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

Comparing Coro/Changes (file contents):
Revision 1.98 by root, Tue Feb 22 19:51:58 2005 UTC vs.
Revision 1.250 by root, Tue Oct 2 23:27:52 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2
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
83.8
9 - incompatibly changed Coro::Storable::freeze.
10 - major new feature: added Coro::Debug, for interactive coroutine
11 debugging, tracing and much more.
12 - added Coro::Semaphore::adjust.
13 - added Coro::Util::fork_eval.
14 - added Coro::Storable::{nfreeze,blocking_{freeze,nfreeze}}.
15 - added Coro::killall.
16 - reduce initial stack sizes to allow for "micro-coroutines".
17 - better async_pool resource management, moved parts of async_pool
18 handling to XS (major speed improvement).
19 - actually croak before modifying important data structures.
20 - refuse to transfer while compiling.
21 - possibly support eval EXPR better now.
22 - enable assembly per default on linux+bsd x86+amd64.
23 - all internal members are named _something for easier subclassing.
24 - many minor tweaks.
25
263.63 Wed May 16 14:10:06 CEST 2007
27 - implement handcoded assembly for x86/amd64 SVR ABI.
28
293.62 Fri Apr 27 21:36:06 CEST 2007
30 - upgrade libcoro (which might set unwind info correctly).
31 - change default on linux to setjmp/longjmp.
32
333.61 Thu Apr 19 12:36:18 CEST 2007
34 - Coro::Storable caused an endless loop when thawing invalid
35 pst's sometimes.
36 - use a Semaphore in Coro::Storable, as Storable doesn't
37 seem to be reentrant (although it is documented to
38 be threadsafe...).
39 - fix Coro::Signal to bring back the original unreliable
40 but stateful semantics.
41 - fixed a lot of typos in Coro.pm (patch submitted by David
42 Steinbrunner, which applied flawlessly).
43
443.6 Sat Apr 14 17:13:31 CEST 2007
45 - added some bugfixes to get eg/myhttpd working again.
46 - added Coro::Storable for often-cede'ing freeze/thaw.
47 - try to do a clean exit when a coroutine calls exit
48 (EXPERIMENTAL).
49 - got rid of indirect call through _coro_init.
50 - updated the partly antique examples in eg/ to
51 work again and be a bit less magic, too.
52 - fixed Coro::Signal semantics to work as documented again.
53
543.55 Sun Mar 25 01:20:47 CET 2007
55 - add SAVE_DEFFH to save the default filehandle and enable
56 it by default.
57 - finally move socket-operations from Coro::Socket to Coro::Handle
58 to be able to unblock foreign sockets.
59 - add Coro::State::save_also and guarded_save.
60 - add count accessor to Coro::Semaphore.
61 - add Coro::State::cctx_stacksize.
62 - just for the fun of it, do not rely on implicit context,
63 which can dramatically improve performance, but people
64 using windows-process-emulation perls do not care much about
65 performance.
66
673.51 Sun Mar 4 14:18:04 CET 2007
68 - fixed a problem when you weakened references to Coro::State's
69 (patch by Yuval Kogman).
70
713.501 Wed Feb 28 12:44:07 CET 2007
72 - rename some global symbols as macosx from hell redefines
73 them without asking.
74
753.5 Tue Feb 13 20:22:53 CET 2007
76 - do AnyEvent model detection earlier, avoiding problems
77 caused by first using AnyEvent and later Coro::Event.
78 - implement and document Coro::Event event objects.
79 - fix a potential problem in Coro::Event causing crashes.
80 - initialise PL_comppad when creating a new coroutine,
81 avoids crashes on early coro destruction.
82
833.41 Mon Jan 22 19:19:49 CET 2007
84 - readline on Coro::Handle did not support undefined $/,
85 nor did it deliver partial lines on EOF or error.
86 - implement malloc fallback for stack allocation because
87 stupid broken idiotic OSX has a stupid broken
88 idiotic fits-the-whole-os mmap "implementation" and
89 my dick feels longer if Coro is portable even to
90 obsolete platforms.
91
923.4 Fri Jan 19 21:52:54 CET 2007
93 - remove t/09_timer.t, as it isn't really testing much
94 but was rather flaky in practise.
95 - async_pool coro would keep arguments and callback alive until
96 it was reused.
97 - cancellation of a coroutine could cause spurious idle calls
98 in cede_notself.
99
1003.3 Sat Jan 6 03:45:00 CET 2007
101 - implement $coro->on_destroy.
102 - Coro::Event blocking semantics have been changed,
103 documented and - hopefully - improved.
104 - fix nice adding, not subtracting, from priority.
105 - fix ->prio and api_is_ready (patch by Mark Hinds).
106 - fixed an assert ("... == prev__cctx->idle_te")
107 that could errronously trigger.
108 - fix various large and small memleaks.
109 - use a (hopefully) more stable cancel implementation
110 that immediately frees the coroutine data.
111 - cede/cede_notself return a status now.
112 - added Coro::guard function.
113 - added a global coroutine pool for jobs (on my machine,
114 I can create and execute 48k simple coros/s with async,
115 and 128k coros with async_pool).
116 - Coro::AIO now uses the coroutine priority as io priority.
117
1183.2 Fri Dec 22 05:07:09 CET 2006
119 - improve portability to slightly older perls.
120 - use cleaner coroutine destruction.
121 - simplify configuration for users.
122 - optionally (unrecommended) prefer perl functions over
123 their coro replacements.
124
1253.11 Tue Dec 5 13:11:24 CET 2006
126 - fixed some bogus assert's, but as perl.h disables assert even
127 without NDEBUG (thank you very much), not too many people should
128 notice (that did include myself). Andreas König noticed, though :)
129 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
130 it. Also noticed by Andreas König :)
131 - save/restore tainted status.
132 - verified to pass the testsuite on my 5.9.5.
133
1343.1 Mon Dec 4 23:03:40 CET 2006
135 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
136 - incompatible change: transfer flags are now per-state.
137 - give a better error message on deadlock.
138 - document Coro::nready.
139 - enhanced testsuite.
140
1413.01 Sun Dec 3 23:47:42 CET 2006
142 - forgot to include Coro::Timer.
143
1443.0 Sun Dec 3 22:57:25 CET 2006
145 - the "FINALLY COMPLETELY STABLE" release (coming soon:
146 the "FAMOUS LAST WORDS" release).
147 - implement a new stack sharing algorithm, which uses a stack
148 pool (size currently hardcoded).
149 - make stack sharing mandatory (it no longer uses a heuristic).
150 - eval/die no longer cause weird problems under heavy use.
151 - Coro::Event could cause livelocks if it was used but
152 no Coro::Event watchers were used.
153 - Coro::Event now uses asynccheck as prepare does not
154 check for changed watchers.
155 - Coro::Event allows multiple waiting coros and will wake up one
156 per event.
157 - Coro::Event should be cleaner and more efficient now.
158 - new utility function Coro::unblock_sub.
159 - document the sad fact that Event is no longer reentrant.
160 - putting a coroutine into the ready queue twice could under
161 some circumstances lead to stack corruption.
162 - minor incompatibility: subclassing Coro::State is supported
163 directly now without going through a _coro_state member.
164 - state/coro switching is much faster now.
165 - very minor optimisations and code/documentation cleanup.
166 - avoid problems due to compiler inlining.
167 - removed timers from Coro::Timer -> use AnyEvent instead.
168 - replaced Coro::idle coroutine by (cleaner) idle handler.
169 - updated to newest libcoro.
170 - implement enhanced support for valgrind.
171 - implement is_ready and return value for ready.
172 - removed Coro::Cont, it was a misnomer (it's generators),
173 and it was somewhat annoying to get it right. Will come back
174 if somebody asks for it :)
175 - many ->wait methods and Coro::Event could return
176 spuriously without the event having happened.
177
1782.5 Tue Nov 7 12:22:33 CET 2006
179 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
180 use AnyEvent, moved them to Coro/.
181 - added Coro::LWP which contains all the uglyness required to
182 make LWP non-blocking.
183 - should work with perl 5.9.x now (Andreas König made me do it).
184 - fixed another bug in Coro::Select when the timeout was undef.
185 - reuse PL_start_env for all coros, saves some memory per coroutine.
186 - manage PL_top_env differently, hopefully to avoid panic: top_env.
187 - timeout argument was not properly used in Coro::Socket.
188 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
189 - emulate undocumented(!) functionality of IO::Socket required
190 by LWP(!!).
191 - updated eg/lwp to work with newer lwp's.
192 - remove "FATAL: uncaught exception" prefix. Coroutines that die
193 kill the whole process, just as exceptions in the main "coroutine"
194 did already.
195
1962.1 Wed Nov 1 23:01:13 CET 2006
197 - fix a long-standing bug in Coro::Select where select with
198 zero timeout would instead change the current default filehandle.
199 - use a simpler and hopefully more robust way to clone padlists
200 (uses less memory and a perl function instead of our own).
201 - coro can now create a stack guard area on many architectures.
202 - Coro::AIO properly reexports additional functions from IO::AIO.
203 - updated libcoro with a workaround for OS X,
204 pach and testing by Michael Schwern.
205
2062.0 Tue Oct 24 05:47:17 CEST 2006
207 - support additional aio requests in Coro::AIO.
208
2091.9 (never properly released due to a glitch)
210
2111.8 Thu Feb 2 00:59:06 CET 2006
212 - applied suggested patch by SAMV to avoid problems during stupid
213 mark & sweep gc run.
214 - applied patch by Scott Walters for 5.9.3 compatibility.
215
2161.7 Tue Dec 27 01:41:58 CET 2005
217 - added Coro::AIO, a thin wrapper around IO::AIO.
218 - improved Makefile.PL explanations.
219
2201.6 Mon Dec 12 21:30:05 CET 2005
221 - additionally save PM_curpm might fix as-of-yet
222 unseen problems with regex matches being attributed to the
223 wrong package.
224 - add t/10_bugs.t, which currently checks against imho broken
225 perls that use 0.26MB of stack space per Perl_magic_get
226 invocation (newer linuxes) as opposed to the 0.0002MB perl
227 normally uses.
228 - make stacksize configurable for the ultimate debian experience.
229
2301.51 Mon Dec 12 18:48:36 CET 2005
231 - remove debugging warn() accidentally left in Coro::Select.
232
2331.5 Tue Nov 29 13:32:44 CET 2005
234 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
235 - fix Coro::Select 'select'.
236 - strict'ify some modules.
237 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
238 - use XSLoader in selected modules.
239 - remove some 5.6 compatibility cruft.
240
2411.4 Tue Sep 6 00:11:05 CEST 2005
242 - libcoro did not take into account the trampoline on amd64,
243 when 'l'inux method was used.
244
2451.31 Tue Aug 30 23:31:33 CEST 2005
246 - some portability fixes/workarounds.
247
2481.3 Sat Aug 20 03:08:56 CEST 2005
249 - no code changes. module seems to work fine.
250 - improve Coro::State docs, remove reference to nonexisting
251 Coro::State::flush method.
252 - no longer autodetect windows, present it as an option instead.
253
2541.2 Mon May 16 02:00:55 CEST 2005
255 - included libcoro.c earlier in State.xs, might improve portability.
256 - use a faster and possibly more stable Coro::Cont implementation.
257 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
258 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
259 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
260
2611.11 Thu Mar 3 18:00:52 CET 2005
262 - change of contact address.
2 263
31.1 Tue Feb 22 20:51:16 CET 2005 2641.1 Tue Feb 22 20:51:16 CET 2005
4 - support [l]inux method on amd64. 265 - support [l]inux method on amd64.
5 - make some contetx switches while compiling work. 266 - allow some context switches while compiling/eval'ing.
6 267
71.0 Tue Aug 10 03:47:27 CEST 2004 2681.0 Tue Aug 10 03:47:27 CEST 2004
8 - Event 0.89 fixes the bug again. 269 - Event 0.89 fixes the bug again.
9 - fix the remaining(??) memleak. I didn't free the cache-AV for 270 - fix the remaining(??) memleak. I didn't free the cache-AV for
10 padlists, so one AV per created anon-sub leaked. 271 padlists, so one AV per created anon-sub leaked.
11 272
120.97 Fri May 14 15:23:32 CEST 2004 2730.97 Fri May 14 15:23:32 CEST 2004
13 - remove stupid gcc extension that was an experiment only anyway. 274 - remove stupid gcc extension that was an experiment only anyway.
14 275
150.96 Thu May 13 18:09:29 CEST 2004 2760.96 Thu May 13 18:09:29 CEST 2004
16 - tell people if they have yet anothe rbroken Event variant. 277 - tell people if they have yet another broken Event variant.
17 - made Coro to work _better_ with threads (it seems that, as long 278 - made Coro to work _better_ with threads (it seems that, as long
18 as you use Coros in one thread only everything will work fine. 279 as you use Coros in one thread only everything will work fine.
19 Looking at the perl source the problem might be missing locking 280 Looking at the perl source the problem might be missing locking
20 between Coro and some perl internal routines. That's not easily 281 between Coro and some perl internal routines. That's not easily
21 fixable). 282 fixable).

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines