ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Coro/Changes
(Generate patch)

Comparing Coro/Changes (file contents):
Revision 1.89 by pcg, Sat Feb 14 13:01:55 2004 UTC vs.
Revision 1.267 by root, Mon Oct 8 00:47:19 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
30.95 Fri Dec 12 20:50:57 CET 2003 3TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO.
4TODO: maybe implement a default message channel, very much like Erlang
5 (which is cool in a lot of important aspects (failures!),
6 but very lacking in others (higher level ipc)).
7
84.1
9 - incompatible change: $SIG{__DIE__} and $SIG{__WARN__} will now
10 be local to each coro (see Coro::State).
11 - possibly bring back 5.10 compatibility (untested).
12 - work around stupid (and wrong) warning on 5.10 :(.
13 - overlay the saved state over the context stack. This saves
14 a few hundred bytes per coroutine on average and also
15 speeds up context switching a bit.
16 - further tune default stack sizes.
17 - (more) correctly calculate stack usage in coro_rss.
18 - Coro::Storable::blocking_* did not properly lock
19 resulting in races between coroutines.
20 - added Coro::Storable::guard.
21
224.03 Sat Oct 6 21:24:00 CEST 2007
23 - added Coro::throw method.
24 - minor code cleanups.
25
264.02 Sat Oct 6 02:36:47 CEST 2007
27 - fix a very minor per-coroutine memleak (a single codereference).
28 - fixed a bug where the currently in-use c context would be freed
29 prematurely (can happen only when programs change the stacksize
30 or use tracing).
31 - tracing can no longer keep a coro alive after it terminated.
32 - do static branch prediction in the common path for gcc. gives
33 about 2-5% speed improvement here.
34
354.01 Fri Oct 5 22:10:49 CEST 2007
36 - instead of recreating *a* standard output handle we simply
37 use STDOUT, which is faster and hopefully more robust.
38
394.0 Fri Oct 5 12:56:00 CEST 2007
40 - incompatibly changed Coro::Storable::freeze.
41 - major new feature: added Coro::Debug, for interactive coroutine
42 debugging, tracing and much more.
43 - major bug fix: unbelievable, but true: $_, $/ and many other
44 "saved" variables actually weren't being saved. This has been fixed,
45 of course, while increasing performance while losing all the save
46 flags.
47 - save flags are gone, and all the api functions dealing with them.
48 - added Coro::Semaphore::adjust.
49 - added Coro::Util::fork_eval.
50 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
51 - added Coro::killall.
52 - reduce initial stack sizes to allow for "micro-coroutines".
53 - better async_pool resource management, moved parts of async_pool
54 handling to XS (major speed improvement).
55 - actually croak before modifying important data structures.
56 - refuse to transfer while compiling.
57 - possibly support eval EXPR better now.
58 - enable assembly per default on linux+bsd x86+amd64.
59 - all internal members were renamed _something for easier subclassing.
60 - many minor tweaks.
61
623.63 Wed May 16 14:10:06 CEST 2007
63 - implement handcoded assembly for x86/amd64 SVR ABI.
64
653.62 Fri Apr 27 21:36:06 CEST 2007
66 - upgrade libcoro (which might set unwind info correctly).
67 - change default on linux to setjmp/longjmp.
68
693.61 Thu Apr 19 12:36:18 CEST 2007
70 - Coro::Storable caused an endless loop when thawing invalid
71 pst's sometimes.
72 - use a Semaphore in Coro::Storable, as Storable doesn't
73 seem to be reentrant (although it is documented to
74 be threadsafe...).
75 - fix Coro::Signal to bring back the original unreliable
76 but stateful semantics.
77 - fixed a lot of typos in Coro.pm (patch submitted by David
78 Steinbrunner, which applied flawlessly).
79
803.6 Sat Apr 14 17:13:31 CEST 2007
81 - added some bugfixes to get eg/myhttpd working again.
82 - added Coro::Storable for often-cede'ing freeze/thaw.
83 - try to do a clean exit when a coroutine calls exit
84 (EXPERIMENTAL).
85 - got rid of indirect call through _coro_init.
86 - updated the partly antique examples in eg/ to
87 work again and be a bit less magic, too.
88 - fixed Coro::Signal semantics to work as documented again.
89
903.55 Sun Mar 25 01:20:47 CET 2007
91 - add SAVE_DEFFH to save the default filehandle and enable
92 it by default.
93 - finally move socket-operations from Coro::Socket to Coro::Handle
94 to be able to unblock foreign sockets.
95 - add Coro::State::save_also and guarded_save.
96 - add count accessor to Coro::Semaphore.
97 - add Coro::State::cctx_stacksize.
98 - just for the fun of it, do not rely on implicit context,
99 which can dramatically improve performance, but people
100 using windows-process-emulation perls do not care much about
101 performance.
102
1033.51 Sun Mar 4 14:18:04 CET 2007
104 - fixed a problem when you weakened references to Coro::State's
105 (patch by Yuval Kogman).
106
1073.501 Wed Feb 28 12:44:07 CET 2007
108 - rename some global symbols as macosx from hell redefines
109 them without asking.
110
1113.5 Tue Feb 13 20:22:53 CET 2007
112 - do AnyEvent model detection earlier, avoiding problems
113 caused by first using AnyEvent and later Coro::Event.
114 - implement and document Coro::Event event objects.
115 - fix a potential problem in Coro::Event causing crashes.
116 - initialise PL_comppad when creating a new coroutine,
117 avoids crashes on early coro destruction.
118
1193.41 Mon Jan 22 19:19:49 CET 2007
120 - readline on Coro::Handle did not support undefined $/,
121 nor did it deliver partial lines on EOF or error.
122 - implement malloc fallback for stack allocation because
123 stupid broken idiotic OSX has a stupid broken
124 idiotic fits-the-whole-os mmap "implementation" and
125 my dick feels longer if Coro is portable even to
126 obsolete platforms.
127
1283.4 Fri Jan 19 21:52:54 CET 2007
129 - remove t/09_timer.t, as it isn't really testing much
130 but was rather flaky in practise.
131 - async_pool coro would keep arguments and callback alive until
132 it was reused.
133 - cancellation of a coroutine could cause spurious idle calls
134 in cede_notself.
135
1363.3 Sat Jan 6 03:45:00 CET 2007
137 - implement $coro->on_destroy.
138 - Coro::Event blocking semantics have been changed,
139 documented and - hopefully - improved.
140 - fix nice adding, not subtracting, from priority.
141 - fix ->prio and api_is_ready (patch by Mark Hinds).
142 - fixed an assert ("... == prev__cctx->idle_te")
143 that could errronously trigger.
144 - fix various large and small memleaks.
145 - use a (hopefully) more stable cancel implementation
146 that immediately frees the coroutine data.
147 - cede/cede_notself return a status now.
148 - added Coro::guard function.
149 - added a global coroutine pool for jobs (on my machine,
150 I can create and execute 48k simple coros/s with async,
151 and 128k coros with async_pool).
152 - Coro::AIO now uses the coroutine priority as io priority.
153
1543.2 Fri Dec 22 05:07:09 CET 2006
155 - improve portability to slightly older perls.
156 - use cleaner coroutine destruction.
157 - simplify configuration for users.
158 - optionally (unrecommended) prefer perl functions over
159 their coro replacements.
160
1613.11 Tue Dec 5 13:11:24 CET 2006
162 - fixed some bogus assert's, but as perl.h disables assert even
163 without NDEBUG (thank you very much), not too many people should
164 notice (that did include myself). Andreas König noticed, though :)
165 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
166 it. Also noticed by Andreas König :)
167 - save/restore tainted status.
168 - verified to pass the testsuite on my 5.9.5.
169
1703.1 Mon Dec 4 23:03:40 CET 2006
171 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
172 - incompatible change: transfer flags are now per-state.
173 - give a better error message on deadlock.
174 - document Coro::nready.
175 - enhanced testsuite.
176
1773.01 Sun Dec 3 23:47:42 CET 2006
178 - forgot to include Coro::Timer.
179
1803.0 Sun Dec 3 22:57:25 CET 2006
181 - the "FINALLY COMPLETELY STABLE" release (coming soon:
182 the "FAMOUS LAST WORDS" release).
183 - implement a new stack sharing algorithm, which uses a stack
184 pool (size currently hardcoded).
185 - make stack sharing mandatory (it no longer uses a heuristic).
186 - eval/die no longer cause weird problems under heavy use.
187 - Coro::Event could cause livelocks if it was used but
188 no Coro::Event watchers were used.
189 - Coro::Event now uses asynccheck as prepare does not
190 check for changed watchers.
191 - Coro::Event allows multiple waiting coros and will wake up one
192 per event.
193 - Coro::Event should be cleaner and more efficient now.
194 - new utility function Coro::unblock_sub.
195 - document the sad fact that Event is no longer reentrant.
196 - putting a coroutine into the ready queue twice could under
197 some circumstances lead to stack corruption.
198 - minor incompatibility: subclassing Coro::State is supported
199 directly now without going through a _coro_state member.
200 - state/coro switching is much faster now.
201 - very minor optimisations and code/documentation cleanup.
202 - avoid problems due to compiler inlining.
203 - removed timers from Coro::Timer -> use AnyEvent instead.
204 - replaced Coro::idle coroutine by (cleaner) idle handler.
205 - updated to newest libcoro.
206 - implement enhanced support for valgrind.
207 - implement is_ready and return value for ready.
208 - removed Coro::Cont, it was a misnomer (it's generators),
209 and it was somewhat annoying to get it right. Will come back
210 if somebody asks for it :)
211 - many ->wait methods and Coro::Event could return
212 spuriously without the event having happened.
213
2142.5 Tue Nov 7 12:22:33 CET 2006
215 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
216 use AnyEvent, moved them to Coro/.
217 - added Coro::LWP which contains all the uglyness required to
218 make LWP non-blocking.
219 - should work with perl 5.9.x now (Andreas König made me do it).
220 - fixed another bug in Coro::Select when the timeout was undef.
221 - reuse PL_start_env for all coros, saves some memory per coroutine.
222 - manage PL_top_env differently, hopefully to avoid panic: top_env.
223 - timeout argument was not properly used in Coro::Socket.
224 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
225 - emulate undocumented(!) functionality of IO::Socket required
226 by LWP(!!).
227 - updated eg/lwp to work with newer lwp's.
228 - remove "FATAL: uncaught exception" prefix. Coroutines that die
229 kill the whole process, just as exceptions in the main "coroutine"
230 did already.
231
2322.1 Wed Nov 1 23:01:13 CET 2006
233 - fix a long-standing bug in Coro::Select where select with
234 zero timeout would instead change the current default filehandle.
235 - use a simpler and hopefully more robust way to clone padlists
236 (uses less memory and a perl function instead of our own).
237 - coro can now create a stack guard area on many architectures.
238 - Coro::AIO properly reexports additional functions from IO::AIO.
239 - updated libcoro with a workaround for OS X,
240 pach and testing by Michael Schwern.
241
2422.0 Tue Oct 24 05:47:17 CEST 2006
243 - support additional aio requests in Coro::AIO.
244
2451.9 (never properly released due to a glitch)
246
2471.8 Thu Feb 2 00:59:06 CET 2006
248 - applied suggested patch by SAMV to avoid problems during stupid
249 mark & sweep gc run.
250 - applied patch by Scott Walters for 5.9.3 compatibility.
251
2521.7 Tue Dec 27 01:41:58 CET 2005
253 - added Coro::AIO, a thin wrapper around IO::AIO.
254 - improved Makefile.PL explanations.
255
2561.6 Mon Dec 12 21:30:05 CET 2005
257 - additionally save PM_curpm might fix as-of-yet
258 unseen problems with regex matches being attributed to the
259 wrong package.
260 - add t/10_bugs.t, which currently checks against imho broken
261 perls that use 0.26MB of stack space per Perl_magic_get
262 invocation (newer linuxes) as opposed to the 0.0002MB perl
263 normally uses.
264 - make stacksize configurable for the ultimate debian experience.
265
2661.51 Mon Dec 12 18:48:36 CET 2005
267 - remove debugging warn() accidentally left in Coro::Select.
268
2691.5 Tue Nov 29 13:32:44 CET 2005
270 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
271 - fix Coro::Select 'select'.
272 - strict'ify some modules.
273 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
274 - use XSLoader in selected modules.
275 - remove some 5.6 compatibility cruft.
276
2771.4 Tue Sep 6 00:11:05 CEST 2005
278 - libcoro did not take into account the trampoline on amd64,
279 when 'l'inux method was used.
280
2811.31 Tue Aug 30 23:31:33 CEST 2005
282 - some portability fixes/workarounds.
283
2841.3 Sat Aug 20 03:08:56 CEST 2005
285 - no code changes. module seems to work fine.
286 - improve Coro::State docs, remove reference to nonexisting
287 Coro::State::flush method.
288 - no longer autodetect windows, present it as an option instead.
289
2901.2 Mon May 16 02:00:55 CEST 2005
291 - included libcoro.c earlier in State.xs, might improve portability.
292 - use a faster and possibly more stable Coro::Cont implementation.
293 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
294 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
295 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
296
2971.11 Thu Mar 3 18:00:52 CET 2005
298 - change of contact address.
299
3001.1 Tue Feb 22 20:51:16 CET 2005
301 - support [l]inux method on amd64.
302 - allow some context switches while compiling/eval'ing.
303
3041.0 Tue Aug 10 03:47:27 CEST 2004
305 - Event 0.89 fixes the bug again.
306 - fix the remaining(??) memleak. I didn't free the cache-AV for
307 padlists, so one AV per created anon-sub leaked.
308
3090.97 Fri May 14 15:23:32 CEST 2004
310 - remove stupid gcc extension that was an experiment only anyway.
311
3120.96 Thu May 13 18:09:29 CEST 2004
313 - tell people if they have yet another broken Event variant.
314 - made Coro to work _better_ with threads (it seems that, as long
315 as you use Coros in one thread only everything will work fine.
316 Looking at the perl source the problem might be missing locking
317 between Coro and some perl internal routines. That's not easily
318 fixable).
319 - (hopefully) got rid of the static xnext inside transfer.
320
3210.95 Sun Feb 15 03:20:28 CET 2004
4 - removed Coro::State::flush, implemented a new and much 322 - removed Coro::State::flush, implemented a new and much
5 faster caching mechanism by attaching magic to the CV. 323 faster caching mechanism by attaching magic to the CV.
6 - a memleak with (real) closures remains. 324 - a memleak with (real) closures remains.
7 - dramatically reduced size of stacks to 65536 (linux/x86), after 325 - dramatically reduced size of stacks to 65536 (linux/x86), after
8 observing that even myhttpd never had more than a kilobyte 326 observing that even myhttpd never had more than a kilobyte
9 of stack in use (stack is still only physically allocated in 327 of stack in use (stack is still only physically allocated in
10 pagesize increments for systems suporting mmap). 328 pagesize increments for systems suporting mmap).
11 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby. 329 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
330 - fix for threaded perl.
12 331
130.9 Sun Nov 23 04:32:39 CET 2003 3320.9 Sun Nov 23 04:32:39 CET 2003
14 - fix a few very hard to track down but (of course) stupid 333 - fix a few very hard to track down but (of course) stupid
15 memory leaks. 334 memory leaks.
16 - stringify version number to avoid locale problems :(. 335 - stringify version number to avoid locale problems :(.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines