| 1 | Revision history for Perl extension Coro. |
1 | Revision history for Perl extension Coro. |
| 2 | |
2 | |
| 3 | TODO: check wether stack-sharing is still effective |
3 | TODO: better (earlier) error messages on broken transfer's |
| 4 | TODO: more effective stack sharing by having one "default stack" for "simple" coroutines? |
4 | TODO: destroy gets only called on cancel: fix(?) |
| 5 | TODO: verify exception frame changes and force stack copies? |
|
|
| 6 | |
5 | |
|
|
6 | - remove t/09_timer.t, as it isn't really testing much |
|
|
7 | but was rather flaky in practise. |
|
|
8 | - async_pool coro would keep arguments and callback alive until |
|
|
9 | it was reused. |
|
|
10 | |
|
|
11 | 3.3 Sat Jan 6 03:45:00 CET 2007 |
|
|
12 | - implement $coro->on_destroy. |
|
|
13 | - Coro::Event blocking semantics have been changed, |
|
|
14 | documented and - hopefully - improved. |
|
|
15 | - fix nice adding, not subtracting, from priority. |
|
|
16 | - fix ->prio and api_is_ready (patch by Mark Hinds). |
|
|
17 | - fixed an assert ("... == prev__cctx->idle_te") |
|
|
18 | that could errronously trigger. |
|
|
19 | - fix various large and small memleaks. |
|
|
20 | - use a (hopefully) more stable cancel implementation |
|
|
21 | that immediately frees the coroutine data. |
|
|
22 | - cede/cede_notself return a status now. |
|
|
23 | - added Coro::guard function. |
|
|
24 | - added a global coroutine pool for jobs (on my machine, |
|
|
25 | I can create and execute 48k simple coros/s with async, |
|
|
26 | and 128k coros with async_pool). |
|
|
27 | - Coro::AIO now uses the coroutine priority as io priority. |
|
|
28 | |
|
|
29 | 3.2 Fri Dec 22 05:07:09 CET 2006 |
|
|
30 | - improve portability to slightly older perls. |
|
|
31 | - use cleaner coroutine destruction. |
|
|
32 | - simplify configuration for users. |
|
|
33 | - optionally (unrecommended) prefer perl functions over |
|
|
34 | their coro replacements. |
|
|
35 | |
|
|
36 | 3.11 Tue Dec 5 13:11:24 CET 2006 |
|
|
37 | - fixed some bogus assert's, but as perl.h disables assert even |
|
|
38 | without NDEBUG (thank you very much), not too many people should |
|
|
39 | notice (that did include myself). Andreas König noticed, though :) |
|
|
40 | - do not save/restore PL_sortcxix on >= 5.9.x, it doesn't seem to have |
|
|
41 | it. Also noticed by Andreas König :) |
|
|
42 | - save/restore tainted status. |
|
|
43 | - verified to pass the testsuite on my 5.9.5. |
|
|
44 | |
|
|
45 | 3.1 Mon Dec 4 23:03:40 CET 2006 |
|
|
46 | - INCOMPATIBLE CHANGE: $/ is now per-coroutine (but slow). |
|
|
47 | - incompatible change: transfer flags are now per-state. |
|
|
48 | - give a better error message on deadlock. |
|
|
49 | - document Coro::nready. |
|
|
50 | - enhanced testsuite. |
|
|
51 | |
|
|
52 | 3.01 Sun Dec 3 23:47:42 CET 2006 |
|
|
53 | - forgot to include Coro::Timer. |
|
|
54 | |
|
|
55 | 3.0 Sun Dec 3 22:57:25 CET 2006 |
|
|
56 | - the "FINALLY COMPLETELY STABLE" release (coming soon: |
|
|
57 | the "FAMOUS LAST WORDS" release). |
| 7 | - reimplement a new stack sharing algorithm. |
58 | - implement a new stack sharing algorithm, which uses a stack |
|
|
59 | pool (size currently hardcoded). |
| 8 | - make stack sharing mandatory (it no longer uses a heuristic). |
60 | - make stack sharing mandatory (it no longer uses a heuristic). |
|
|
61 | - eval/die no longer cause weird problems under heavy use. |
| 9 | - Coro::Event could cause livelocks if it was used but |
62 | - Coro::Event could cause livelocks if it was used but |
| 10 | Event watchers were used directly. |
63 | no Coro::Event watchers were used. |
|
|
64 | - Coro::Event now uses asynccheck as prepare does not |
|
|
65 | check for changed watchers. |
|
|
66 | - Coro::Event allows multiple waiting coros and will wake up one |
|
|
67 | per event. |
| 11 | - Coro::Event should be cleaner and more efficient now. |
68 | - Coro::Event should be cleaner and more efficient now. |
|
|
69 | - new utility function Coro::unblock_sub. |
|
|
70 | - document the sad fact that Event is no longer reentrant. |
|
|
71 | - putting a coroutine into the ready queue twice could under |
|
|
72 | some circumstances lead to stack corruption. |
| 12 | - minor incompatibility: subclassing is supported directly now |
73 | - minor incompatibility: subclassing Coro::State is supported |
| 13 | without going through a _coro_state member. |
74 | directly now without going through a _coro_state member. |
| 14 | - process swicthing is much faster now. |
75 | - state/coro switching is much faster now. |
| 15 | - very minor optimisations and code/documentation cleanup. |
76 | - very minor optimisations and code/documentation cleanup. |
| 16 | - avoid problems due to compiler inlining. |
77 | - avoid problems due to compiler inlining. |
| 17 | - possibly better support for threads. |
|
|
| 18 | - removed timers from Coro::Timer -> use AnyEvent instead. |
78 | - removed timers from Coro::Timer -> use AnyEvent instead. |
| 19 | - replaced Coro::idle coroutine by (cleaner) idle handler. |
79 | - replaced Coro::idle coroutine by (cleaner) idle handler. |
|
|
80 | - updated to newest libcoro. |
|
|
81 | - implement enhanced support for valgrind. |
|
|
82 | - implement is_ready and return value for ready. |
|
|
83 | - removed Coro::Cont, it was a misnomer (it's generators), |
|
|
84 | and it was somewhat annoying to get it right. Will come back |
|
|
85 | if somebody asks for it :) |
|
|
86 | - many ->wait methods and Coro::Event could return |
|
|
87 | spuriously without the event having happened. |
| 20 | |
88 | |
| 21 | 2.5 Tue Nov 7 12:22:33 CET 2006 |
89 | 2.5 Tue Nov 7 12:22:33 CET 2006 |
| 22 | - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket |
90 | - made Coro::Util, Coro::Select, Coro::Handle and Coro::Socket |
| 23 | use AnyEvent, moved them to Coro/. |
91 | use AnyEvent, moved them to Coro/. |
| 24 | - added Coro::LWP which contains all the uglyness required to |
92 | - added Coro::LWP which contains all the uglyness required to |
