| 1 | Revision history for Perl extension Coro. |
1 | Revision history for Perl extension Coro. |
| 2 | |
2 | |
| 3 | 0.49 |
3 | TODO: transfer flags must be per state |
|
|
4 | TODO: better (earlier) error messages on broken transfer's |
|
|
5 | |
|
|
6 | 3. |
|
|
7 | - give a better error message on deadlock. |
|
|
8 | - enhanced testsuite. |
|
|
9 | |
|
|
10 | 3.01 Sun Dec 3 23:47:42 CET 2006 |
|
|
11 | - forgot to include Coro::Timer. |
|
|
12 | |
|
|
13 | 3.0 Sun Dec 3 22:57:25 CET 2006 |
|
|
14 | - the "FINALLY COMPLETELY STABLE" release (coming soon: |
|
|
15 | the "FAMOUS LAST WORDS" release). |
|
|
16 | - implement a new stack sharing algorithm, which uses a stack |
|
|
17 | pool (size currently hardcoded). |
|
|
18 | - make stack sharing mandatory (it no longer uses a heuristic). |
|
|
19 | - eval/die no longer cause weird problems under heavy use. |
|
|
20 | - Coro::Event could cause livelocks if it was used but |
|
|
21 | no Coro::Event watchers were used. |
|
|
22 | - Coro::Event now uses asynccheck as prepare does not |
|
|
23 | check for changed watchers. |
|
|
24 | - Coro::Event allows multiple waiting coros and will wake up one |
|
|
25 | per event. |
|
|
26 | - Coro::Event should be cleaner and more efficient now. |
|
|
27 | - new utility function Coro::unblock_sub. |
|
|
28 | - document the sad fact that Event is no longer reentrant. |
|
|
29 | - putting a coroutine into the ready queue twice could under |
|
|
30 | some circumstances lead to stack corruption. |
|
|
31 | - minor incompatibility: subclassing Coro::State is supported |
|
|
32 | directly now without going through a _coro_state member. |
|
|
33 | - state/coro switching is much faster now. |
|
|
34 | - very minor optimisations and code/documentation cleanup. |
|
|
35 | - avoid problems due to compiler inlining. |
|
|
36 | - removed timers from Coro::Timer -> use AnyEvent instead. |
|
|
37 | - replaced Coro::idle coroutine by (cleaner) idle handler. |
|
|
38 | - updated to newest libcoro. |
|
|
39 | - implement enhanced support for valgrind. |
|
|
40 | - implement is_ready and return value for ready. |
|
|
41 | - removed Coro::Cont, it was a misnomer (it's generators), |
|
|
42 | and it was somewhat annoying to get it right. Will come back |
|
|
43 | if somebody asks for it :) |
|
|
44 | - many ->wait methods and Coro::Event could return |
|
|
45 | spuriously without the event having happened. |
|
|
46 | |
|
|
47 | 2.5 Tue Nov 7 12:22:33 CET 2006 |
|
|
48 | - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket |
|
|
49 | use AnyEvent, moved them to Coro/. |
|
|
50 | - added Coro::LWP which contains all the uglyness required to |
|
|
51 | make LWP non-blocking. |
|
|
52 | - should work with perl 5.9.x now (Andreas König made me do it). |
|
|
53 | - fixed another bug in Coro::Select when the timeout was undef. |
|
|
54 | - reuse PL_start_env for all coros, saves some memory per coroutine. |
|
|
55 | - manage PL_top_env differently, hopefully to avoid panic: top_env. |
|
|
56 | - timeout argument was not properly used in Coro::Socket. |
|
|
57 | - allow limited forms of subclassing in Coro::Handle/Coro::Socket. |
|
|
58 | - emulate undocumented(!) functionality of IO::Socket required |
|
|
59 | by LWP(!!). |
|
|
60 | - updated eg/lwp to work with newer lwp's. |
|
|
61 | - remove "FATAL: uncaught exception" prefix. Coroutines that die |
|
|
62 | kill the whole process, just as exceptions in the main "coroutine" |
|
|
63 | did already. |
|
|
64 | |
|
|
65 | 2.1 Wed Nov 1 23:01:13 CET 2006 |
|
|
66 | - fix a long-standing bug in Coro::Select where select with |
|
|
67 | zero timeout would instead change the current default filehandle. |
|
|
68 | - use a simpler and hopefully more robust way to clone padlists |
|
|
69 | (uses less memory and a perl function instead of our own). |
|
|
70 | - coro can now create a stack guard area on many architectures. |
|
|
71 | - Coro::AIO properly reexports additional functions from IO::AIO. |
|
|
72 | - updated libcoro with a workaround for OS X, |
|
|
73 | pach and testing by Michael Schwern. |
|
|
74 | |
|
|
75 | 2.0 Tue Oct 24 05:47:17 CEST 2006 |
|
|
76 | - support additional aio requests in Coro::AIO. |
|
|
77 | |
|
|
78 | 1.9 (never properly released due to a glitch) |
|
|
79 | |
|
|
80 | 1.8 Thu Feb 2 00:59:06 CET 2006 |
|
|
81 | - applied suggested patch by SAMV to avoid problems during stupid |
|
|
82 | mark & sweep gc run. |
|
|
83 | - applied patch by Scott Walters for 5.9.3 compatibility. |
|
|
84 | |
|
|
85 | 1.7 Tue Dec 27 01:41:58 CET 2005 |
|
|
86 | - added Coro::AIO, a thin wrapper around IO::AIO. |
|
|
87 | - improved Makefile.PL explanations. |
|
|
88 | |
|
|
89 | 1.6 Mon Dec 12 21:30:05 CET 2005 |
|
|
90 | - additionally save PM_curpm might fix as-of-yet |
|
|
91 | unseen problems with regex matches being attributed to the |
|
|
92 | wrong package. |
|
|
93 | - add t/10_bugs.t, which currently checks against imho broken |
|
|
94 | perls that use 0.26MB of stack space per Perl_magic_get |
|
|
95 | invocation (newer linuxes) as opposed to the 0.0002MB perl |
|
|
96 | normally uses. |
|
|
97 | - make stacksize configurable for the ultimate debian experience. |
|
|
98 | |
|
|
99 | 1.51 Mon Dec 12 18:48:36 CET 2005 |
|
|
100 | - remove debugging warn() accidentally left in Coro::Select. |
|
|
101 | |
|
|
102 | 1.5 Tue Nov 29 13:32:44 CET 2005 |
|
|
103 | - use Coro::Event inside Coro::Select to avoid spurious deadlocks. |
|
|
104 | - fix Coro::Select 'select'. |
|
|
105 | - strict'ify some modules. |
|
|
106 | - libcoro errornously restored the SIGUSR2 handler to SIGUSR1. |
|
|
107 | - use XSLoader in selected modules. |
|
|
108 | - remove some 5.6 compatibility cruft. |
|
|
109 | |
|
|
110 | 1.4 Tue Sep 6 00:11:05 CEST 2005 |
|
|
111 | - libcoro did not take into account the trampoline on amd64, |
|
|
112 | when 'l'inux method was used. |
|
|
113 | |
|
|
114 | 1.31 Tue Aug 30 23:31:33 CEST 2005 |
|
|
115 | - some portability fixes/workarounds. |
|
|
116 | |
|
|
117 | 1.3 Sat Aug 20 03:08:56 CEST 2005 |
|
|
118 | - no code changes. module seems to work fine. |
|
|
119 | - improve Coro::State docs, remove reference to nonexisting |
|
|
120 | Coro::State::flush method. |
|
|
121 | - no longer autodetect windows, present it as an option instead. |
|
|
122 | |
|
|
123 | 1.2 Mon May 16 02:00:55 CEST 2005 |
|
|
124 | - included libcoro.c earlier in State.xs, might improve portability. |
|
|
125 | - use a faster and possibly more stable Coro::Cont implementation. |
|
|
126 | - accept x86_64 in addition to amd64 for optimized linux-amd64 support. |
|
|
127 | - fix bugs Coro::SemaphoreSet that could cause locks never to be freed. |
|
|
128 | - fix bugs in CoroAPI.h and document it in Coro::MakeMaker. |
|
|
129 | |
|
|
130 | 1.11 Thu Mar 3 18:00:52 CET 2005 |
|
|
131 | - change of contact address. |
|
|
132 | |
|
|
133 | 1.1 Tue Feb 22 20:51:16 CET 2005 |
|
|
134 | - support [l]inux method on amd64. |
|
|
135 | - allow some context switches while compiling/eval'ing. |
|
|
136 | |
|
|
137 | 1.0 Tue Aug 10 03:47:27 CEST 2004 |
|
|
138 | - Event 0.89 fixes the bug again. |
|
|
139 | - fix the remaining(??) memleak. I didn't free the cache-AV for |
|
|
140 | padlists, so one AV per created anon-sub leaked. |
|
|
141 | |
|
|
142 | 0.97 Fri May 14 15:23:32 CEST 2004 |
|
|
143 | - remove stupid gcc extension that was an experiment only anyway. |
|
|
144 | |
|
|
145 | 0.96 Thu May 13 18:09:29 CEST 2004 |
|
|
146 | - tell people if they have yet another broken Event variant. |
|
|
147 | - made Coro to work _better_ with threads (it seems that, as long |
|
|
148 | as you use Coros in one thread only everything will work fine. |
|
|
149 | Looking at the perl source the problem might be missing locking |
|
|
150 | between Coro and some perl internal routines. That's not easily |
|
|
151 | fixable). |
|
|
152 | - (hopefully) got rid of the static xnext inside transfer. |
|
|
153 | |
|
|
154 | 0.95 Sun Feb 15 03:20:28 CET 2004 |
|
|
155 | - removed Coro::State::flush, implemented a new and much |
|
|
156 | faster caching mechanism by attaching magic to the CV. |
|
|
157 | - a memleak with (real) closures remains. |
|
|
158 | - dramatically reduced size of stacks to 65536 (linux/x86), after |
|
|
159 | observing that even myhttpd never had more than a kilobyte |
|
|
160 | of stack in use (stack is still only physically allocated in |
|
|
161 | pagesize increments for systems suporting mmap). |
|
|
162 | - fix for a "cygwin-64" architecture added, proposed by Stefan Traby. |
|
|
163 | - fix for threaded perl. |
|
|
164 | |
|
|
165 | 0.9 Sun Nov 23 04:32:39 CET 2003 |
|
|
166 | - fix a few very hard to track down but (of course) stupid |
|
|
167 | memory leaks. |
|
|
168 | - stringify version number to avoid locale problems :(. |
|
|
169 | - perl 5.9.0 does not have SvPADBUSY, reported by Scott Walters, |
|
|
170 | hopefully ignoring it "just works". |
|
|
171 | |
|
|
172 | 0.8 Wed Nov 5 19:38:40 CET 2003 |
|
|
173 | - port to perl5.005_03, but only 5.8.x+ is supported! |
|
|
174 | - honor LocalAddr even without LocalPort in Coro::Socket. |
|
|
175 | |
|
|
176 | 0.7 Tue May 27 03:12:38 CEST 2003 |
|
|
177 | - the version jump indicates some level of testing, not gobs |
|
|
178 | of new features. |
|
|
179 | - uh... I found the prompt function in ExtUtils::MakeMaker. |
|
|
180 | Highly correct stuff, that is... |
|
|
181 | - fixed(?) a bug with die's in coroutines causing "panic: top_env". |
|
|
182 | the fix is not well understood by the author, so beware :(. |
|
|
183 | |
|
|
184 | 0.652 Thu May 8 02:54:46 CEST 2003 |
|
|
185 | - Applied patch by Slaven Rezic to set default to "s" on FreeBSD, |
|
|
186 | cause version 4 doesn't have ucontext. |
|
|
187 | - Benjamin Reed reported that setjmp works fine on darwin, |
|
|
188 | so preselect it. |
|
|
189 | |
|
|
190 | 0.651 Sat Mar 29 15:00:23 CET 2003 |
|
|
191 | - fix a bug in Coro::Handle where some bytes could get lost |
|
|
192 | on reading, reported by jason@nichego.net. |
|
|
193 | |
|
|
194 | 0.65 Sun Mar 23 00:08:26 CET 2003 |
|
|
195 | - added a README.linux-glibc. |
|
|
196 | - new module Coro::Select. |
|
|
197 | - also save/restore PL_comppad, fixes segfaults with 5.8.1. |
|
|
198 | (I never claimed that I know what I am doing ;). |
|
|
199 | - set default to setjmp/longjmp on non-x86-linux. |
|
|
200 | |
|
|
201 | 0.6 Thu Nov 21 11:09:06 CET 2002 |
|
|
202 | - made lazy context switching the default. |
|
|
203 | - fixed the bug where SemaphoreSet::guard |
|
|
204 | was the same as timed_guard and timed_guard was missing. |
|
|
205 | - fix a memleak in Coro::Timer. |
|
|
206 | - reclassified context sharing as not experimental. |
|
|
207 | |
|
|
208 | 0.534 Sun Apr 14 03:05:12 CEST 2002 |
|
|
209 | - fixed backspace => tab in header parsing (myhttpd). |
|
|
210 | - added eg/readline, for elmex. |
|
|
211 | - Coro::Event now calls ->start on first call instead of |
|
|
212 | ->again, which created wrong timeouts for the first ->next. |
|
|
213 | - fixed a bug where Coro::Socket returned a closed socket instead |
|
|
214 | of nothing on ECONNREFUSED and similar error conditions. |
|
|
215 | |
|
|
216 | 0.533 Mon Feb 18 18:49:40 CET 2002 |
|
|
217 | - adapt to finally working Event-0.85. |
|
|
218 | - fixed buggy event initialization (trapped under windows, but how |
|
|
219 | could it possibly have worked anywhere?) |
|
|
220 | |
|
|
221 | 0.532 Wed Jan 16 02:45:32 CET 2002 |
|
|
222 | - removed the reference to pp_entersub, might work on |
|
|
223 | (some) win32 perls now (testsuite works under cygwin, |
|
|
224 | without Event, which freezes). |
|
|
225 | |
|
|
226 | 0.531 Mon Dec 10 22:18:44 CET 2001 |
|
|
227 | - Fixed a bug in SemaphoreSet::timed_down. |
|
|
228 | |
|
|
229 | 0.53 Tue Nov 27 21:11:13 CET 2001 |
|
|
230 | - some tests for Event. |
|
|
231 | - slightly different internal architecture to get refcnt into the object |
|
|
232 | and not the reference. might fix a few bugs, certainly did introduce |
|
|
233 | new ones. |
|
|
234 | - Coro::Timer (independent of Event). |
|
|
235 | - new timed_wait functions for Coro::Signal, Semaphore, SemaphoreSet. |
|
|
236 | |
|
|
237 | 0.52 Tue Nov 6 21:36:18 CET 2001 |
|
|
238 | - ported to cygwin (trivial). |
|
|
239 | patch by Gerrit P. Haase <gp@familiehaase.de>. |
|
|
240 | - small setjmp code fix by Sullivan.DanielJ@epamail.epa.gov. |
|
|
241 | |
|
|
242 | 0.51 Thu Nov 1 20:39:01 CET 2001 |
|
|
243 | - terminate/cancel now work properly (otherwise termination |
|
|
244 | could cause a "next coroutine is not and contains not..." |
|
|
245 | error. |
|
|
246 | - added Coro::Socket::shutdown. |
|
|
247 | - Coro::Event::loop no is the same as Event::loop. |
|
|
248 | - implemented terminate with args + join. |
|
|
249 | |
|
|
250 | 0.5 Fri Sep 28 16:15:35 CEST 2001 |
|
|
251 | - fixed "print" on a Coro::Handle. The print method worked. |
|
|
252 | - small tweaks (seem to reduce memory consumption a lot) |
|
|
253 | in various modules. |
|
|
254 | - splendid use of "no warnings" scattered throughout. |
|
|
255 | - added Coro::Handle::rbuf, fixed Coro::Handle::fh. |
|
|
256 | |
|
|
257 | 0.49 Sun Sep 16 02:42:45 CEST 2001 |
| 4 | - changed some method calls to function calls for speed |
258 | - changed some method calls to function calls for speed |
| 5 | inside Coro::Handle. |
259 | inside Coro::Handle. |
| 6 | - make Coro::Handle use an array instead of a hash for |
260 | - make Coro::Handle use an array instead of a hash for |
| 7 | speed reasons. |
261 | speed reasons. |
| 8 | - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?". |
262 | - IRIX mystery solved: it's SGI's NT, after all: "standard, huh?". |
