| 1 |
2007-05-24 Stephane Thiell <mbuna@bugged.org> |
| 2 |
|
| 3 |
* Add IA-64 arch support |
| 4 |
* Cleanup in peak/arch to allow proper AC substitions |
| 5 |
* Rebuild autotools files |
| 6 |
* Added missing engine_mod_devpoll.c |
| 7 |
* Release version 0.1.2 |
| 8 |
|
| 9 |
2007-03-06 Stephane Thiell <mbuna@bugged.org> |
| 10 |
|
| 11 |
* Main for-loop for epoll_wait was wrong |
| 12 |
Submitted by Pascal Gloor <pascal.gloor@spale.com> |
| 13 |
|
| 14 |
2005-02-28 Stephane Thiell <mbuna@bugged.org> |
| 15 |
|
| 16 |
* Fixed a bug when the last timer is removed |
| 17 |
* Changed engine private-extern function names |
| 18 |
|
| 19 |
2004-01-18 Stephane Thiell <mbuna@bugged.org> |
| 20 |
|
| 21 |
* Release of libpeak 0.1.0 (first public release). |
| 22 |
* configure.ac: Fixed NetBSD compilation. |
| 23 |
* CREDITS: New file with some info and credits. |
| 24 |
* INSTALL: New file with some info about libpeak's installation. |
| 25 |
* README.platforms: New file with some advices and info about |
| 26 |
currently supported platforms. |
| 27 |
|
| 28 |
2004-01-17 Stephane Thiell <mbuna@bugged.org> |
| 29 |
|
| 30 |
* peak/engine_mod_devpoll.c: Fixed a bug when registering |
| 31 |
signals in the engine (clients array wasn't allocated...), also |
| 32 |
when we want to remove a signal. |
| 33 |
* peak/arch/sparc: Added support for sparc processors. Tested a |
| 34 |
bit on sparc-sun-solaris2.6, 2.8 and 2.9. |
| 35 |
|
| 36 |
2004-01-12 Stephane Thiell <mbuna@bugged.org> |
| 37 |
|
| 38 |
* peak/timer.c, peak/task.c: Fixed a bug when reconfiguring a |
| 39 |
timer. |
| 40 |
* peak/semaphore.h: Fixed header includes to avoid a warning on |
| 41 |
FreeBSD 4.5. |
| 42 |
|
| 43 |
2004-01-10 Stephane Thiell <mbuna@bugged.org> |
| 44 |
|
| 45 |
* peak/task.c: Fixed a bug in __peak_task_timer_insert() when |
| 46 |
updating task->_tifirst. |
| 47 |
* peak/stream.c: Internal timer bug fix, following the changes in |
| 48 |
peak/task.c. |
| 49 |
* peak/task.h: Improved documentation for peak_task_timer_add() |
| 50 |
and peak_task_timer_remove(). |
| 51 |
* peak/task.c: Fixed peak_task_timer_remove(): for convenience, it |
| 52 |
does nothing if the timer is not added to the task. Missing code |
| 53 |
in peak_task_timer_add() and remove() so that they respectively |
| 54 |
retain and release the timer. |
| 55 |
|
| 56 |
2004-01-09 Stephane Thiell <mbuna@bugged.org> |
| 57 |
|
| 58 |
* peak/engine_mod_poll.c, peak/engine_mod_kqueue.c: Fixed code |
| 59 |
when removing handled signals. |
| 60 |
* tests/t_atomic1.c, tests/t_atomic2.c, tests/t_signal2.c: New |
| 61 |
tests. |
| 62 |
* peak/engine_mod_epoll.c: Fix bug (signal issue). Tested on |
| 63 |
Linux 2.6.1. |
| 64 |
|
| 65 |
2004-01-08 Stephane Thiell <mbuna@bugged.org> |
| 66 |
|
| 67 |
* tests/t_timer3.c: Modified threaded test program to use |
| 68 |
peak_atomic_dec(). Increased timeout delay too. |
| 69 |
* configure.ac, peak/peak.h: Added peak/atomic.h |
| 70 |
auto-generation. Added the peak_atomic's primitives to the API. |
| 71 |
* peak/engine_mod_epoll.c: Added new epoll() task's underlying |
| 72 |
engine (available on Linux 2.6). |
| 73 |
* peak/errors.c: Removed use of strerror_r() as it's broken |
| 74 |
on some Linux systems (glibc)... |
| 75 |
* peak/signal.h: Added documentation for peak_signal objects. |
| 76 |
Rebuilt docs. |
| 77 |
* configure.ac, peak/*: Fixed some 64-bit portability warnings |
| 78 |
on AIX. |
| 79 |
|
| 80 |
2004-01-07 Stephane Thiell <mbuna@bugged.org> |
| 81 |
|
| 82 |
* peak/semaphore.h, peak/arch/powerpc/spinlock.h: Fixed things |
| 83 |
to make libpeak compile on Linux PPC (YDL 3.0). |
| 84 |
* configure.ac, peak/Makefile.am, tests/Makefile.am, |
| 85 |
peak/arch/powerpc/atomic.h peak/arch/powerpc/spinlock.h: Added |
| 86 |
AIX support. Not tested with AIX < 5. |
| 87 |
|
| 88 |
2004-01-06 Stephane Thiell <mbuna@bugged.org> |
| 89 |
|
| 90 |
* acinclude.m4, tz.c: Added better struct tzhead availability. |
| 91 |
* peak/task.c, peak/timer_tree.c: Changed internal task's pending |
| 92 |
timers management from a linked list to an ordered splay tree |
| 93 |
(featuring a special "first access" in O(1)). Now instead of |
| 94 |
O(1)/O(n)/O(n) we have O(1)/O(log(N))/O(log(N)) for first |
| 95 |
accessor/any add/any remove complexity. And in practice, timers |
| 96 |
insertions and removals are often in constant time. |
| 97 |
|
| 98 |
2004-01-05 Stephane Thiell <mbuna@bugged.org> |
| 99 |
|
| 100 |
* tests: Added timer tests; more tests to come... |
| 101 |
* peak/task.c peak/timer.c: code cleaning for timers |
| 102 |
|
| 103 |
2004-01-04 Stephane Thiell <mbuna@bugged.org> |
| 104 |
|
| 105 |
* peak/task.c: Added TASK_FLAG_RUNNING to avoid setting |
| 106 |
TASK_FLAG_ON_TIME when it's not running. |
| 107 |
* doc: rebuilt docs |
| 108 |
* peak/timer.h, peak/timer.c: Added peak_timer_set_context() |
| 109 |
* tests: New tests directory |
| 110 |
* tests/test-suite.pl: New test suite script (to do) |
| 111 |
* tests/t_timer1.c: New little timer test program |
| 112 |
* Enabled make check to compile and run the (future) test suite. |
| 113 |
|
| 114 |
2004-01-03 Stephane Thiell <mbuna@bugged.org> |
| 115 |
|
| 116 |
* peak/time.c: Fixed a bad array subscript type |
| 117 |
|
| 118 |
2004-01-02 Stephane Thiell <mbuna@bugged.org> |
| 119 |
|
| 120 |
* peak/arch/test/spinlocktest.c: Added this little spinlock test |
| 121 |
program. Improved a bit the atomic test program and added an |
| 122 |
option to enable normal code instead of atomic one (to see that's |
| 123 |
the program fails). |
| 124 |
* peak/arch/m68k: Added support for m68k (probably only 68020+), |
| 125 |
assembly code only tested on m68k-unknown-netbsdelf1.6ZC. |
| 126 |
* peak/stream.c: Removed fatal error from |
| 127 |
__peak_stream_send_queued() and peak_stream_write_buffered(), as |
| 128 |
asynchronous errors can happen there. For now, ignore any errors, |
| 129 |
it should be handled by the engine at the next event loop. |
| 130 |
|
| 131 |
2003-12-30 Stephane Thiell <mbuna@bugged.org> |
| 132 |
|
| 133 |
* Sourceforge.net import |
| 134 |
|
