c6726 # NODOC |
91c261ec |
1187 | Author: Nicholas Clark <nick@ccl4.org> |
1188 | Date: Fri May 28 10:56:21 2010 +0100 |
1189 | |
1190 | Create S_append_flags() from a common code pattern in dump.c. |
1191 | |
1192 | Convert repetitive sequences of "if this bit is set, append that string" into |
1193 | structures and a function call. Use a custom macro append_flags() to make |
1194 | calling it easer. |
1195 | |
1196 | This makes the object code slightly smaller. |
1197 | |
216cbae1 |
1198 | commit eff5b9d539e47421a784cb6a5fa9366c6522a4eb # NODOC |
91c261ec |
1199 | Author: Nicholas Clark <nick@ccl4.org> |
1200 | Date: Tue Jan 5 13:06:52 2010 +0000 |
1201 | |
1202 | InPerl_boot_core_UNIVERSAL() use a data structure for calls to newXS{,proto} |
1203 | |
1204 | Replacing the longhand list of calls to newXS{,proto} with loop over a data |
1205 | structure reduces the object size by over 1K. |
1206 | |
d60f0907 |
1207 | commit dc1f0c2041dab5a98af69338d1fa501b5e90ac70 # HALPHALP |
91c261ec |
1208 | Author: Salvador Ortiz Garcia <sog@msg.com.mx> |
1209 | Date: Fri May 28 10:25:52 2010 +0200 |
1210 | |
1211 | Naif calls segfault T_PRTOBJ of the stock typemap |
1212 | |
1213 | The T_PTROBJ INPUT in stock typemap only uses 'sv_derived_from' for |
1214 | input validation, that cause a segfault when the argument passed match |
1215 | the class name. |
1216 | |
05300b4b |
1217 | commit 02927be17db68ef768a4a953e754b78a4350b226 # NODOC (verify needed) |
91c261ec |
1218 | Author: Nicholas Clark <nick@ccl4.org> |
1219 | Date: Fri May 28 07:58:19 2010 +0100 |
1220 | |
1221 | Add simple_bitmask and varies_bitmask to globvar.sym. |
1222 | |
1223 | global.sym is generated; is there a way to automate globvar.sym? |
1224 | |
05300b4b |
1225 | commit ded4dd2add376b302c561318612805c584ef9e6a # NODOC |
91c261ec |
1226 | Author: Nicholas Clark <nick@ccl4.org> |
1227 | Date: Wed Mar 17 17:16:24 2010 +0000 |
1228 | |
1229 | Convert REGNODE_{SIMPLE,VARIES} to a bitmask lookup, from a strchr() lookup. |
1230 | |
1231 | This is O(1) with no branching, instead of O(n) with branching. |
1232 | Deprecate the old implementation's externally visible variables |
1233 | PL_simple and PL_varies. Google codesearch suggests that nothing outside the |
1234 | core regexp code was using these. |
1235 | |
05300b4b |
1236 | commit 93882df08ca38a02b0381419b29fff019ec87ef7 # NODOC |
91c261ec |
1237 | Author: Nicholas Clark <nick@ccl4.org> |
1238 | Date: Wed Mar 17 16:41:27 2010 +0000 |
1239 | |
1240 | Change regcomp.pl to 0-based indexing for its arrays and loops. |
1241 | |
05300b4b |
1242 | commit e52fc5395a9d11f134b6e4ecacde7782d1af6b26 # NBODOC |
91c261ec |
1243 | Author: Nicholas Clark <nick@ccl4.org> |
1244 | Date: Wed Mar 17 13:58:25 2010 +0000 |
1245 | |
1246 | Encapsulate lookups in PL_{varies,simple} within macros REGNODE_{VARIES,SIMPLE} |
1247 | |
1248 | This allows the implementation of the lookup mechanism to change. |
1249 | |
05300b4b |
1250 | commit f9ef50a71935a8e93b4030c12dcd1206ccab71ab # NODOC |
91c261ec |
1251 | Author: Nicholas Clark <nick@ccl4.org> |
1252 | Date: Wed Mar 17 13:33:48 2010 +0000 |
1253 | |
1254 | Generate PL_simple[] and PL_varies[] with regcomp.pl, rather than hard-coding. |
1255 | |
1256 | Add a new flags column to regcomp.sym, with V if the node type is in PL_varies, |
1257 | S if it is in PL_simple, and . if a placeholder is needed because subsequent |
1258 | optional columns are present. |
1259 | |
05300b4b |
1260 | commit f8abb37e5b952f76a9e019137369e3f8ef5a58ae # NODOC |
91c261ec |
1261 | Author: Nicholas Clark <nick@ccl4.org> |
1262 | Date: Tue Mar 16 10:22:04 2010 +0000 |
1263 | |
1264 | Re-work the regcomp.sym to remove use of hard tabs. No data change. |
1265 | |
1266 | The tab separating name and type is replaced with whitespace, the tab marking |
1267 | the start of the description is replaced by a semicolon. |
1268 | |
05300b4b |
1269 | commit 5014f7dab1e8d50a8f9ba9b88aa6446cfc8189ab # NODOC |
91c261ec |
1270 | Author: Nicholas Clark <nick@ccl4.org> |
1271 | Date: Tue Mar 16 09:50:29 2010 +0000 |
1272 | |
1273 | Correct the node numbers in comments. |
1274 | |
1275 | Really, should we be maintaining these manually? |
1276 | |
05300b4b |
1277 | commit 20832bc55b5c8a1f61c1270e40b964ff1488847d # NODOC |
91c261ec |
1278 | Author: Nicholas Clark <nick@ccl4.org> |
1279 | Date: Tue Mar 16 09:48:35 2010 +0000 |
1280 | |
1281 | Remove stray tab character in definition for VERB. |
1282 | |
1283 | As VERB is "Used only for the type field of verbs" this is only a cosmetic |
1284 | change, causing that correct description to appear in the comment in |
1285 | regnodes.h. The change to regarglen doesn't affect anything, as the VERB type |
1286 | is never actually used for compiled nodes. |
1287 | |
c5075608 |
1288 | commit 1404b48744cfa915e3f54775feb4e9b6c10f3b91 # NODOC |
91c261ec |
1289 | Author: George Greer <perl@greerga.m-l.org> |
1290 | Date: Wed May 26 21:12:37 2010 -0400 |
1291 | |
1292 | Fix clang "incompatible operand types" error in ternary expressions. |
1293 | |
d4a8a3ce |
1294 | commit a6f87d8c282ba94b0308898be052d1ac956d0ea9 # HALPHALP |
91c261ec |
1295 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1296 | Date: Fri May 14 08:56:46 2010 -0600 |
1297 | |
1298 | PATCH: user defined special casing for non utf8 |
1299 | |
1300 | Users can define their own case changing mappings to replace the |
1301 | standard ones. Prior to this patch, any mappings on characters whose |
1302 | ordinals are 0-222, 224-255 that resulted in multiple characters were |
1303 | ignored. |
1304 | |
1305 | Note that there still is a deficiency in that the mappings will be |
1306 | applied only to strings in utf8 format. |
1307 | |
e33229c4 |
1308 | commit 8a38a8369536df0f4eac69e5a26e9b86f9123d1d # NODOC |
91c261ec |
1309 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1310 | Date: Mon May 24 11:01:10 2010 -0600 |
1311 | |
1312 | Change a few tests to use new test.pl EBCDIC stuff |
1313 | |
1314 | A few .t's are delivered here which use the new conversion functions in |
1315 | test.pl. Obviously, this commit doesn't have to be incorporated, but it |
1316 | does simplify those tests. |
1317 | |
72bc2074 |
1318 | commit f69d9fdf3f3cf041b8398a5d897daef24f0e5fa0 # NODOC |
91c261ec |
1319 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1320 | Date: Mon May 24 10:50:13 2010 -0600 |
1321 | |
1322 | Add functions to test.pl for easier EBCDIC testing |
1323 | |
1324 | This patch adds functions native_to_latin1($) and latin1_to_native($) to |
1325 | test.pl. Use of these in test scripts allows for less special case code |
1326 | (or deciding to blow it off and just skip EBCDIC platforms). If a |
1327 | string is all ASCII, one has always been able to just use it and not |
1328 | worry about EBCDIC. But non-ASCII, latin1 characters have been a |
1329 | different matter. This allows one to wrap those characters in a |
1330 | function call and get automatic EBCDIC compatibility. |
1331 | |
1332 | The functions do not handle UTF-8/UTF-EBCDIC conversions. This is more |
1333 | complicated, and I was worried that that part was not appropriate for |
1334 | test.pl which should use the bare minimum of Perl functionality. These |
1335 | functions use tr///, beyond what other parts of test.pl use. |
1336 | |
69c0769c |
1337 | commit 04518cc3f43b495f85caf2ec89c8b06540a60f8c # ALREADY |
91c261ec |
1338 | Author: Nicholas Clark <nick@ccl4.org> |
1339 | Date: Tue May 25 17:23:10 2010 +0100 |
1340 | |
1341 | Fix CLONE/weakref bug revealed by adf8f095c5881bce. |
1342 | |
1343 | The AV unreferenced in the clone_params needs to be reference counted, rather |
1344 | than not referenced counted, because the fixup to ensure that all otherwise |
1345 | 0-reference count scalars have a reference (on the temps stack) happens after |
1346 | CLONE is run, and CLONE can run Perl code that causes their reference counts |
1347 | to increase from then return to zero, which prematurely triggers sv_free(). |
1348 | |
85f6e172 |
1349 | commit a77cd7b8173d6381e0eb4d3296627023900d0cb4 # NODOC |
91c261ec |
1350 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
1351 | Date: Tue May 25 18:26:37 2010 +0200 |
1352 | |
1353 | Update rsync sources in perlhack |
1354 | |
d5c47a3d |
1355 | commit 775f1d61966fb2c5300c9c08fa931980430ed0da # NODOC |
91c261ec |
1356 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1357 | Date: Mon May 24 23:40:48 2010 +0300 |
1358 | |
1359 | Add the perl equivalent for av_make. |
1360 | |
1361 | Signed-off-by: David Golden <dagolden@cpan.org> |
1362 | |
d5c47a3d |
1363 | commit a8676f70bbd19ba630d2e76288f31b4a2d7a63f1 # NODOC |
91c261ec |
1364 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1365 | Date: Wed May 19 20:11:03 2010 +0300 |
1366 | |
1367 | Add the Perl equivalent for av_len. |
1368 | |
1369 | Signed-off-by: David Golden <dagolden@cpan.org> |
1370 | |
d5c47a3d |
1371 | commit 1a3362a5d4e9f2d3548e2937ecf5a9a8a1ac7898 # NODOC |
91c261ec |
1372 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1373 | Date: Wed May 19 17:46:43 2010 +0300 |
1374 | |
1375 | Add a missing comma in the av_fill() docs. |
1376 | |
1377 | Minor, but still good enough for a commit. |
1378 | |
1379 | Signed-off-by: David Golden <dagolden@cpan.org> |
1380 | |
d5c47a3d |
1381 | commit 1a32886282d60539000205670909069b85d9eedd # NODOC |
91c261ec |
1382 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1383 | Date: Tue May 18 17:42:52 2010 +0300 |
1384 | |
1385 | Clarify the av_fetch() documentation. |
1386 | |
1387 | Thanks to LeoNerd and Zefram on #p5p on IRC for some insights and |
1388 | suggesting versions for the modified text. I ended up using Zefram's |
1389 | version. |
1390 | |
1391 | Signed-off-by: David Golden <dagolden@cpan.org> |
1392 | |
d5c47a3d |
1393 | commit b7ff7ff211d0d5fc593efe5c9d305d197ecb8d86 # NODOC |
91c261ec |
1394 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1395 | Date: Tue May 18 15:31:41 2010 +0300 |
1396 | |
1397 | Add Perl equivalent for av_exists(). |
1398 | |
1399 | This mentions that it's equivalent to exists($myarray[$key]). |
1400 | |
1401 | Signed-off-by: David Golden <dagolden@cpan.org> |
1402 | |
d5c47a3d |
1403 | commit 71282cab69578876db765e8d8aadb9cc2331ebf2 # NODOC |
91c261ec |
1404 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1405 | Date: Sat May 15 22:13:59 2010 +0300 |
1406 | |
1407 | Add a Perl equivalent to av_delete(). |
1408 | |
1409 | Signed-off-by: David Golden <dagolden@cpan.org> |
1410 | |
d5c47a3d |
1411 | commit 31bde0acd044dc2a6e4ccd40fa4660a6009fcb9d # NODOC |
91c261ec |
1412 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1413 | Date: Sat May 15 21:44:58 2010 +0300 |
1414 | |
1415 | Add the Perl equivalent example to av_clear. |
1416 | |
1417 | Signed-off-by: David Golden <dagolden@cpan.org> |
1418 | |
d5c47a3d |
1419 | commit 683d9ad15a8b09b17249f2efd39d10a9a251cd95 # NODOC |
91c261ec |
1420 | Author: Shlomi Fish <shlomif@iglu.org.il> |
1421 | Date: Sat May 15 21:37:46 2010 +0300 |
1422 | |
1423 | Link from GIMME_V to perlcall for a usage example. |
1424 | |
1425 | Added a link from the GIMME_V entry to perlcall for a usage example. |
1426 | |
1427 | Signed-off-by: David Golden <dagolden@cpan.org> |
1428 | |
9c4e2ac7 |
1429 | commit d5cd9e7bba185db6dc6b1e6fa215978a38ae9ea8 # NODOC |
91c261ec |
1430 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1431 | Date: Mon May 24 08:08:51 2010 -0600 |
1432 | |
1433 | PATCH: correct misstatement, formats in perlunicode |
1434 | |
1435 | This is suitable for 5.12.2, but not many people use this feature. |
1436 | |
ce3fe28d |
1437 | commit 24e28fd799c14b3895ad81b7f27a6a547ac2dd44 # ALREADY |
91c261ec |
1438 | Author: Nicholas Clark <nick@ccl4.org> |
1439 | Date: Tue May 25 16:30:43 2010 +0100 |
1440 | |
1441 | Fix for typo in 0824d66743a706cd, in the expected output with threads. |
1442 | |
ce3fe28d |
1443 | commit 0824d66743a706cd268ace8fc9df03d7374c6886 # ALREADY |
91c261ec |
1444 | Author: David Mitchell <davem@iabyn.com> |
1445 | Date: Tue May 25 11:38:35 2010 +0100 |
1446 | |
1447 | add OPpDEREFed flag to avoid double mg_get() |
1448 | |
1449 | The previous commit made various ops such as rv2av unconditionally do |
1450 | an SvGETMAGIC(). Under some circumstances this could cause a double |
1451 | mg_get() (and hence double FETCH etc). In particular, when the |
1452 | proceeding op was something like aelem with OPpDEREF, the aelem would |
1453 | call vivify_ref(), which would call magic. So in peep(), mark |
1454 | OP_RV2[SAH]V ops with the new OPpDEREFed flag if the preceding op was |
1455 | OPpDEREF. Then use this flag to avoid a second dose of magic. |
1456 | |
1457 | Note that RV2GV probably needs this flag too, but there weren't any |
1458 | spare private flag bits left for that op (I think). |
1459 | |
ce3fe28d |
1460 | commit bb1bc619ea68d9703fbd3fe5bc65ae000f90151f # ALREADY |
91c261ec |
1461 | Author: Father Chrysostomos (via RT) <perlbug-followup@perl.org> |
1462 | Date: Sun Jan 17 14:32:24 2010 -0800 |
1463 | |
1464 | Deref ops ignore get-magic when SvROK(sv) |
1465 | |
1466 | This is just like bug 68192, except in this case it’s a different set |
1467 | of operators that have had this problem for much longer. |
1468 | |
ce3fe28d |
1469 | commit 6a5f8cbd14b4a44b35830907e944f1af0caeea90 # ALREADY |
91c261ec |
1470 | Author: Father Chrysostomos <sprout@cpan.org> |
1471 | Date: Mon May 24 11:56:25 2010 +0100 |
1472 | |
1473 | Just the tests from a proposed fix for 68192 |
1474 | |
1475 | The bug was fixed in a different way by davem, but the tests |
1476 | are needed as the base for a commit to follow |
1477 | |
6f9cbd70 |
1478 | commit 5f26d5fd39994d2ecb568aeb7efdba685fe9a350 # DOCed |
91c261ec |
1479 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1480 | Date: Tue May 18 08:43:31 2010 -0600 |
1481 | |
1482 | Replace wrong tr/// table in perlebcdic.pod |
1483 | |
1484 | perlebcdic.pod contains a helpful table to use in tr/// to convert |
1485 | between EBCDIC and Latin1/ASCII. Unfortunately, the table is the |
1486 | inverse of the one it describes, as I discovered in following the |
1487 | directions on how it was extracted. |
1488 | |
1489 | The actual code that uses the table uses it in the inverse way, so it |
1490 | actually works correctly, but if you follow the instructions to get the |
1491 | tables corresponding to the other recognized code pages, you will get |
1492 | the inverse of the ones you wanted, and things won't work out. |
1493 | |
1494 | This patch changes the table to its inverse, and changes the sample code |
1495 | to correspond, as this is easier for the person trying to follow the |
1496 | instructions since deriving the old table is somewhat more complicated. |
1497 | |
1498 | I also changed the table to hex from octal, as that is more the norm |
1499 | these days, and I changed to recipes in the pod to print out leading |
1500 | zeros to make all the values the same length, as the table that they can |
1501 | generate has them. |
1502 | |
2f4f4fc9 |
1503 | commit 8590c068107627f63321a9736653a498cadf3a27 # NODOC |
91c261ec |
1504 | Author: Karl Williamson <khw@khw-desktop.(none)> |
1505 | Date: Tue May 18 10:02:26 2010 -0600 |
1506 | |
1507 | Remove unused, wrong #define in utf8.h |
1508 | |
1509 | is unused in the code, and is wrong for EBCDIC platforms, as there can |
1510 | be invariants there that aren't ASCII. I simply removed it. |
1511 | |
9a382194 |
1512 | commit 5df417d0a2691b4e62ac4c2403e6cf71612289f6 # NODOC |
91c261ec |
1513 | Author: Jerry D. Hedden <jdhedden@cpan.org> |
1514 | Date: Fri May 7 12:14:05 2010 -0400 |
1515 | |
1516 | Fix build warnings introduced by v5.13.0-139-ge0fa7e2 |
1517 | |
07601a2e |
1518 | commit 7c425e3cac4112a3489a559bbfc18e717d87d070 # NODOC |
91c261ec |
1519 | Author: Nicholas Clark <nick@ccl4.org> |
1520 | Date: Mon May 24 21:01:29 2010 +0100 |
1521 | |
1522 | Add Perl_clone_params_{del,new} and Perl_sv_dup to the "no threads" exclusions. |
1523 | |
ef4ed187 |
1524 | commit adf8f095c5881bcedf07b8e41072f8125e00b5a6 # NODOC |
91c261ec |
1525 | Author: Nicholas Clark <nick@ccl4.org> |
1526 | Date: Fri Feb 26 09:18:44 2010 +0000 |
1527 | |
1528 | Set PADSTALE on all lexicals at the end of sub creation. |
1529 | |
1530 | The PADSTALEness of lexicals between the 0th and 1st call to a subroutine is now |
1531 | consistent with the state between the nth and (n + 1)th call. |
1532 | |
1533 | This permits a work around in Perl_padlist_dup() to avoid leaking active pad |
1534 | data into a new thread, whilst still correctly bodging the external references |
1535 | needed by the current ?{} implementation. Fix that, and this can be removed. |
1536 | |
674803eb |
1537 | commit 05d04d9c74ee968bace5e063c9ded74f94b3df24 # DOCed |
91c261ec |
1538 | Author: Nicholas Clark <nick@ccl4.org> |
1539 | Date: Thu Feb 25 21:35:39 2010 +0000 |
1540 | |
1541 | Don't clone the contents of lexicals in pads. |
1542 | |
1543 | This stops the values of lexicals in active stack frames in the parent leaking |
1544 | into the lexicals in the child thread. |
1545 | |
1546 | With an exception for lexicals with a reference count of > 1, to cope with the |
1547 | implementation of ?{{ ... }} blocks in regexps. :-( |
1548 | |
960deebc |
1549 | commit 4cee4ca8b28e9dadc530df8ce100439bc4a78e73 # NODOC |
91c261ec |
1550 | Author: Nicholas Clark <nick@ccl4.org> |
1551 | Date: Thu Feb 25 16:56:53 2010 +0000 |
1552 | |
1553 | In Perl_pad_tidy(), merge the SvPADTMP_on() loops for padtidy_SUB and _FORMAT. |
1554 | |
173e4c4c |
1555 | commit 6de654a5795b6f7915432ff16bcdac0688492a9b # DOCed |
91c261ec |
1556 | Author: Nicholas Clark <nick@ccl4.org> |
1557 | Date: Thu Feb 25 14:21:18 2010 +0000 |
1558 | |
1559 | In Perl_padlist_dup() don't duplicate @_ or pads caused by recursion. |
1560 | |
1561 | CvDEPTH() is 0 in a new thread, so duplicating pads beyond the always-present |
1562 | first level is a waste of effort and memory. |
1563 | |
d383dd92 |
1564 | commit d5b1589c09b534ccfeb2eae26b3de9339c1bf22b # NODOC |
91c261ec |
1565 | Author: Nicholas Clark <nick@ccl4.org> |
1566 | Date: Wed Feb 10 09:57:23 2010 +0000 |
1567 | |
1568 | Convert PAD_DUP to a function Perl_padlist_dup(). |
1569 | |
1570 | assert() that pads are never AvREAL(). |
1571 | |
25e6a476 |
1572 | commit e42956688f2e0df936f1a42811962946e4e185bf # DOC (verify needed) |
91c261ec |
1573 | Author: Nicholas Clark <nick@ccl4.org> |
1574 | Date: Thu Feb 25 11:12:03 2010 +0000 |
1575 | |
1576 | When deleting CLONE_PARAMS, push any unreferenced SVs onto the temps stack. |
1577 | |
1578 | Effectively this leaves the cloned-into interpreter in a consistent state. |
1579 | In the cloned-from interpreter, the SV targets of non-reference owning pointers |
1580 | *are* referenced and managed by other pointers. SvREFCNT() == 0 SVs in the |
1581 | cloned-into interpreter result from the non-reference owning pointers being |
1582 | found and followed, but the reference owning and managing pointers not being |
1583 | part of the subsection of interpreter state cloned over. Hence, this change |
1584 | creates reference owning pointers to this SVs on the temps stack, which ensures |
1585 | that they are correctly cleaned up, and don't "leak" until interpreter |
1586 | destruction. (Which might be some time away, in a persistent process.) |
1587 | |
25e6a476 |
1588 | commit d08d57ef17162c52e2024a3ba6755f778acbc697 # NODOC |
91c261ec |
1589 | Author: Nicholas Clark <nick@ccl4.org> |
1590 | Date: Wed Feb 24 17:15:41 2010 +0000 |
1591 | |
1592 | Better ithreads cloning - add all SVs with a 0 refcnt to the temps stack. |
1593 | |
1594 | Track all SVs created by sv_dup() that have a 0 reference count. If they still |
1595 | have a 0 reference count at the end of cloning, assign a reference to each to |
1596 | the temps stack. As the temps stack is cleared at thread exit, SVs book keeping |
1597 | will be correct and consistent before perl_destruct() makes its check for |
1598 | leaked scalars. |
1599 | |
1600 | Remove special case code for checking each @_ and the parent's temp stack. |
1601 | |
32b57737 |
1602 | commit 1db366cc74404c47243e1d86efa59c6559db818e # NODOC |
91c261ec |
1603 | Author: Nicholas Clark <nick@ccl4.org> |
1604 | Date: Mon May 24 15:48:06 2010 +0100 |
1605 | |
1606 | Cleaner implementations for Perl_clone_params_{new,del} |
1607 | |
1608 | Not source or binary compatible with maint-5.12. |
1609 | |
d2d72832 |
1610 | commit f7abe70be985cb9179c2e728a593cb8a5c8e049d # NODOC (verify) |
91c261ec |
1611 | Author: Nicholas Clark <nick@ccl4.org> |
1612 | Date: Wed Feb 24 11:47:08 2010 +0000 |
1613 | |
1614 | Abstract *correct* initialisation of CLONE_PARAMS into Perl_clone_params_new(). |
1615 | |
1616 | As it allocates memory dynamically, add Perl_clone_params_del(). This will |
1617 | allow CLONE_PARAMS to be expand in future in a source and binary compatible |
1618 | fashion. |
1619 | |
1620 | These implementations of Perl_clone_params_new()/Perl_clone_params_del() jump |
1621 | through hoops to remain source and binary compatible, in particular, by not |
1622 | assuming that the structure member is present and correctly initialised. Hence |
1623 | they should be suitable for inclusion into Devel::PPPort. |
1624 | |
1625 | Convert threads.xs to use them, resolving RT #73046. |
1626 | |
9aa6229b |
1627 | commit a09252eb79f700c93c37322c1ad831cf3193571b # HALPHALP |
91c261ec |
1628 | Author: Nicholas Clark <nick@ccl4.org> |
1629 | Date: Tue Feb 23 14:48:17 2010 +0000 |
1630 | |
1631 | Convert Perl_sv_dup_inc() from a macro to a real function. |
1632 | |
cfbf992c |
1633 | commit 842c41230043ce99d4bf7b2c79aed85ce2908e89 # NODOC |
91c261ec |
1634 | Author: Nicholas Clark <nick@ccl4.org> |
1635 | Date: Wed Feb 24 15:07:51 2010 +0000 |
1636 | |
1637 | In perl_clone_using(), turn off AvREAL() on param->stashes. |
1638 | |
1639 | This is an optimisation, not a bug fix. |
1640 | |
850c82be |
1641 | commit f5d8aca1e51e677519775b879a4d41b6f78d07d2 # NODOC |
91c261ec |
1642 | Author: Zefram <zefram@fysh.org> |
1643 | Date: Mon May 24 07:38:24 2010 -0400 |
1644 | |
1645 | Add package block syntax to perldelta |
1646 | |
850c82be |
1647 | commit c4a6534165d0545c5a1273d7febaa02f28023972 # NODOC |
91c261ec |
1648 | Author: David Golden <dagolden@cpan.org> |
1649 | Date: Sun May 23 20:04:38 2010 -0400 |
1650 | |
1651 | Add perldelta entry for non-destructive substitution |
1652 | |
01154e8f |
1653 | commit 196007d1d65f1413c42844a9656643290de75eb5 # NODOC |
91c261ec |
1654 | Author: Nicholas Clark <nick@ccl4.org> |
1655 | Date: Sun May 23 21:09:13 2010 +0100 |
1656 | |
1657 | Change the API documentation from sv_2nv to sv_2nv_flags. |
1658 | |
1659 | Addendum to 39d5de13bc6d138b. |
1660 | |
7cd782e5 |
1661 | commit 702646f25c38c963340344fc4125d616ff143857 # ALREADY |
91c261ec |
1662 | Author: Zefram <zefram@fysh.org> |
1663 | Date: Sat May 22 23:10:16 2010 +0100 |
1664 | |
1665 | fix MAD handling of package block syntax |
1666 | |
1667 | There was a broken TOKEN_GETMAD attempting to handle the label preceding a |
1668 | package-block statement, where no label is actually possible. The correct |
1669 | behaviour for no label is a no-op, so just remove the TOKEN_GETMAD. |
1670 | |
01154e8f |
1671 | commit ebe8e111d12294e0144e3f276418ccacc4ff3fe2 # NODOC |
91c261ec |
1672 | Author: David Mitchell <davem@iabyn.com> |
1673 | Date: Sun May 23 15:04:37 2010 +0100 |
1674 | |
1675 | add perldelta entry for overload fixes |
1676 | |
541ee700 |
1677 | commit 847428bd7826cc0cc873c7440aba67dffca8e984 # NODOC |
91c261ec |
1678 | Author: David Golden <dagolden@cpan.org> |
1679 | Date: Sun May 23 07:51:04 2010 -0400 |
1680 | |
1681 | Format perlhist consistently by maint/dev track |
1682 | |
541ee700 |
1683 | commit f83c51e56b41b58bf562b416ffc86a6548df8f88 # NODOC |
91c261ec |
1684 | Author: David Golden <dagolden@cpan.org> |
1685 | Date: Sun May 23 07:22:32 2010 -0400 |
1686 | |
1687 | create perl5132delta |
1688 | |
541ee700 |
1689 | commit 000c65fce914409ad42f49763dbced48187b5baf # ALREADY |
91c261ec |
1690 | Author: David Golden <dagolden@cpan.org> |
1691 | Date: Sat May 22 22:46:50 2010 -0400 |
1692 | |
1693 | Rephrase perlop for non-destructive substitution |
1694 | |
541ee700 |
1695 | commit 5901ab6e45f075cff484db49565241f42c2202a5 # NODOC |
91c261ec |
1696 | Author: David Golden <dagolden@cpan.org> |
1697 | Date: Sat May 22 22:31:38 2010 -0400 |
1698 | |
1699 | Add David Caldwell to AUTHORS |
1700 | |
541ee700 |
1701 | commit 4f4d7508b0c2c114e5f52420e0e87a853c5f642a # ALREADY |
91c261ec |
1702 | Author: David Caldwell <david@porkrind.org> |
1703 | Date: Mon Nov 23 17:24:25 2009 -0800 |
1704 | |
1705 | Add s///r (non-destructive substitution). |
1706 | |
1707 | This changes s/// so that it doesn't act destructively on its target. |
1708 | Instead it returns the result of the substitution (or the original string if |
1709 | there was no match). |
1710 | |
1711 | In addition this patch: |
1712 | |
1713 | * Adds a new warning when s///r happens in void context. |
1714 | * Adds a error when you try to use s///r with !~ |
1715 | * Makes it so constant strings can be bound to s///r with =~ |
1716 | * Adds documentation. |
1717 | * Adds some tests. |
1718 | * Updates various debug code so it knows about the /r flag. |
1719 | * Adds some new 'r' words to B::Deparse. |
1720 | |
541ee700 |
1721 | commit dd9035cd5bdeced1187df399d27d526f3b30194b # NODOC |
91c261ec |
1722 | Author: Nicholas Clark <nick@ccl4.org> |
1723 | Date: Tue Nov 17 10:17:11 2009 +0000 |
1724 | |
1725 | Restore strict refs on stashes, removed by ce10b5d1ec5b5f68b0811018a415bc37. |
1726 | |
1727 | Strictures on stashes were special-cased, because the lexer was special-cased |
1728 | to avoid autovivifying stashes (in adc51b978ed1b2e9d4512c9bfa80386ac917d05a, |
1729 | now reverted), to keep the behaviour of defined %stash:: consistent. |
1730 | |
1731 | defined %stash:: is deprecated. |
1732 | |
1733 | Keep the new regression tests from ce10b5d1ec5b5f68b0811018a415bc37bb6bfe5e. |
1734 | |
dee6ca0d |
1735 | commit 223f0fb7579941d15c6185ada7e98fb53bf60852 # HALPHALP |
91c261ec |
1736 | Author: Nicholas Clark <nick@ccl4.org> |
1737 | Date: Sat Oct 24 12:22:01 2009 +0100 |
1738 | |
1739 | Re-instate the use of gv_stashpvn_flags(), and the correct non-boolean argument. |
1740 | |
1741 | This restores the change of 9bde8eb087a2c05d4c8b0394a59d28a09fe5f529. |
1742 | |
a7f63c6f |
1743 | commit 0cc522c3af5e9943da81cc0c962abc67d3df6925 # NODOC (verify) |
91c261ec |
1744 | Author: Nicholas Clark <nick@ccl4.org> |
1745 | Date: Sat Oct 24 11:36:06 2009 +0100 |
1746 | |
1747 | Remove the tokeniser hack that prevents compile-time vivification of %stash:: |
1748 | |
1749 | This was put in to ensure that defined %stash:: continued to return false after |
1750 | the implementation of hashes was changed, such that stashes were always defined. |
1751 | defined %stash:: is deprecated. |
1752 | |
1753 | This reverts the tokeniser changes of adc51b978ed1b2e9d4512c9bfa80386ac917d05a, |
1754 | 76138434928a968a390c791aec92e5f00017d01d, |
1755 | d6069db2e52f58ef65bf59f2fd453604270d2205 and part of |
1756 | 9bde8eb087a2c05d4c8b0394a59d28a09fe5f529, and updates the tests added with those |
1757 | commits to reflect the restored (but as yet unreleased) behaviour. |
1758 | |
1759 | I don't think that this should be merged to blead until after 5.12.0 ships, |
1760 | with the enabled deprecation warnings on defined %hash, as it changes subtle |
1761 | behaviour that all current released stable perls accept without warning. |
1762 | |
901c2e6a |
1763 | commit 4fa7c2bfe0a54ceffcc7c56cdc072eaeaf54cad9 # NODOC |
91c261ec |
1764 | Author: Nicholas Clark <nick@ccl4.org> |
1765 | Date: Fri May 21 20:32:28 2010 +0100 |
1766 | |
1767 | PL_in_load_module only has values 0 and 1, so can be a bool instead of int. |
1768 | |
e54740d0 |
1769 | commit fcd24582e9751804a977b6d4ef227de5a3b0343b # DOCed |
91c261ec |
1770 | Author: Nicholas Clark <nick@ccl4.org> |
1771 | Date: Sat Mar 13 11:23:46 2010 +0000 |
1772 | |
1773 | Perl_hv_fill(), count empty chains down, rather than used chains up. |
1774 | |
1775 | The assumption is that most chains of a hash are in use. |
1776 | |
1777 | Suggestion and initial patch by Ruslan Zakirov. |
1778 | |
e54740d0 |
1779 | commit 359164a0b0ca9d7142b89ef0b09c1d01437e4471 # DOCed |
91c261ec |
1780 | Author: Nicholas Clark <nick@ccl4.org> |
1781 | Date: Sun Jan 24 15:39:53 2010 +0000 |
1782 | |
1783 | Eliminate xhv_fill from struct xpvhv. |
1784 | |
e54740d0 |
1785 | commit 4d0fbddde6c5dcb972786d09de0cab6e93056b88 # DOCed |
91c261ec |
1786 | Author: Nicholas Clark <nick@ccl4.org> |
1787 | Date: Sun Jan 24 15:07:50 2010 +0000 |
1788 | |
1789 | Make HvFILL() count the allocated buckets, instead of reading a stored value. |
1790 | |
1791 | Add a function Perl_hv_fill to perform the count. This will save 1 IV per hash, |
1792 | and on some systems cause struct xpvhv to become cache aligned. |
1793 | |
e54740d0 |
1794 | commit f4431c56525a8650559872ff19c75f109a5d1190 # DOCed |
91c261ec |
1795 | Author: Nicholas Clark <nick@ccl4.org> |
1796 | Date: Sun Jan 24 10:50:02 2010 +0000 |
1797 | |
1798 | Replace boolean use of HvFILL(hv) with HvTOTALKEYS(hv), which is equivalent. |
1799 | |
cd975843 |
1800 | commit ed3b9b3c212f717939207379cdb328156dd4a01e # ALREADY |
91c261ec |
1801 | Author: David Mitchell <davem@iabyn.com> |
1802 | Date: Fri May 21 17:07:40 2010 +0100 |
1803 | |
1804 | followup to magic/overload fix |
1805 | |
1806 | 6f1401dc2acd2a2b85df22b0a74e5f7e6e0a33aa was over-enthusiastic |
1807 | on removing redundant code in the comparison ops. This code was only used |
1808 | on 64-bit #ifdef branches which is why I failed to spot it earlier. |
1809 | So restore that code! |
1810 | |
cd975843 |
1811 | commit 6f1401dc2acd2a2b85df22b0a74e5f7e6e0a33aa # ALREADY |
91c261ec |
1812 | Author: David Mitchell <davem@iabyn.com> |
1813 | Date: Fri May 21 14:18:21 2010 +0100 |
1814 | |
1815 | make overload respect get magic |
1816 | |
1817 | In most places, ops checked their args for overload *before* doing |
1818 | mg_get(). This meant that, among other issues, tied vars that |
1819 | returned overloaded objects wouldn't trigger calling the |
1820 | overloaded method. (Actually, for tied and arrays and hashes, it |
1821 | still often would since mg_get gets called beforehand in rvalue |
1822 | context). |
1823 | |
1824 | This patch does the following: |
1825 | |
1826 | Makes sure get magic is called first. |
1827 | |
1828 | Moves most of the overload code formerly included by macros at the |
1829 | start of each pp function into the separate helper functions |
1830 | Perl_try_amagic_bin, Perl_try_amagic_un, S_try_amagic_ftest, |
1831 | with 3 new wrapper macros: |
1832 | tryAMAGICbin_MG, tryAMAGICun_MG, tryAMAGICftest_MG. |
1833 | This made the code 3800 bytes smaller. |
1834 | |
1835 | Makes sure that FETCH is not called multiple times. Much of this |
1836 | bit was helped by some earlier work from Father Chrysostomos. |
1837 | |
1838 | Added new functions and macros sv_inc_nomg(), sv_dec_nomg(), |
1839 | dPOPnv_nomg, dPOPXiirl_ul_nomg, dPOPTOPnnrl_nomg, dPOPTOPiirl_ul_nomg |
1840 | dPOPTOPiirl_nomg, SvIV_please_nomg, SvNV_nomg (again, some of |
1841 | these were based on Father Chrysostomos's work). |
1842 | |
1843 | Fixed the list version of the repeat operator (x): it now only |
1844 | calls overloaded methods for the scalar version: |
1845 | (1,2,$overloaded) x 10 |
1846 | no longer erroneously calls |
1847 | x_method($overloaded,10)) |
1848 | |
1849 | The only thing I haven't checked/fixed yet is overloading the |
1850 | iterator operator, <>. |
1851 | |
078a9fc1 |
1852 | commit c4648999f2aa0b971b46a580c1258b719394072a # HALPHALP |
91c261ec |
1853 | Author: Nicholas Clark <nick@ccl4.org> |
1854 | Date: Tue Jan 26 17:03:41 2010 +0000 |
1855 | |
1856 | Remove union _xivu from _XPVCV_COMMON, and hence structs xpvcv and xpvfm |
1857 | |
1858 | Replaced with xcv_depth and xfm_lines respectively. Both structures might |
1859 | benefit from some field re-ordering. |
1860 | |
1861 | Update the descriptive comments in the definition of union _xivu. |
1862 | |
078a9fc1 |
1863 | commit 89f6f2871ea81b2c6a09311796395c610f3e3464 # HALPHALP |
91c261ec |
1864 | Author: Nicholas Clark <nick@ccl4.org> |
1865 | Date: Tue Jan 26 16:28:59 2010 +0000 |
1866 | |
1867 | Remove union _xivu from struct regexp - replace it with a non-union paren_names. |
1868 | |
1869 | This was the only user of xivu_hv in union _xivu, so remove that too. |
1870 | |
078a9fc1 |
1871 | commit de0a224a057997a65d38856f1981702fca5d7c18 # HALPHALP |
91c261ec |
1872 | Author: Nicholas Clark <nick@ccl4.org> |
1873 | Date: Mon Jan 25 14:28:30 2010 +0000 |
1874 | |
1875 | Remove union _xivu from struct xpvhv - replace it with a non-union xav_keys. |
1876 | |
078a9fc1 |
1877 | commit 4f7003f5e8adb225a16f3788ec81e193021f41e5 # HALPHALP |
91c261ec |
1878 | Author: Nicholas Clark <nick@ccl4.org> |
1879 | Date: Mon Jan 25 13:55:28 2010 +0000 |
1880 | |
1881 | Remove union _xivu from struct xpvav - replace it with a non-union xav_alloc. |
1882 | |
1883 | This was the only user of xivu_p1 in union _xivu, so remove that too. |
1884 | |
078a9fc1 |
1885 | commit 889d28b2ea2c17517fae97cf4a92bd3563aaa74f # HALPHALP |
91c261ec |
1886 | Author: Nicholas Clark <nick@ccl4.org> |
1887 | Date: Mon Jan 25 13:30:16 2010 +0000 |
1888 | |
1889 | Reinstate space optimisations to SV body structures. |
1890 | |
078a9fc1 |
1891 | commit 6e1287864cd02fb6f429f12a995e1c08e03572b9 # HALPHALP |
91c261ec |
1892 | Author: Nicholas Clark <nick@ccl4.org> |
1893 | Date: Mon Jan 25 09:27:56 2010 +0000 |
1894 | |
1895 | In the SV body, exchange the positions of the NV and stash/magic. |
1896 | |
078a9fc1 |
1897 | commit 601dfd0af0604fa7ed2cabe7f33a232b9b3908c0 # HALPHALP |
91c261ec |
1898 | Author: Nicholas Clark <nick@ccl4.org> |
1899 | Date: Fri Jan 22 15:31:49 2010 +0000 |
1900 | |
1901 | Remove all space optimisations from SV body structures. |
1902 | |
c5cc31c3 |
1903 | commit 89c1d9143a1f7d921350481bd9e3f617a2988f9c # NODOC |
91c261ec |
1904 | Author: H.Merijn Brand <h.m.brand@xs4all.nl> |
1905 | Date: Fri May 21 08:01:54 2010 +0200 |
1906 | |
1907 | The size of a character in C is per definition 1 |
1908 | |
04c16a5e |
1909 | commit 23ae7f173577cbe2d62a4f7d64340f6457c75ee3 # DOCed |
91c261ec |
1910 | Author: Sisyphus <sisyphus1@optusnet.com.au> |
1911 | Date: Thu May 20 17:56:07 2010 -0700 |
1912 | |
1913 | Fix CCINCDIR and CCLIBDIR for mingw64 cross compiler |
1914 | |
1915 | When building perl with the mingw64 x64 cross-compiler 'incpath', 'libpth', |
1916 | 'ldflags', 'lddlflags' and 'ldflags_nolargefiles' values in Config.pm and |
1917 | Config_heavy.pl are not being set correctly because, with that compiler, the |
1918 | include and lib directories are not immediately below $(CCHOME). |
1919 | |
3afe9b45 |
1920 | commit 36f77d7116441b32ff6c10307a116e105b7eabb4 # ALREADY |
91c261ec |
1921 | Author: Zefram <zefram@fysh.org> |
1922 | Date: Thu May 20 21:48:03 2010 +0100 |
1923 | |
1924 | fully test package-version-block syntax |
1925 | |
1926 | Extend the exhaustive package-version tests in t/op/packagev.t |
1927 | to test each case using package-block syntax in addition to the |
1928 | package-declaration syntax. |
1929 | |
3afe9b45 |
1930 | commit 3afb2f14ba09da7b54ce62a6f12d9703a7776666 # ALREADY |
91c261ec |
1931 | Author: Zefram <zefram@fysh.org> |
1932 | Date: Sat May 1 17:51:21 2010 +0100 |
1933 | |
1934 | additional tests for package block syntax |
1935 | |
1936 | Test that __PACKAGE__ propagates into string eval correctly. Test that |
1937 | __LINE__ is correct. Test that goto into and out of package blocks |
1938 | works correctly. |
1939 | |
3afe9b45 |
1940 | commit af5050ac60ae083ebc5a0b2a1b4289c51c622f95 # ALREADY |
91c261ec |
1941 | Author: Zefram <zefram@fysh.org> |
1942 | Date: Sat May 1 17:30:26 2010 +0100 |
1943 | |
1944 | fix SEGV with eval("package Foo {") |
1945 | |
1946 | OPs relating to the package name and version were subject to double |
1947 | freeing during error recovery from an incomplete package block. Fixed by |
1948 | using the op_latefree mechanism to shift the op free time. |
1949 | |
3afe9b45 |
1950 | commit 4e4da3acc11d96d134ed1dc0effd641e7bedb0ca # ALREADY |
91c261ec |
1951 | Author: Zefram <zefram@fysh.org> |
1952 | Date: Mon Apr 19 20:37:25 2010 +0100 |
1953 | |
1954 | support "package Foo { ... }" |
1955 | |
1956 | Package block syntax limits the scope of the package declaration to the |
1957 | attached block. It's cleaner than requiring the declaration to come |
1958 | inside the block. |
1959 | |
1acc4234 |
1960 | commit 39f3f7f442aed93239540238d19a15f6020da747 # DOCed as perf enhancement |
91c261ec |
1961 | Author: Nicholas Clark <nick@ccl4.org> |
1962 | Date: Tue Feb 23 20:35:29 2010 +0000 |
1963 | |
1964 | PL_endav can be NULL, so in S_ithread_create() no need to set it to newAV(). |
1965 | |
1acc4234 |
1966 | commit b1faab355535118c9fb99fcc343501012923ece # DOCed as perf enhancement |
91c261ec |
1967 | Author: Nicholas Clark <nick@ccl4.org> |
1968 | Date: Sat Feb 13 09:05:18 2010 +0000 |
1969 | |
1970 | Remove redundant hv_exists() calls from ithread_create()'s spec parser. |
1971 | |
1972 | hv_fetch(..., 0) won't create the element if it doesn't exist, returning a NULL |
1973 | pointer, so hv_exists() and hv_fetch() is doing two hash lookups where one would |
1974 | suffice. |
1975 | |
1976 | On this machine, reduces the object code by 3K, about 7%. Everyone's a winner. |
1977 | |
1acc4234 |
1978 | commit 4cf5eae5e58faebb6ae9e68e493cd85343f7a76c # DOCed as perf enhancement |
91c261ec |
1979 | Author: Nicholas Clark <nick@ccl4.org> |
1980 | Date: Sat Feb 13 08:31:55 2010 +0000 |
1981 | |
1982 | Change S_ithread_create() params from a single AV* to a pair of SV** pointers. |
1983 | |
1984 | This saves creating, duplicating and freeing and AV, which is only ever used for |
1985 | an internal calling convention. |
1986 | |
1acc4234 |
1987 | commit 78b7eff23b1757f7fccef902e7e706080a997e2c # DOCed as perf enhancement |
91c261ec |
1988 | Author: Nicholas Clark <nick@ccl4.org> |
1989 | Date: Sat Feb 13 08:01:45 2010 +0000 |
1990 | |
1991 | In threads.xs, convert thread->params from RV to AV. |
1992 | |
1993 | Pass around and store the array directly, rather than creating, holding and |
1994 | dereferencing a reference to it. |
1995 | |
012fc879 |
1996 | commit e77da3b2183d0f9c5ea81dc5780d9a480fc4fa88 # DOCed (verify) |
91c261ec |
1997 | Author: Nicholas Clark <nick@ccl4.org> |
1998 | Date: Tue Jan 26 12:00:56 2010 +0000 |
1999 | |
2000 | SvIVX() isn't valid on SVt_REGEXP |
2001 | |
05d5c3cb |
2002 | commit 65ac1738675fbcf49a4c9d625c0c43dd73e6ff2f # NODOC (verify) |
91c261ec |
2003 | Author: Nicholas Clark <nick@ccl4.org> |
2004 | Date: Mon Jan 25 10:53:33 2010 +0000 |
2005 | |
2006 | In sv.c, _all_ {new,del}_X* macros can be *_body_allocated. |
2007 | |
2008 | Previously those where bodies_by_type[sv_type].offset was zero were using |
2009 | {new,del}_body_typed. However, the optimiser can spot this, and generates the |
2010 | same object code. This allows simplification of the C code, and more |
2011 | flexibility to rearrange the structures without generating bugs. |
2012 | |
73a7bec3 |
2013 | commit ef01b67f31004ca818c73987384945a0fea605d6 # NODOC |
91c261ec |
2014 | Author: Ricardo Signes <rjbs@cpan.org> |
2015 | Date: Thu May 20 10:14:39 2010 -0400 |
2016 | |
2017 | update Ricardo Signes's email |
2018 | |
73a7bec3 |
2019 | commit faee19b51573e81abe8811f1256a1d27777d6d04 # DOCed |
91c261ec |
2020 | Author: Rafael Garcia-Suarez <rgs@consttype.org> |
2021 | Date: Wed May 19 20:36:12 2010 +0200 |
2022 | |
2023 | Do not try to load a feature bundle when doing "no VERSION" |