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

Comparing Coro/Changes (file contents):
Revision 1.102 by root, Mon Mar 21 17:33:49 2005 UTC vs.
Revision 1.223 by root, Mon Apr 16 13:26:42 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
3TODO: better (earlier) error messages on broken transfer's
4TODO: destroy gets only called on cancel: fix(?)
5TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO.
6
7TODO: fix Coro::Signal once and for all
8
93.61
10 - use a Semaphore in Coro::Storable, as Storable doesn't
11 seem to be reentrant (although it should be...).
12
133.6 Sat Apr 14 17:13:31 CEST 2007
14 - added some bugfixes to get eg/myhttpd working again.
15 - added Coro::Storable for often-cede'ing freeze/thaw.
16 - try to do a clean exit when a coroutine calls exit
17 (EXPERIMENTAL).
18 - got rid of indirect call through _coro_init.
19 - updated the partly antique examples in eg/ to
20 work again and be a bit less magic, too.
21 - fixed Coro::Signal semantics to work as documented again.
22
233.55 Sun Mar 25 01:20:47 CET 2007
24 - add SAVE_DEFFH to save the default filehandle and enable
25 it by default.
26 - finally move socket-operations from Coro::Socket to Coro::Handle
27 to be able to unblock foreign sockets.
28 - add Coro::State::save_also and guarded_save.
29 - add count accessor to Coro::Semaphore.
30 - add Coro::State::cctx_stacksize.
31 - just for the fun of it, do not rely on implicit context,
32 which can dramatically improve performance, but people
33 using windows-process-emulation perls do not care much about
34 performance.
35
363.51 Sun Mar 4 14:18:04 CET 2007
37 - fixed a problem when you weakened references to Coro::State's
38 (patch by Yuval Kogman).
39
403.501 Wed Feb 28 12:44:07 CET 2007
41 - rename some global symbols as macosx from hell redefines
42 them without asking.
43
443.5 Tue Feb 13 20:22:53 CET 2007
45 - do AnyEvent model detection earlier, avoiding problems
46 caused by first using AnyEvent and later Coro::Event.
47 - implement and document Coro::Event event objects.
48 - fix a potential problem in Coro::Event causing crashes.
49 - initialise PL_comppad when creating a new coroutine,
50 avoids crashes on early coro destruction.
51
523.41 Mon Jan 22 19:19:49 CET 2007
53 - readline on Coro::Handle did not support undefined $/,
54 nor did it deliver partial lines on EOF or error.
55 - implement malloc fallback for stack allocation because
56 stupid broken idiotic OSX has a stupid broken
57 idiotic fits-the-whole-os mmap "implementation" and
58 my dick feels longer if Coro is portable even to
59 obsolete platforms.
60
613.4 Fri Jan 19 21:52:54 CET 2007
62 - remove t/09_timer.t, as it isn't really testing much
63 but was rather flaky in practise.
64 - async_pool coro would keep arguments and callback alive until
65 it was reused.
66 - cancellation of a coroutine could cause spurious idle calls
67 in cede_notself.
68
693.3 Sat Jan 6 03:45:00 CET 2007
70 - implement $coro->on_destroy.
71 - Coro::Event blocking semantics have been changed,
72 documented and - hopefully - improved.
73 - fix nice adding, not subtracting, from priority.
74 - fix ->prio and api_is_ready (patch by Mark Hinds).
75 - fixed an assert ("... == prev__cctx->idle_te")
76 that could errronously trigger.
77 - fix various large and small memleaks.
78 - use a (hopefully) more stable cancel implementation
79 that immediately frees the coroutine data.
80 - cede/cede_notself return a status now.
81 - added Coro::guard function.
82 - added a global coroutine pool for jobs (on my machine,
83 I can create and execute 48k simple coros/s with async,
84 and 128k coros with async_pool).
85 - Coro::AIO now uses the coroutine priority as io priority.
86
873.2 Fri Dec 22 05:07:09 CET 2006
88 - improve portability to slightly older perls.
89 - use cleaner coroutine destruction.
90 - simplify configuration for users.
91 - optionally (unrecommended) prefer perl functions over
92 their coro replacements.
93
943.11 Tue Dec 5 13:11:24 CET 2006
95 - fixed some bogus assert's, but as perl.h disables assert even
96 without NDEBUG (thank you very much), not too many people should
97 notice (that did include myself). Andreas König noticed, though :)
98 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
99 it. Also noticed by Andreas König :)
100 - save/restore tainted status.
101 - verified to pass the testsuite on my 5.9.5.
102
1033.1 Mon Dec 4 23:03:40 CET 2006
104 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
105 - incompatible change: transfer flags are now per-state.
106 - give a better error message on deadlock.
107 - document Coro::nready.
108 - enhanced testsuite.
109
1103.01 Sun Dec 3 23:47:42 CET 2006
111 - forgot to include Coro::Timer.
112
1133.0 Sun Dec 3 22:57:25 CET 2006
114 - the "FINALLY COMPLETELY STABLE" release (coming soon:
115 the "FAMOUS LAST WORDS" release).
116 - implement a new stack sharing algorithm, which uses a stack
117 pool (size currently hardcoded).
118 - make stack sharing mandatory (it no longer uses a heuristic).
119 - eval/die no longer cause weird problems under heavy use.
120 - Coro::Event could cause livelocks if it was used but
121 no Coro::Event watchers were used.
122 - Coro::Event now uses asynccheck as prepare does not
123 check for changed watchers.
124 - Coro::Event allows multiple waiting coros and will wake up one
125 per event.
126 - Coro::Event should be cleaner and more efficient now.
127 - new utility function Coro::unblock_sub.
128 - document the sad fact that Event is no longer reentrant.
129 - putting a coroutine into the ready queue twice could under
130 some circumstances lead to stack corruption.
131 - minor incompatibility: subclassing Coro::State is supported
132 directly now without going through a _coro_state member.
133 - state/coro switching is much faster now.
134 - very minor optimisations and code/documentation cleanup.
135 - avoid problems due to compiler inlining.
136 - removed timers from Coro::Timer -> use AnyEvent instead.
137 - replaced Coro::idle coroutine by (cleaner) idle handler.
138 - updated to newest libcoro.
139 - implement enhanced support for valgrind.
140 - implement is_ready and return value for ready.
141 - removed Coro::Cont, it was a misnomer (it's generators),
142 and it was somewhat annoying to get it right. Will come back
143 if somebody asks for it :)
144 - many ->wait methods and Coro::Event could return
145 spuriously without the event having happened.
146
1472.5 Tue Nov 7 12:22:33 CET 2006
148 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
149 use AnyEvent, moved them to Coro/.
150 - added Coro::LWP which contains all the uglyness required to
151 make LWP non-blocking.
152 - should work with perl 5.9.x now (Andreas König made me do it).
153 - fixed another bug in Coro::Select when the timeout was undef.
154 - reuse PL_start_env for all coros, saves some memory per coroutine.
155 - manage PL_top_env differently, hopefully to avoid panic: top_env.
156 - timeout argument was not properly used in Coro::Socket.
157 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
158 - emulate undocumented(!) functionality of IO::Socket required
159 by LWP(!!).
160 - updated eg/lwp to work with newer lwp's.
161 - remove "FATAL: uncaught exception" prefix. Coroutines that die
162 kill the whole process, just as exceptions in the main "coroutine"
163 did already.
164
1652.1 Wed Nov 1 23:01:13 CET 2006
166 - fix a long-standing bug in Coro::Select where select with
167 zero timeout would instead change the current default filehandle.
168 - use a simpler and hopefully more robust way to clone padlists
169 (uses less memory and a perl function instead of our own).
170 - coro can now create a stack guard area on many architectures.
171 - Coro::AIO properly reexports additional functions from IO::AIO.
172 - updated libcoro with a workaround for OS X,
173 pach and testing by Michael Schwern.
174
1752.0 Tue Oct 24 05:47:17 CEST 2006
176 - support additional aio requests in Coro::AIO.
177
1781.9 (never properly released due to a glitch)
179
1801.8 Thu Feb 2 00:59:06 CET 2006
181 - applied suggested patch by SAMV to avoid problems during stupid
182 mark & sweep gc run.
183 - applied patch by Scott Walters for 5.9.3 compatibility.
184
1851.7 Tue Dec 27 01:41:58 CET 2005
186 - added Coro::AIO, a thin wrapper around IO::AIO.
187 - improved Makefile.PL explanations.
188
1891.6 Mon Dec 12 21:30:05 CET 2005
190 - additionally save PM_curpm might fix as-of-yet
191 unseen problems with regex matches being attributed to the
192 wrong package.
193 - add t/10_bugs.t, which currently checks against imho broken
194 perls that use 0.26MB of stack space per Perl_magic_get
195 invocation (newer linuxes) as opposed to the 0.0002MB perl
196 normally uses.
197 - make stacksize configurable for the ultimate debian experience.
198
1991.51 Mon Dec 12 18:48:36 CET 2005
200 - remove debugging warn() accidentally left in Coro::Select.
201
2021.5 Tue Nov 29 13:32:44 CET 2005
203 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
204 - fix Coro::Select 'select'.
205 - strict'ify some modules.
206 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
207 - use XSLoader in selected modules.
208 - remove some 5.6 compatibility cruft.
209
2101.4 Tue Sep 6 00:11:05 CEST 2005
211 - libcoro did not take into account the trampoline on amd64,
212 when 'l'inux method was used.
213
2141.31 Tue Aug 30 23:31:33 CEST 2005
215 - some portability fixes/workarounds.
216
2171.3 Sat Aug 20 03:08:56 CEST 2005
218 - no code changes. module seems to work fine.
219 - improve Coro::State docs, remove reference to nonexisting
220 Coro::State::flush method.
221 - no longer autodetect windows, present it as an option instead.
222
2231.2 Mon May 16 02:00:55 CEST 2005
3 - included libcoro.c earlier in State.xs, might improve portability. 224 - included libcoro.c earlier in State.xs, might improve portability.
4 - use a faster and possibly more stable Coro::Cont implementation. 225 - use a faster and possibly more stable Coro::Cont implementation.
5 - accept x86_64 in addition to amd64 for optimized linux-amd64 support. 226 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
227 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
228 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
6 229
71.11 Thu Mar 3 18:00:52 CET 2005 2301.11 Thu Mar 3 18:00:52 CET 2005
8 - change of contact address. 231 - change of contact address.
9 232
101.1 Tue Feb 22 20:51:16 CET 2005 2331.1 Tue Feb 22 20:51:16 CET 2005
11 - support [l]inux method on amd64. 234 - support [l]inux method on amd64.
12 - make some context switches while compiling/eval'ing work. 235 - allow some context switches while compiling/eval'ing.
13 236
141.0 Tue Aug 10 03:47:27 CEST 2004 2371.0 Tue Aug 10 03:47:27 CEST 2004
15 - Event 0.89 fixes the bug again. 238 - Event 0.89 fixes the bug again.
16 - fix the remaining(??) memleak. I didn't free the cache-AV for 239 - fix the remaining(??) memleak. I didn't free the cache-AV for
17 padlists, so one AV per created anon-sub leaked. 240 padlists, so one AV per created anon-sub leaked.
18 241
190.97 Fri May 14 15:23:32 CEST 2004 2420.97 Fri May 14 15:23:32 CEST 2004
20 - remove stupid gcc extension that was an experiment only anyway. 243 - remove stupid gcc extension that was an experiment only anyway.
21 244
220.96 Thu May 13 18:09:29 CEST 2004 2450.96 Thu May 13 18:09:29 CEST 2004
23 - tell people if they have yet anothe rbroken Event variant. 246 - tell people if they have yet another broken Event variant.
24 - made Coro to work _better_ with threads (it seems that, as long 247 - made Coro to work _better_ with threads (it seems that, as long
25 as you use Coros in one thread only everything will work fine. 248 as you use Coros in one thread only everything will work fine.
26 Looking at the perl source the problem might be missing locking 249 Looking at the perl source the problem might be missing locking
27 between Coro and some perl internal routines. That's not easily 250 between Coro and some perl internal routines. That's not easily
28 fixable). 251 fixable).

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines