[00:02:03] <CIA-110> mmlr * r41693 /haiku/branches/releases/r1alpha3/src/add-ons/kernel/busses/usb/ehci.cpp:
[00:02:03] <CIA-110> Manually apply r41690 to the r1alpha3 branch as it fixes EHCI interrupt issues
[00:02:03] <CIA-110> on some chipsets.
[00:48:44] <CIA-110> bonefish * r41694 /haiku/branches/developer/bonefish/signals/src/bin/debug/strace/ (MemoryReader.cpp MemoryReader.h strace.cpp):
[00:48:44] <CIA-110> * MemoryReader: Replaced the constructor by a default constructor and a new
[00:48:44] <CIA-110> Init() method.
[00:48:44] <CIA-110> * strace.cpp:
[00:55:10] <CIA-110> bonefish * r41695 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp: handle_signals(): Fixed incorrect check whether signals are debugged.
[01:12:57] <CIA-110> bonefish * r41696 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp:
[01:12:57] <CIA-110> Removed SIGNAL_FLAG_DONT_RESTART_SYSCALL support for the send_signal_to_*()
[01:12:57] <CIA-110> team and process group variants. The flag is only used for resume_thread() and
[01:12:57] <CIA-110> is not needed at team level.
[01:28:47] <CIA-110> bonefish * r41697 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp:
[01:28:47] <CIA-110> deliver_signal_locked(), send_signal_to_team_locked(): If SIGCONT isn't blocked
[01:28:47] <CIA-110> it needs to interrupt the target.
[02:49:32] <CIA-110> bonefish * r41698 /haiku/branches/developer/bonefish/signals/ (4 files in 3 dirs):
[02:49:32] <CIA-110> Removed the SIGNAL_FLAG_DONT_RESTART_SYSCALL flag entirely and introduced the
[02:49:32] <CIA-110> kernel internal signal SIGNAL_CONTINUE_THREAD, which is used by resume_thread()
[02:49:32] <CIA-110> instead. This prevents the Haiku thread management from interfering with POSIX
[02:49:32] <CIA-110> semantics.
[03:06:23] <CIA-110> bonefish * r41699 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/sys/wait.cpp: waitpid(): Fixed returns.
[03:07:16] <CIA-110> bonefish * r41700 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/sys/wait.cpp: waitpid(): Fixed the broken fix. In case of success the child pid is returned.
[03:10:58] <CIA-110> bonefish * r41701 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp:
[03:10:58] <CIA-110> The SIGNAL_CONTINUE_THREAD priority must not be > that of SIGSTOP, since it
[03:10:58] <CIA-110> can otherwise overtake the latter.
[03:18:51] <CIA-110> bonefish * r41702 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp:
[03:18:51] <CIA-110> _user_sigaction(): Call sigaction_internal() instead of sigaction(). The
[03:18:51] <CIA-110> former returns the error code directly, which is what we expect here. Fixes
[03:18:51] <CIA-110> incorrect error codes.
[03:36:31] <CIA-110> bonefish * r41703 /haiku/branches/developer/bonefish/signals/headers/posix/unistd.h: We do support realtime signals now.
[03:37:15] <CIA-110> bonefish * r41704 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/signal/sigset_accessors.cpp:
[03:37:15] <CIA-110> * Fixed sigdelset() implementations.
[03:37:15] <CIA-110> * Mini style cleanup.
[04:04:31] <CIA-110> bonefish * r41705 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/signal/sigignore.cpp: Fixed copy and paste bug: The sigignore() symbol wasn't defined correctly.
[05:19:39] <CIA-110> anevilyak * r41706 /haiku/trunk/ (9 files in 9 dirs): (log message trimmed)
[05:19:39] <CIA-110> * Slightly alter the semantics of the input gathering function to take the
[05:19:39] <CIA-110> menu item it's associated with rather than an input string. This allows it
[05:19:39] <CIA-110> to calculate the position to start the input at, as well as the correct
[05:19:39] <CIA-110> line to place it on. The previous solution always put the input at the
[05:19:40] <CIA-110> center line, which happened to be the right place by happy coincidence
[05:19:41] <CIA-110> unless one also had the menu items for viewing/saving the debug syslog
[14:08:43] <CIA-110> anevilyak * r41707 /haiku/trunk/src/system/boot/loader/menu.cpp: Cleanup: no need for parsing since the kernel settings will handle semicolon separated options directly. Thanks Axel!
[14:31:51] <CIA-110> anevilyak * r41708 /haiku/trunk/src/system/boot/loader/menu.cpp: Build fix, removed one line too many.
[14:51:12] *** kurain has joined #haiku-jp
[14:51:22] *** kurain has left #haiku-jp
[15:31:46] <CIA-110> bonefish * r41709 /haiku/branches/developer/bonefish/signals/ (3 files in 2 dirs): (log message trimmed)
[15:31:46] <CIA-110> * handle_signals(): When handling a stop signal suspend the thread right there
[15:31:46] <CIA-110> instead of returning and letting the caller (thread_at_kernel_exit()) do that.
[15:31:46] <CIA-110> * Before suspending the thread, but when already holding the scheduler lock,
[15:31:46] <CIA-110> check whether there are pending continue/kill signals. This remedies a race
[15:31:46] <CIA-110> condition with the signal delivery code, which, upon delivery of such a
[15:31:47] <CIA-110> signal, unsuspends the thread, but of course only when it was already
[15:54:02] <CIA-110> bonefish * r41710 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/errno.c:
[15:54:02] <CIA-110> _to_positive_error(): B_POSIX_ENOMEM is negative, so we need to negate the
[15:54:02] <CIA-110> value.
[17:46:32] <CIA-110> scottmc * r41711 /haiku/trunk/build/jam/ (OptionalBuildFeatures OptionalPackages):
[17:46:32] <CIA-110> Updated a few dozen gcc4 OptionalPackages for r1a3. Note that these were all
[17:46:32] <CIA-110> built in an r1a3 gcc4 only environment.
[17:46:32] <CIA-110> +alpha3
[18:02:29] <CIA-110> bonefish * r41712 /haiku/branches/developer/bonefish/signals/ (3 files in 2 dirs): (log message trimmed)
[18:02:29] <CIA-110> Fixed sigsuspend() (hopefully for good now):
[18:02:29] <CIA-110> * We keep the changed signal mask after returning from sigsuspend_internal().
[18:02:29] <CIA-110> That's the one that needs to stay in effect also while the signal handler is
[18:02:29] <CIA-110> invoked (potentially plus additionally blocked signals).
[18:02:30] <CIA-110> * Renamed Thread::sig_temp_enabled to sigsuspend_original_unblocked_mask. It
[18:02:31] <CIA-110> stores the (inverted) thread's original signal mask now.
[18:42:47] <CIA-110> bonefish * r41713 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/sigqueue/ (12-1.c 3-1.c): Do not assume that root's name is "root".
[18:45:15] <CIA-110> bonefish * r41714 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp: deliver_signal_locked(), send_signal_to_team_locked(): Added permission checks.
[18:59:21] <CIA-110> mmlr * r41715 /haiku/trunk/src/add-ons/kernel/bus_managers/ata/ATADevice.cpp: Fix typo in error output and a little automatic whitespace cleanup.
[20:00:02] <CIA-110> humdinger * r41716 /haiku/trunk/docs/welcome/images/flags/ (ru.png zh_CN.png): Added Chinese and Russian flags. Fixes #7566. Thanks taos for noticing.
[20:36:49] <CIA-110> korli * r41717 /haiku/trunk/src/kits/midi/MidiSynth.cpp: Patch from Pete with changes by myself (bug #7558): BMidiSynth destructor should delete be_synth.
[21:11:06] <CIA-110> pulkomandy * r41718 /haiku/branches/releases/r1alpha3/ (34 files in 27 dirs): Merge from trunk.
[21:21:42] <CIA-110> mmlr * r41719 /haiku/trunk/build/scripts/determine_haiku_revision:
[21:21:42] <CIA-110> Use gsed instead of sed on Mac OS X as the builtin sed doesn't support all the
[21:21:42] <CIA-110> options we need. This still requires gsed to be installed from ports.
[21:21:42] <CIA-110> Fixes #7563.
[22:34:05] <CIA-110> scottmc * r41720 /haiku/trunk/build/jam/OptionalPackages: Updated several more gcc4 OptionalPackages, including sed, tar, vision, pe and others.
[23:12:27] <CIA-110> mmlr * r41721 /haiku/trunk/build/scripts/ (build_cross_tools_gcc4 determine_haiku_revision): Replace the $(...) syntax with backticks instead.





top