| 1 | Revision history for Perl extension Coro. |
1 | Revision history for Perl extension Coro. |
| 2 | |
2 | |
| 3 | TODO: should explore PerlIO::via::CoroCede and PerlIO::via::CoroAIO. |
3 | TODO: should explore PerlIO::coroaio (perl leaks like hell). |
| 4 | TODO: maybe implement a default message channel, very much like Erlang's |
4 | TODO: maybe implement a default message channel, very much like Erlang's |
| 5 | actor model (which is cool in a lot of important aspects (failures!), |
5 | actor model (which is cool in a lot of important aspects (failures!), |
| 6 | but very lacking in others (higher level ipc)). |
6 | but very lacking in others (higher level ipc)). |
| 7 | |
7 | |
|
|
8 | 5.0 |
|
|
9 | - NEW ARCHITECTURE: use 4.912 for the stable version. |
|
|
10 | - bump API version to 7 - all dependents must be recompiled. |
|
|
11 | - Coro::Semaphore->adjust didn't correctly wake up enough waiters. |
|
|
12 | - async_pool did free a scalar value twice |
|
|
13 | ("Attempt to unreference..."). |
|
|
14 | - perl's process emulation is now not even theoretically supported |
|
|
15 | anymore. |
|
|
16 | - directly patch the entersub opcode calling SLF functions (cede, |
|
|
17 | transfer and so on). this does speed up context switching, but |
|
|
18 | more importanly, it frees us from the hardcoded behaviour of |
|
|
19 | entersub, so we might actually be able to return something from |
|
|
20 | those functions and atcually create new ones. |
|
|
21 | - take advantage of __builtin_frame_address on gcc. |
|
|
22 | - expose THX in coroapi (not sure whether this was a wise decision, |
|
|
23 | as "threaded" perls are running at half speed anyways). |
|
|
24 | - implement execute_slf (schedule-like-function) interface that makes |
|
|
25 | it possible to implement schedule-like-functions in XS. |
|
|
26 | - use new SLF interface to massively speed up Coro::EV by roughly a factor |
|
|
27 | of two. |
|
|
28 | - removed timed_down and timed_guard functions - they were not being used |
|
|
29 | anyways and should be replaced by a more generic mechanism - and were |
|
|
30 | annoying to support anyways : |
|
|
31 | - used new SLF interface to massively speed up Coro::Semaphore by a factor |
|
|
32 | of three. |
|
|
33 | - forgot to include Coro/libcoro/README in the dist for all these years. |
|
|
34 | |
|
|
35 | 4.913 Sat Nov 15 07:58:28 CET 2008 |
|
|
36 | - async_pool did free a scalar value twice |
|
|
37 | ("Attempt to unreference..."). |
|
|
38 | |
|
|
39 | 4.912 Thu Nov 13 18:31:23 CET 2008 |
|
|
40 | - minor cleanups. |
|
|
41 | - use much larger stacks on linux and perl < 5.8.8. |
|
|
42 | - Coro::Debug::new_unix_server did not unlink the socket |
|
|
43 | when destroyed. |
|
|
44 | |
|
|
45 | 4.911 Tue Nov 11 04:26:17 CET 2008 |
|
|
46 | - "port" to threaded perls. |
|
|
47 | |
|
|
48 | 4.91 Mon Nov 10 05:36:38 CET 2008 |
|
|
49 | - the ->throw exception object no longer leaks. |
|
|
50 | - creating a new cctx leaked a scopestack entry (memleak). |
|
|
51 | - new coroutines didn't get created with a zero flags field |
|
|
52 | (unknown impact). |
|
|
53 | - calling ->throw on a not-yet-started coroutine should now work |
|
|
54 | instead of being ignored. |
|
|
55 | - ->throw is now supported on Coro::State objects. |
|
|
56 | - clean up cctx creation code a bit. |
|
|
57 | - entersub is actually a UNOP, not a LOGOP (not a bugfix). |
|
|
58 | |
|
|
59 | 4.9 Sat Nov 8 17:45:27 CET 2008 |
|
|
60 | - (libcoro) did not preserve rbp with CORO_ASM (we are getting there). |
|
|
61 | - (libcoro) no longer leak threads in the experimental pthread backend, |
|
|
62 | also speed it up considerably. |
|
|
63 | - (libcoro) do not rely on makecontext passing void *'s unscathed. |
|
|
64 | - fix compiletime dependencies on libcoro in the Makefile. |
|
|
65 | - cctx_count wasn't always updated properly. |
|
|
66 | - Coro::State::cctx_stacksize wasn't applied correctly. |
|
|
67 | - new function Coro::State::cctx_max_idle. |
|
|
68 | - the default max number of idle C contexts is now 4. |
|
|
69 | - (libcoro) try harder to get _setjmp/_longjmp. |
|
|
70 | - (libcoro) cleanup and extend the libcoro API to officially |
|
|
71 | allow the creation of empty source contexts. |
|
|
72 | - very experimental workaround for the totally broken netbsd platform. |
|
|
73 | - tried to hack around openbsd bugs. |
|
|
74 | |
|
|
75 | 4.804 Wed Nov 5 16:36:00 CET 2008 |
|
|
76 | - Coro::Debug::new_unix_server would not create a non-blocking listening |
|
|
77 | socket, sometimes causing freezes. |
|
|
78 | - (libcoro) fix misaligned stack points for setjmp and assembly |
|
|
79 | methods, which can cause crashes on x86/x86_64 with a sufficiently |
|
|
80 | aggressive compiler. |
|
|
81 | - new function: Coro::Debug::new_tcp_server. |
|
|
82 | - move ->throw into the Coro class because it only works on coro objects. |
|
|
83 | |
|
|
84 | 4.803 Mon Nov 3 17:16:12 CET 2008 |
|
|
85 | - (libcoro) use a global asm statement to become independent of gcc |
|
|
86 | otpimisations for CORO_ASM (thanks to pippijn for the idea). |
|
|
87 | - try to workaround yet another broken bsd, this time dragonfly. |
|
|
88 | |
|
|
89 | 4.802 Thu Oct 30 10:56:12 CET 2008 |
|
|
90 | - support -fno-omit-frame-pointer on x86 with the assembly method. |
|
|
91 | - tune 01_unblock.t tests a bit. |
|
|
92 | |
|
|
93 | 4.801 Wed Oct 22 18:33:37 CEST 2008 |
|
|
94 | - improve readline speed for very long "lines". |
|
|
95 | - backport to 5.8.8. |
|
|
96 | |
|
|
97 | 4.8 Thu Oct 2 13:34:40 CEST 2008 |
|
|
98 | - new function Coro::AIO::aio_wait. |
|
|
99 | - Coro.:AIO and Coro::BDB now "use Coro::AnyEvent". |
|
|
100 | - greatly speed up and reduce memory usage of Coro::AIO requests. |
|
|
101 | - implement some other ยต-optimisations. |
|
|
102 | |
|
|
103 | 4.749 Mon Sep 29 14:40:12 CEST 2008 |
|
|
104 | - port to slow and broken pseudo-threaded perls. (courtesy pippijn). |
|
|
105 | |
|
|
106 | 4.748 Sat Sep 27 14:03:19 CEST 2008 |
|
|
107 | - implement, but do not document, PerlIO::cede(granularity). |
|
|
108 | - Coro::Storable forgot to wrap Storable::pstore. |
|
|
109 | - work around the multitude of leaks and memory corruption |
|
|
110 | bugs in PerlIO::via by using our own C-level perliol. As a side |
|
|
111 | effect, Coro::Storable is now much, much, much faster. |
|
|
112 | |
|
|
113 | 4.747 Tue Sep 23 01:59:41 CEST 2008 |
|
|
114 | - fix a per-cv memleak (one empty array was leaked per |
|
|
115 | code reference). |
|
|
116 | - avoid a crash in coro->call|rss when the coroutine was already |
|
|
117 | destroyed (most noticably when using Coro::Debug::ps :) |
|
|
118 | - also protect *Storable::FILE. |
|
|
119 | - push up default storable granularity to 20ms. |
|
|
120 | |
|
|
121 | 4.746 Sun Sep 21 03:22:20 CEST 2008 |
|
|
122 | - be more insistent on locking Storable against reentrancy |
|
|
123 | in Coro::Storable. |
|
|
124 | - move swap_def?v and throw to Coro::State, as documented. |
|
|
125 | |
|
|
126 | 4.745 Thu Jul 24 00:14:38 CEST 2008 |
|
|
127 | - remove debugging code related to MgPV_nolen_const, also try to |
|
|
128 | make it compile with perl 5.8.6 (yes, apple apparently loves |
|
|
129 | outdated software). Reported by John S. |
|
|
130 | |
|
|
131 | 4.744 Tue Jul 8 22:06:35 CEST 2008 |
| 8 | - correctly provide default DIE/WARN handlers as documented. |
132 | - correctly provide default DIE/WARN handlers as documented. |
| 9 | - also overwrite PL_vtbl_sigelem.svt_clear, even though current |
133 | - also overwrite PL_vtbl_sigelem.svt_clear, even though current |
| 10 | implementations inside perl work fine for us. |
134 | implementations inside perl work fine for us. |
| 11 | |
135 | |
| 12 | 4.743 Mon Jun 16 00:21:57 CEST 2008 |
136 | 4.743 Mon Jun 16 00:21:57 CEST 2008 |
