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

Comparing Coro/Changes (file contents):
Revision 1.94 by pcg, Fri May 14 13:25:08 2004 UTC vs.
Revision 1.221 by root, Sat Apr 14 15:08:15 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
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: eg/event
8 - added some bugfixes to get eg/myhttpd working again.
9 - added Coro::Storable for often-cede'ing freeze/thaw.
10 - try to do a clean exit when a coroutine calls exit
11 (EXPERIMENTAL).
12 - got rid of indirect call through _coro_init.
13 - updated the partly antique examples in eg/ to
14 work again and be a bit less magic, too.
15 - fixed Coro::Signal semantics to work as documented again.
16
173.55 Sun Mar 25 01:20:47 CET 2007
18 - add SAVE_DEFFH to save the default filehandle and enable
19 it by default.
20 - finally move socket-operations from Coro::Socket to Coro::Handle
21 to be able to unblock foreign sockets.
22 - add Coro::State::save_also and guarded_save.
23 - add count accessor to Coro::Semaphore.
24 - add Coro::State::cctx_stacksize.
25 - just for the fun of it, do not rely on implicit context,
26 which can dramatically improve performance, but people
27 using windows-process-emulation perls do not care much about
28 performance.
29
303.51 Sun Mar 4 14:18:04 CET 2007
31 - fixed a problem when you weakened references to Coro::State's
32 (patch by Yuval Kogman).
33
343.501 Wed Feb 28 12:44:07 CET 2007
35 - rename some global symbols as macosx from hell redefines
36 them without asking.
37
383.5 Tue Feb 13 20:22:53 CET 2007
39 - do AnyEvent model detection earlier, avoiding problems
40 caused by first using AnyEvent and later Coro::Event.
41 - implement and document Coro::Event event objects.
42 - fix a potential problem in Coro::Event causing crashes.
43 - initialise PL_comppad when creating a new coroutine,
44 avoids crashes on early coro destruction.
45
463.41 Mon Jan 22 19:19:49 CET 2007
47 - readline on Coro::Handle did not support undefined $/,
48 nor did it deliver partial lines on EOF or error.
49 - implement malloc fallback for stack allocation because
50 stupid broken idiotic OSX has a stupid broken
51 idiotic fits-the-whole-os mmap "implementation" and
52 my dick feels longer if Coro is portable even to
53 obsolete platforms.
54
553.4 Fri Jan 19 21:52:54 CET 2007
56 - remove t/09_timer.t, as it isn't really testing much
57 but was rather flaky in practise.
58 - async_pool coro would keep arguments and callback alive until
59 it was reused.
60 - cancellation of a coroutine could cause spurious idle calls
61 in cede_notself.
62
633.3 Sat Jan 6 03:45:00 CET 2007
64 - implement $coro->on_destroy.
65 - Coro::Event blocking semantics have been changed,
66 documented and - hopefully - improved.
67 - fix nice adding, not subtracting, from priority.
68 - fix ->prio and api_is_ready (patch by Mark Hinds).
69 - fixed an assert ("... == prev__cctx->idle_te")
70 that could errronously trigger.
71 - fix various large and small memleaks.
72 - use a (hopefully) more stable cancel implementation
73 that immediately frees the coroutine data.
74 - cede/cede_notself return a status now.
75 - added Coro::guard function.
76 - added a global coroutine pool for jobs (on my machine,
77 I can create and execute 48k simple coros/s with async,
78 and 128k coros with async_pool).
79 - Coro::AIO now uses the coroutine priority as io priority.
80
813.2 Fri Dec 22 05:07:09 CET 2006
82 - improve portability to slightly older perls.
83 - use cleaner coroutine destruction.
84 - simplify configuration for users.
85 - optionally (unrecommended) prefer perl functions over
86 their coro replacements.
87
883.11 Tue Dec 5 13:11:24 CET 2006
89 - fixed some bogus assert's, but as perl.h disables assert even
90 without NDEBUG (thank you very much), not too many people should
91 notice (that did include myself). Andreas König noticed, though :)
92 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
93 it. Also noticed by Andreas König :)
94 - save/restore tainted status.
95 - verified to pass the testsuite on my 5.9.5.
96
973.1 Mon Dec 4 23:03:40 CET 2006
98 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
99 - incompatible change: transfer flags are now per-state.
100 - give a better error message on deadlock.
101 - document Coro::nready.
102 - enhanced testsuite.
103
1043.01 Sun Dec 3 23:47:42 CET 2006
105 - forgot to include Coro::Timer.
106
1073.0 Sun Dec 3 22:57:25 CET 2006
108 - the "FINALLY COMPLETELY STABLE" release (coming soon:
109 the "FAMOUS LAST WORDS" release).
110 - implement a new stack sharing algorithm, which uses a stack
111 pool (size currently hardcoded).
112 - make stack sharing mandatory (it no longer uses a heuristic).
113 - eval/die no longer cause weird problems under heavy use.
114 - Coro::Event could cause livelocks if it was used but
115 no Coro::Event watchers were used.
116 - Coro::Event now uses asynccheck as prepare does not
117 check for changed watchers.
118 - Coro::Event allows multiple waiting coros and will wake up one
119 per event.
120 - Coro::Event should be cleaner and more efficient now.
121 - new utility function Coro::unblock_sub.
122 - document the sad fact that Event is no longer reentrant.
123 - putting a coroutine into the ready queue twice could under
124 some circumstances lead to stack corruption.
125 - minor incompatibility: subclassing Coro::State is supported
126 directly now without going through a _coro_state member.
127 - state/coro switching is much faster now.
128 - very minor optimisations and code/documentation cleanup.
129 - avoid problems due to compiler inlining.
130 - removed timers from Coro::Timer -> use AnyEvent instead.
131 - replaced Coro::idle coroutine by (cleaner) idle handler.
132 - updated to newest libcoro.
133 - implement enhanced support for valgrind.
134 - implement is_ready and return value for ready.
135 - removed Coro::Cont, it was a misnomer (it's generators),
136 and it was somewhat annoying to get it right. Will come back
137 if somebody asks for it :)
138 - many ->wait methods and Coro::Event could return
139 spuriously without the event having happened.
140
1412.5 Tue Nov 7 12:22:33 CET 2006
142 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
143 use AnyEvent, moved them to Coro/.
144 - added Coro::LWP which contains all the uglyness required to
145 make LWP non-blocking.
146 - should work with perl 5.9.x now (Andreas König made me do it).
147 - fixed another bug in Coro::Select when the timeout was undef.
148 - reuse PL_start_env for all coros, saves some memory per coroutine.
149 - manage PL_top_env differently, hopefully to avoid panic: top_env.
150 - timeout argument was not properly used in Coro::Socket.
151 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
152 - emulate undocumented(!) functionality of IO::Socket required
153 by LWP(!!).
154 - updated eg/lwp to work with newer lwp's.
155 - remove "FATAL: uncaught exception" prefix. Coroutines that die
156 kill the whole process, just as exceptions in the main "coroutine"
157 did already.
158
1592.1 Wed Nov 1 23:01:13 CET 2006
160 - fix a long-standing bug in Coro::Select where select with
161 zero timeout would instead change the current default filehandle.
162 - use a simpler and hopefully more robust way to clone padlists
163 (uses less memory and a perl function instead of our own).
164 - coro can now create a stack guard area on many architectures.
165 - Coro::AIO properly reexports additional functions from IO::AIO.
166 - updated libcoro with a workaround for OS X,
167 pach and testing by Michael Schwern.
168
1692.0 Tue Oct 24 05:47:17 CEST 2006
170 - support additional aio requests in Coro::AIO.
171
1721.9 (never properly released due to a glitch)
173
1741.8 Thu Feb 2 00:59:06 CET 2006
175 - applied suggested patch by SAMV to avoid problems during stupid
176 mark & sweep gc run.
177 - applied patch by Scott Walters for 5.9.3 compatibility.
178
1791.7 Tue Dec 27 01:41:58 CET 2005
180 - added Coro::AIO, a thin wrapper around IO::AIO.
181 - improved Makefile.PL explanations.
182
1831.6 Mon Dec 12 21:30:05 CET 2005
184 - additionally save PM_curpm might fix as-of-yet
185 unseen problems with regex matches being attributed to the
186 wrong package.
187 - add t/10_bugs.t, which currently checks against imho broken
188 perls that use 0.26MB of stack space per Perl_magic_get
189 invocation (newer linuxes) as opposed to the 0.0002MB perl
190 normally uses.
191 - make stacksize configurable for the ultimate debian experience.
192
1931.51 Mon Dec 12 18:48:36 CET 2005
194 - remove debugging warn() accidentally left in Coro::Select.
195
1961.5 Tue Nov 29 13:32:44 CET 2005
197 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
198 - fix Coro::Select 'select'.
199 - strict'ify some modules.
200 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
201 - use XSLoader in selected modules.
202 - remove some 5.6 compatibility cruft.
203
2041.4 Tue Sep 6 00:11:05 CEST 2005
205 - libcoro did not take into account the trampoline on amd64,
206 when 'l'inux method was used.
207
2081.31 Tue Aug 30 23:31:33 CEST 2005
209 - some portability fixes/workarounds.
210
2111.3 Sat Aug 20 03:08:56 CEST 2005
212 - no code changes. module seems to work fine.
213 - improve Coro::State docs, remove reference to nonexisting
214 Coro::State::flush method.
215 - no longer autodetect windows, present it as an option instead.
216
2171.2 Mon May 16 02:00:55 CEST 2005
218 - included libcoro.c earlier in State.xs, might improve portability.
219 - use a faster and possibly more stable Coro::Cont implementation.
220 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
221 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
222 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
223
2241.11 Thu Mar 3 18:00:52 CET 2005
225 - change of contact address.
226
2271.1 Tue Feb 22 20:51:16 CET 2005
228 - support [l]inux method on amd64.
229 - allow some context switches while compiling/eval'ing.
230
2311.0 Tue Aug 10 03:47:27 CEST 2004
232 - Event 0.89 fixes the bug again.
233 - fix the remaining(??) memleak. I didn't free the cache-AV for
234 padlists, so one AV per created anon-sub leaked.
2 235
30.97 Fri May 14 15:23:32 CEST 2004 2360.97 Fri May 14 15:23:32 CEST 2004
4 - remove stupid gcc extension that was an experiment only anyway. 237 - remove stupid gcc extension that was an experiment only anyway.
5 238
60.96 Thu May 13 18:09:29 CEST 2004 2390.96 Thu May 13 18:09:29 CEST 2004
7 - tell people if they have yet anothe rbroken Event variant. 240 - tell people if they have yet another broken Event variant.
8 - made Coro to work _better_ with threads (it seems that, as long 241 - made Coro to work _better_ with threads (it seems that, as long
9 as you use Coros in one thread only everything will work fine. 242 as you use Coros in one thread only everything will work fine.
10 Looking at the perl source the problem might be missing locking 243 Looking at the perl source the problem might be missing locking
11 between Coro and some perl internal routines. That's not easily 244 between Coro and some perl internal routines. That's not easily
12 fixable). 245 fixable).

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines