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

Comparing Coro/Changes (file contents):
Revision 1.44 by root, Mon Sep 3 02:50:18 2001 UTC vs.
Revision 1.212 by root, Mon Mar 12 21:40:14 2007 UTC

1Revision history for Perl extension Coro. 1Revision history for Perl extension Coro.
2 2
30.49 3TODO: better (earlier) error messages on broken transfer's
4TODO: destroy gets only called on cancel: fix(?)
5
63.52
7 - add count accessor to Coro::Semaphore.
8 - add Coro::State::cctx_stacksize.
9
103.51 Sun Mar 4 14:18:04 CET 2007
11 - fixed a problem when you weakened references to Coro::State's
12 (patch by Yuval Kogman).
13
143.501 Wed Feb 28 12:44:07 CET 2007
15 - rename some global symbols as macosx from hell redefines
16 them without asking.
17
183.5 Tue Feb 13 20:22:53 CET 2007
19 - do AnyEvent model detection earlier, avoiding problems
20 caused by first using AnyEvent and later Coro::Event.
21 - implement and document Coro::Event event objects.
22 - fix a potential problem in Coro::Event causing crashes.
23 - initialise PL_comppad when creating a new coroutine,
24 avoids crashes on early coro destruction.
25
263.41 Mon Jan 22 19:19:49 CET 2007
27 - readline on Coro::Handle did not support undefined $/,
28 nor did it deliver partial lines on EOF or error.
29 - implement malloc fallback for stack allocation because
30 stupid broken idiotic OSX has a stupid broken
31 idiotic fits-the-whole-os mmap "implementation" and
32 my dick feels longer if Coro is portable even to
33 obsolete platforms.
34
353.4 Fri Jan 19 21:52:54 CET 2007
36 - remove t/09_timer.t, as it isn't really testing much
37 but was rather flaky in practise.
38 - async_pool coro would keep arguments and callback alive until
39 it was reused.
40 - cancellation of a coroutine could cause spurious idle calls
41 in cede_notself.
42
433.3 Sat Jan 6 03:45:00 CET 2007
44 - implement $coro->on_destroy.
45 - Coro::Event blocking semantics have been changed,
46 documented and - hopefully - improved.
47 - fix nice adding, not subtracting, from priority.
48 - fix ->prio and api_is_ready (patch by Mark Hinds).
49 - fixed an assert ("... == prev__cctx->idle_te")
50 that could errronously trigger.
51 - fix various large and small memleaks.
52 - use a (hopefully) more stable cancel implementation
53 that immediately frees the coroutine data.
54 - cede/cede_notself return a status now.
55 - added Coro::guard function.
56 - added a global coroutine pool for jobs (on my machine,
57 I can create and execute 48k simple coros/s with async,
58 and 128k coros with async_pool).
59 - Coro::AIO now uses the coroutine priority as io priority.
60
613.2 Fri Dec 22 05:07:09 CET 2006
62 - improve portability to slightly older perls.
63 - use cleaner coroutine destruction.
64 - simplify configuration for users.
65 - optionally (unrecommended) prefer perl functions over
66 their coro replacements.
67
683.11 Tue Dec 5 13:11:24 CET 2006
69 - fixed some bogus assert's, but as perl.h disables assert even
70 without NDEBUG (thank you very much), not too many people should
71 notice (that did include myself). Andreas König noticed, though :)
72 - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have
73 it. Also noticed by Andreas König :)
74 - save/restore tainted status.
75 - verified to pass the testsuite on my 5.9.5.
76
773.1 Mon Dec 4 23:03:40 CET 2006
78 - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow).
79 - incompatible change: transfer flags are now per-state.
80 - give a better error message on deadlock.
81 - document Coro::nready.
82 - enhanced testsuite.
83
843.01 Sun Dec 3 23:47:42 CET 2006
85 - forgot to include Coro::Timer.
86
873.0 Sun Dec 3 22:57:25 CET 2006
88 - the "FINALLY COMPLETELY STABLE" release (coming soon:
89 the "FAMOUS LAST WORDS" release).
90 - implement a new stack sharing algorithm, which uses a stack
91 pool (size currently hardcoded).
92 - make stack sharing mandatory (it no longer uses a heuristic).
93 - eval/die no longer cause weird problems under heavy use.
94 - Coro::Event could cause livelocks if it was used but
95 no Coro::Event watchers were used.
96 - Coro::Event now uses asynccheck as prepare does not
97 check for changed watchers.
98 - Coro::Event allows multiple waiting coros and will wake up one
99 per event.
100 - Coro::Event should be cleaner and more efficient now.
101 - new utility function Coro::unblock_sub.
102 - document the sad fact that Event is no longer reentrant.
103 - putting a coroutine into the ready queue twice could under
104 some circumstances lead to stack corruption.
105 - minor incompatibility: subclassing Coro::State is supported
106 directly now without going through a _coro_state member.
107 - state/coro switching is much faster now.
108 - very minor optimisations and code/documentation cleanup.
109 - avoid problems due to compiler inlining.
110 - removed timers from Coro::Timer -> use AnyEvent instead.
111 - replaced Coro::idle coroutine by (cleaner) idle handler.
112 - updated to newest libcoro.
113 - implement enhanced support for valgrind.
114 - implement is_ready and return value for ready.
115 - removed Coro::Cont, it was a misnomer (it's generators),
116 and it was somewhat annoying to get it right. Will come back
117 if somebody asks for it :)
118 - many ->wait methods and Coro::Event could return
119 spuriously without the event having happened.
120
1212.5 Tue Nov 7 12:22:33 CET 2006
122 - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket
123 use AnyEvent, moved them to Coro/.
124 - added Coro::LWP which contains all the uglyness required to
125 make LWP non-blocking.
126 - should work with perl 5.9.x now (Andreas König made me do it).
127 - fixed another bug in Coro::Select when the timeout was undef.
128 - reuse PL_start_env for all coros, saves some memory per coroutine.
129 - manage PL_top_env differently, hopefully to avoid panic: top_env.
130 - timeout argument was not properly used in Coro::Socket.
131 - allow limited forms of subclassing in Coro::Handle/Coro::Socket.
132 - emulate undocumented(!) functionality of IO::Socket required
133 by LWP(!!).
134 - updated eg/lwp to work with newer lwp's.
135 - remove "FATAL: uncaught exception" prefix. Coroutines that die
136 kill the whole process, just as exceptions in the main "coroutine"
137 did already.
138
1392.1 Wed Nov 1 23:01:13 CET 2006
140 - fix a long-standing bug in Coro::Select where select with
141 zero timeout would instead change the current default filehandle.
142 - use a simpler and hopefully more robust way to clone padlists
143 (uses less memory and a perl function instead of our own).
144 - coro can now create a stack guard area on many architectures.
145 - Coro::AIO properly reexports additional functions from IO::AIO.
146 - updated libcoro with a workaround for OS X,
147 pach and testing by Michael Schwern.
148
1492.0 Tue Oct 24 05:47:17 CEST 2006
150 - support additional aio requests in Coro::AIO.
151
1521.9 (never properly released due to a glitch)
153
1541.8 Thu Feb 2 00:59:06 CET 2006
155 - applied suggested patch by SAMV to avoid problems during stupid
156 mark & sweep gc run.
157 - applied patch by Scott Walters for 5.9.3 compatibility.
158
1591.7 Tue Dec 27 01:41:58 CET 2005
160 - added Coro::AIO, a thin wrapper around IO::AIO.
161 - improved Makefile.PL explanations.
162
1631.6 Mon Dec 12 21:30:05 CET 2005
164 - additionally save PM_curpm might fix as-of-yet
165 unseen problems with regex matches being attributed to the
166 wrong package.
167 - add t/10_bugs.t, which currently checks against imho broken
168 perls that use 0.26MB of stack space per Perl_magic_get
169 invocation (newer linuxes) as opposed to the 0.0002MB perl
170 normally uses.
171 - make stacksize configurable for the ultimate debian experience.
172
1731.51 Mon Dec 12 18:48:36 CET 2005
174 - remove debugging warn() accidentally left in Coro::Select.
175
1761.5 Tue Nov 29 13:32:44 CET 2005
177 - use Coro::Event inside Coro::Select to avoid spurious deadlocks.
178 - fix Coro::Select 'select'.
179 - strict'ify some modules.
180 - libcoro errornously restored the SIGUSR2 handler to SIGUSR1.
181 - use XSLoader in selected modules.
182 - remove some 5.6 compatibility cruft.
183
1841.4 Tue Sep 6 00:11:05 CEST 2005
185 - libcoro did not take into account the trampoline on amd64,
186 when 'l'inux method was used.
187
1881.31 Tue Aug 30 23:31:33 CEST 2005
189 - some portability fixes/workarounds.
190
1911.3 Sat Aug 20 03:08:56 CEST 2005
192 - no code changes. module seems to work fine.
193 - improve Coro::State docs, remove reference to nonexisting
194 Coro::State::flush method.
195 - no longer autodetect windows, present it as an option instead.
196
1971.2 Mon May 16 02:00:55 CEST 2005
198 - included libcoro.c earlier in State.xs, might improve portability.
199 - use a faster and possibly more stable Coro::Cont implementation.
200 - accept x86_64 in addition to amd64 for optimized linux-amd64 support.
201 - fix bugs Coro::SemaphoreSet that could cause locks never to be freed.
202 - fix bugs in CoroAPI.h and document it in Coro::MakeMaker.
203
2041.11 Thu Mar 3 18:00:52 CET 2005
205 - change of contact address.
206
2071.1 Tue Feb 22 20:51:16 CET 2005
208 - support [l]inux method on amd64.
209 - allow some context switches while compiling/eval'ing.
210
2111.0 Tue Aug 10 03:47:27 CEST 2004
212 - Event 0.89 fixes the bug again.
213 - fix the remaining(??) memleak. I didn't free the cache-AV for
214 padlists, so one AV per created anon-sub leaked.
215
2160.97 Fri May 14 15:23:32 CEST 2004
217 - remove stupid gcc extension that was an experiment only anyway.
218
2190.96 Thu May 13 18:09:29 CEST 2004
220 - tell people if they have yet another broken Event variant.
221 - made Coro to work _better_ with threads (it seems that, as long
222 as you use Coros in one thread only everything will work fine.
223 Looking at the perl source the problem might be missing locking
224 between Coro and some perl internal routines. That's not easily
225 fixable).
226 - (hopefully) got rid of the static xnext inside transfer.
227
2280.95 Sun Feb 15 03:20:28 CET 2004
229 - removed Coro::State::flush, implemented a new and much
230 faster caching mechanism by attaching magic to the CV.
231 - a memleak with (real) closures remains.
232 - dramatically reduced size of stacks to 65536 (linux/x86), after
233 observing that even myhttpd never had more than a kilobyte
234 of stack in use (stack is still only physically allocated in
235 pagesize increments for systems suporting mmap).
236 - fix for a "cygwin-64" architecture added, proposed by Stefan Traby.
237 - fix for threaded perl.
238
2390.9 Sun Nov 23 04:32:39 CET 2003
240 - fix a few very hard to track down but (of course) stupid
241 memory leaks.
242 - stringify version number to avoid locale problems :(.
243 - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters,
244 hopefully ignoring it "just works".
245
2460.8 Wed Nov 5 19:38:40 CET 2003
247 - port to perl5.005_03, but only 5.8.x+ is supported!
248 - honor LocalAddr even without LocalPort in Coro::Socket.
249
2500.7 Tue May 27 03:12:38 CEST 2003
251 - the version jump indicates some level of testing, not gobs
252 of new features.
253 - uh... I found the prompt function in ExtUtils::MakeMaker.
254 Highly correct stuff, that is...
255 - fixed(?) a bug with die's in coroutines causing "panic: top_env".
256 the fix is not well understood by the author, so beware :(.
257
2580.652 Thu May 8 02:54:46 CEST 2003
259 - Applied patch by Slaven Rezic to set default to "s" on FreeBSD,
260 cause version 4 doesn't have ucontext.
261 - Benjamin Reed reported that setjmp works fine on darwin,
262 so preselect it.
263
2640.651 Sat Mar 29 15:00:23 CET 2003
265 - fix a bug in Coro::Handle where some bytes could get lost
266 on reading, reported by jason@nichego.net.
267
2680.65 Sun Mar 23 00:08:26 CET 2003
269 - added a README.linux-glibc.
270 - new module Coro::Select.
271 - also save/restore PL_comppad, fixes segfaults with 5.8.1.
272 (I never claimed that I know what I am doing ;).
273 - set default to setjmp/longjmp on non-x86-linux.
274
2750.6 Thu Nov 21 11:09:06 CET 2002
276 - made lazy context switching the default.
277 - fixed the bug where SemaphoreSet::guard
278 was the same as timed_guard and timed_guard was missing.
279 - fix a memleak in Coro::Timer.
280 - reclassified context sharing as not experimental.
281
2820.534 Sun Apr 14 03:05:12 CEST 2002
283 - fixed backspace => tab in header parsing (myhttpd).
284 - added eg/readline, for elmex.
285 - Coro::Event now calls ->start on first call instead of
286 ->again, which created wrong timeouts for the first ->next.
287 - fixed a bug where Coro::Socket returned a closed socket instead
288 of nothing on ECONNREFUSED and similar error conditions.
289
2900.533 Mon Feb 18 18:49:40 CET 2002
291 - adapt to finally working Event-0.85.
292 - fixed buggy event initialization (trapped under windows, but how
293 could it possibly have worked anywhere?)
294
2950.532 Wed Jan 16 02:45:32 CET 2002
296 - removed the reference to pp_entersub, might work on
297 (some) win32 perls now (testsuite works under cygwin,
298 without Event, which freezes).
299
3000.531 Mon Dec 10 22:18:44 CET 2001
301 - Fixed a bug in SemaphoreSet::timed_down.
302
3030.53 Tue Nov 27 21:11:13 CET 2001
304 - some tests for Event.
305 - slightly different internal architecture to get refcnt into the object
306 and not the reference. might fix a few bugs, certainly did introduce
307 new ones.
308 - Coro::Timer (independent of Event).
309 - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet.
310
3110.52 Tue Nov 6 21:36:18 CET 2001
312 - ported to cygwin (trivial).
313 patch by Gerrit P. Haase <gp@familiehaase.de>.
314 - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov.
315
3160.51 Thu Nov 1 20:39:01 CET 2001
317 - terminate/cancel now work properly (otherwise termination
318 could cause a "next coroutine is not and contains not..."
319 error.
320 - added Coro::Socket::shutdown.
321 - Coro::Event::loop no is the same as Event::loop.
322 - implemented terminate with args + join.
323
3240.5 Fri Sep 28 16:15:35 CEST 2001
325 - fixed "print" on a Coro::Handle. The print method worked.
326 - small tweaks (seem to reduce memory consumption a lot)
327 in various modules.
328 - splendid use of "no warnings" scattered throughout.
329 - added Coro::Handle::rbuf, fixed Coro::Handle::fh.
330
3310.49 Sun Sep 16 02:42:45 CEST 2001
4 - changed some method calls to function calls for speed 332 - changed some method calls to function calls for speed
5 inside Coro::Handle. 333 inside Coro::Handle.
6 - make Coro::Handle use an array instead of a hash for 334 - make Coro::Handle use an array instead of a hash for
7 speed reasons. 335 speed reasons.
8 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?". 336 - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?".
9 sjlj and ucontext should now work. 337 sjlj and ucontext should now work.
10 - IRIX-specific port for libcoro. 338 - IRIX-specific port for libcoro.
339 - swapped order of accept results to match IO::Socket.
340 - changed getsock/peername to sock/peername to match IO::Socket.
341 - fixed a bug that caused segfault when returning to main task
342 under some circumstances.
343 - other bugfixes.
11 344
120.45 Sun Sep 2 02:54:01 CEST 2001 3450.45 Sun Sep 2 02:54:01 CEST 2001
13 - new method Coro::Handle::timeout. 346 - new method Coro::Handle::timeout.
14 - corrected speling of set...name to get...name in Coro::Socket. 347 - corrected speling of set...name to get...name in Coro::Socket.
15 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle. 348 - Coro::Socket::accept now returns a Coro::Socket, not a Coro::Handle.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines