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