| 1 | Revision history for Perl extension Coro. |
1 | Revision history for Perl extension Coro. |
| 2 | |
2 | |
|
|
3 | TODO: better (earlier) error messages on broken transfer's |
|
|
4 | TODO: destroy gets only called on cancel: fix(?) |
|
|
5 | |
|
|
6 | 3.501 Wed Feb 28 12:44:07 CET 2007 |
|
|
7 | - rename some global symbols as macosx from hell redefines |
|
|
8 | them without asking. |
|
|
9 | |
|
|
10 | 3.5 Tue Feb 13 20:22:53 CET 2007 |
|
|
11 | - do AnyEvent model detection earlier, avoiding problems |
|
|
12 | caused by first using AnyEvent and later Coro::Event. |
|
|
13 | - implement and document Coro::Event event objects. |
|
|
14 | - fix a potential problem in Coro::Event causing crashes. |
|
|
15 | - initialise PL_comppad when creating a new coroutine, |
|
|
16 | avoids crashes on early coro destruction. |
|
|
17 | |
|
|
18 | 3.41 Mon Jan 22 19:19:49 CET 2007 |
|
|
19 | - readline on Coro::Handle did not support undefined $/, |
|
|
20 | nor did it deliver partial lines on EOF or error. |
|
|
21 | - implement malloc fallback for stack allocation because |
|
|
22 | stupid broken idiotic OSX has a stupid broken |
|
|
23 | idiotic fits-the-whole-os mmap "implementation" and |
|
|
24 | my dick feels longer if Coro is portable even to |
|
|
25 | obsolete platforms. |
|
|
26 | |
|
|
27 | 3.4 Fri Jan 19 21:52:54 CET 2007 |
|
|
28 | - remove t/09_timer.t, as it isn't really testing much |
|
|
29 | but was rather flaky in practise. |
|
|
30 | - async_pool coro would keep arguments and callback alive until |
|
|
31 | it was reused. |
|
|
32 | - cancellation of a coroutine could cause spurious idle calls |
|
|
33 | in cede_notself. |
|
|
34 | |
|
|
35 | 3.3 Sat Jan 6 03:45:00 CET 2007 |
|
|
36 | - implement $coro->on_destroy. |
|
|
37 | - Coro::Event blocking semantics have been changed, |
|
|
38 | documented and - hopefully - improved. |
|
|
39 | - fix nice adding, not subtracting, from priority. |
|
|
40 | - fix ->prio and api_is_ready (patch by Mark Hinds). |
|
|
41 | - fixed an assert ("... == prev__cctx->idle_te") |
|
|
42 | that could errronously trigger. |
|
|
43 | - fix various large and small memleaks. |
|
|
44 | - use a (hopefully) more stable cancel implementation |
|
|
45 | that immediately frees the coroutine data. |
|
|
46 | - cede/cede_notself return a status now. |
|
|
47 | - added Coro::guard function. |
|
|
48 | - added a global coroutine pool for jobs (on my machine, |
|
|
49 | I can create and execute 48k simple coros/s with async, |
|
|
50 | and 128k coros with async_pool). |
|
|
51 | - Coro::AIO now uses the coroutine priority as io priority. |
|
|
52 | |
|
|
53 | 3.2 Fri Dec 22 05:07:09 CET 2006 |
|
|
54 | - improve portability to slightly older perls. |
|
|
55 | - use cleaner coroutine destruction. |
|
|
56 | - simplify configuration for users. |
|
|
57 | - optionally (unrecommended) prefer perl functions over |
|
|
58 | their coro replacements. |
|
|
59 | |
|
|
60 | 3.11 Tue Dec 5 13:11:24 CET 2006 |
|
|
61 | - fixed some bogus assert's, but as perl.h disables assert even |
|
|
62 | without NDEBUG (thank you very much), not too many people should |
|
|
63 | notice (that did include myself). Andreas König noticed, though :) |
|
|
64 | - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have |
|
|
65 | it. Also noticed by Andreas König :) |
|
|
66 | - save/restore tainted status. |
|
|
67 | - verified to pass the testsuite on my 5.9.5. |
|
|
68 | |
|
|
69 | 3.1 Mon Dec 4 23:03:40 CET 2006 |
|
|
70 | - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow). |
|
|
71 | - incompatible change: transfer flags are now per-state. |
|
|
72 | - give a better error message on deadlock. |
|
|
73 | - document Coro::nready. |
|
|
74 | - enhanced testsuite. |
|
|
75 | |
|
|
76 | 3.01 Sun Dec 3 23:47:42 CET 2006 |
|
|
77 | - forgot to include Coro::Timer. |
|
|
78 | |
|
|
79 | 3.0 Sun Dec 3 22:57:25 CET 2006 |
|
|
80 | - the "FINALLY COMPLETELY STABLE" release (coming soon: |
|
|
81 | the "FAMOUS LAST WORDS" release). |
|
|
82 | - implement a new stack sharing algorithm, which uses a stack |
|
|
83 | pool (size currently hardcoded). |
|
|
84 | - make stack sharing mandatory (it no longer uses a heuristic). |
|
|
85 | - eval/die no longer cause weird problems under heavy use. |
|
|
86 | - Coro::Event could cause livelocks if it was used but |
|
|
87 | no Coro::Event watchers were used. |
|
|
88 | - Coro::Event now uses asynccheck as prepare does not |
|
|
89 | check for changed watchers. |
|
|
90 | - Coro::Event allows multiple waiting coros and will wake up one |
|
|
91 | per event. |
|
|
92 | - Coro::Event should be cleaner and more efficient now. |
|
|
93 | - new utility function Coro::unblock_sub. |
|
|
94 | - document the sad fact that Event is no longer reentrant. |
|
|
95 | - putting a coroutine into the ready queue twice could under |
|
|
96 | some circumstances lead to stack corruption. |
|
|
97 | - minor incompatibility: subclassing Coro::State is supported |
|
|
98 | directly now without going through a _coro_state member. |
|
|
99 | - state/coro switching is much faster now. |
|
|
100 | - very minor optimisations and code/documentation cleanup. |
|
|
101 | - avoid problems due to compiler inlining. |
|
|
102 | - removed timers from Coro::Timer -> use AnyEvent instead. |
|
|
103 | - replaced Coro::idle coroutine by (cleaner) idle handler. |
|
|
104 | - updated to newest libcoro. |
|
|
105 | - implement enhanced support for valgrind. |
|
|
106 | - implement is_ready and return value for ready. |
|
|
107 | - removed Coro::Cont, it was a misnomer (it's generators), |
|
|
108 | and it was somewhat annoying to get it right. Will come back |
|
|
109 | if somebody asks for it :) |
|
|
110 | - many ->wait methods and Coro::Event could return |
|
|
111 | spuriously without the event having happened. |
|
|
112 | |
|
|
113 | 2.5 Tue Nov 7 12:22:33 CET 2006 |
|
|
114 | - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket |
|
|
115 | use AnyEvent, moved them to Coro/. |
|
|
116 | - added Coro::LWP which contains all the uglyness required to |
|
|
117 | make LWP non-blocking. |
|
|
118 | - should work with perl 5.9.x now (Andreas König made me do it). |
|
|
119 | - fixed another bug in Coro::Select when the timeout was undef. |
|
|
120 | - reuse PL_start_env for all coros, saves some memory per coroutine. |
|
|
121 | - manage PL_top_env differently, hopefully to avoid panic: top_env. |
|
|
122 | - timeout argument was not properly used in Coro::Socket. |
|
|
123 | - allow limited forms of subclassing in Coro::Handle/Coro::Socket. |
|
|
124 | - emulate undocumented(!) functionality of IO::Socket required |
|
|
125 | by LWP(!!). |
|
|
126 | - updated eg/lwp to work with newer lwp's. |
|
|
127 | - remove "FATAL: uncaught exception" prefix. Coroutines that die |
|
|
128 | kill the whole process, just as exceptions in the main "coroutine" |
|
|
129 | did already. |
|
|
130 | |
|
|
131 | 2.1 Wed Nov 1 23:01:13 CET 2006 |
|
|
132 | - fix a long-standing bug in Coro::Select where select with |
|
|
133 | zero timeout would instead change the current default filehandle. |
|
|
134 | - use a simpler and hopefully more robust way to clone padlists |
|
|
135 | (uses less memory and a perl function instead of our own). |
|
|
136 | - coro can now create a stack guard area on many architectures. |
|
|
137 | - Coro::AIO properly reexports additional functions from IO::AIO. |
|
|
138 | - updated libcoro with a workaround for OS X, |
|
|
139 | pach and testing by Michael Schwern. |
|
|
140 | |
|
|
141 | 2.0 Tue Oct 24 05:47:17 CEST 2006 |
|
|
142 | - support additional aio requests in Coro::AIO. |
|
|
143 | |
|
|
144 | 1.9 (never properly released due to a glitch) |
|
|
145 | |
|
|
146 | 1.8 Thu Feb 2 00:59:06 CET 2006 |
|
|
147 | - applied suggested patch by SAMV to avoid problems during stupid |
|
|
148 | mark & sweep gc run. |
|
|
149 | - applied patch by Scott Walters for 5.9.3 compatibility. |
|
|
150 | |
|
|
151 | 1.7 Tue Dec 27 01:41:58 CET 2005 |
|
|
152 | - added Coro::AIO, a thin wrapper around IO::AIO. |
|
|
153 | - improved Makefile.PL explanations. |
|
|
154 | |
|
|
155 | 1.6 Mon Dec 12 21:30:05 CET 2005 |
|
|
156 | - additionally save PM_curpm might fix as-of-yet |
|
|
157 | unseen problems with regex matches being attributed to the |
|
|
158 | wrong package. |
|
|
159 | - add t/10_bugs.t, which currently checks against imho broken |
|
|
160 | perls that use 0.26MB of stack space per Perl_magic_get |
|
|
161 | invocation (newer linuxes) as opposed to the 0.0002MB perl |
|
|
162 | normally uses. |
|
|
163 | - make stacksize configurable for the ultimate debian experience. |
|
|
164 | |
|
|
165 | 1.51 Mon Dec 12 18:48:36 CET 2005 |
|
|
166 | - remove debugging warn() accidentally left in Coro::Select. |
|
|
167 | |
|
|
168 | 1.5 Tue Nov 29 13:32:44 CET 2005 |
|
|
169 | - use Coro::Event inside Coro::Select to avoid spurious deadlocks. |
|
|
170 | - fix Coro::Select 'select'. |
|
|
171 | - strict'ify some modules. |
|
|
172 | - libcoro errornously restored the SIGUSR2 handler to SIGUSR1. |
|
|
173 | - use XSLoader in selected modules. |
|
|
174 | - remove some 5.6 compatibility cruft. |
|
|
175 | |
|
|
176 | 1.4 Tue Sep 6 00:11:05 CEST 2005 |
|
|
177 | - libcoro did not take into account the trampoline on amd64, |
|
|
178 | when 'l'inux method was used. |
|
|
179 | |
|
|
180 | 1.31 Tue Aug 30 23:31:33 CEST 2005 |
|
|
181 | - some portability fixes/workarounds. |
|
|
182 | |
|
|
183 | 1.3 Sat Aug 20 03:08:56 CEST 2005 |
|
|
184 | - no code changes. module seems to work fine. |
|
|
185 | - improve Coro::State docs, remove reference to nonexisting |
|
|
186 | Coro::State::flush method. |
|
|
187 | - no longer autodetect windows, present it as an option instead. |
|
|
188 | |
|
|
189 | 1.2 Mon May 16 02:00:55 CEST 2005 |
|
|
190 | - included libcoro.c earlier in State.xs, might improve portability. |
|
|
191 | - use a faster and possibly more stable Coro::Cont implementation. |
|
|
192 | - accept x86_64 in addition to amd64 for optimized linux-amd64 support. |
|
|
193 | - fix bugs Coro::SemaphoreSet that could cause locks never to be freed. |
|
|
194 | - fix bugs in CoroAPI.h and document it in Coro::MakeMaker. |
|
|
195 | |
|
|
196 | 1.11 Thu Mar 3 18:00:52 CET 2005 |
|
|
197 | - change of contact address. |
|
|
198 | |
|
|
199 | 1.1 Tue Feb 22 20:51:16 CET 2005 |
|
|
200 | - support [l]inux method on amd64. |
|
|
201 | - allow some context switches while compiling/eval'ing. |
|
|
202 | |
|
|
203 | 1.0 Tue Aug 10 03:47:27 CEST 2004 |
|
|
204 | - Event 0.89 fixes the bug again. |
|
|
205 | - fix the remaining(??) memleak. I didn't free the cache-AV for |
|
|
206 | padlists, so one AV per created anon-sub leaked. |
|
|
207 | |
|
|
208 | 0.97 Fri May 14 15:23:32 CEST 2004 |
|
|
209 | - remove stupid gcc extension that was an experiment only anyway. |
|
|
210 | |
| 3 | 0.96 Thu May 13 18:09:29 CEST 2004 |
211 | 0.96 Thu May 13 18:09:29 CEST 2004 |
| 4 | - tell people if they have yet anothe rbroken Event variant. |
212 | - tell people if they have yet another broken Event variant. |
| 5 | - made Coro to work _better_ with threads (it seems that, as long |
213 | - made Coro to work _better_ with threads (it seems that, as long |
| 6 | as you use Coros in one thread only everything will work fine. |
214 | as you use Coros in one thread only everything will work fine. |
| 7 | Looking at the perl source the problem might be missing locking |
215 | Looking at the perl source the problem might be missing locking |
| 8 | between Coro and some perl internal routines. That's not easily |
216 | between Coro and some perl internal routines. That's not easily |
| 9 | fixable). |
217 | fixable). |
