[00:21:02] <CIA-111> korli * r41986 /haiku/trunk/src/kits/storage/disk_device/Partition.cpp: Second try at fixing #7572 (after r41842). Please review.
[00:22:47] <CIA-111> bonefish * r41987 /haiku/branches/developer/bonefish/signals/headers/private/shared/syscall_utils.h: Added macros RETURN_AND_TEST_CANCEL() and RETURN_AND_SET_ERRNO_TEST_CANCEL().
[00:24:42] <CIA-111> bonefish * r41988 /haiku/branches/developer/bonefish/signals/src/system/kernel/lib/ (Jamfile kernel_lib.h):
[00:24:42] <CIA-111> Added header kernel_lib.h, which is auto-included in all kernel/lib sources.
[00:24:42] <CIA-111> Currently only contains an empty pthread_testcancel() macro.
[00:25:44] <CIA-111> bonefish * r41989 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp:
[00:25:44] <CIA-111> handle_signals(): Prevent syscall restarts when handling the thread
[00:25:44] <CIA-111> cancellation signal.
[00:30:10] <CIA-111> bonefish * r41990 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/pthread/pthread_cancel.cpp:
[00:30:10] <CIA-111> pthread_cancel(): Call _kernel_cancel_thread() also when the thread hasn't set
[00:30:10] <CIA-111> the asynchronous cancellation flag. Otherwise we wouldn't interrupt a syscall.
[00:34:47] <CIA-111> bonefish * r41991 /haiku/branches/developer/bonefish/signals/src/ (22 files in 7 dirs): Made all functions cancellation points that POSIX requires to be ones.
[00:44:43] <CIA-111> mmlr * r41992 /haiku/trunk/src/add-ons/kernel/drivers/input/usb_hid/Driver.cpp:
[00:44:43] <CIA-111> Add devices we have explicit quirky device support for to the support
[00:44:43] <CIA-111> descriptor. This allows us to support devices that aren't strictly HID or only
[00:44:44] <CIA-111> provide vendor specific interfaces by constructing report descriptors for their
[00:44:44] <CIA-111> data format.
[00:57:58] <CIA-111> bonefish * r41993 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/timer_delete/speculative/:
[00:57:58] <CIA-111> Removed useless tests. There's really no point in observing what happens when
[00:57:58] <CIA-111> an invalid timer is passed to timer_delete().
[01:00:53] <CIA-111> bonefish * r41994 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/timer_getoverrun/speculative/: Removed more bogus tests with invalid timers.
[01:03:12] <CIA-111> bonefish * r41995 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/timer_gettime/speculative/: Removed more bogus tests with invalid timers.
[01:05:13] <CIA-111> bonefish * r41996 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/timer_settime/speculative/: Removed even more bogus tests with invalid timers.
[01:23:55] <CIA-111> bonefish * r41997 /haiku/branches/developer/bonefish/signals/src/system/kernel/UserTimer.cpp:
[01:23:55] <CIA-111> _user_create_timer(): Only try to get the thread creation attributes when the
[01:23:55] <CIA-111> notification method is SIGEV_THREAD.
[01:29:04] <CIA-111> bonefish * r41998 /haiku/branches/developer/bonefish/signals/headers/private/shared/syscall_utils.h: Made C-friendly.
[01:30:06] <CIA-111> bonefish * r41999 /haiku/branches/developer/bonefish/signals/src/system/libroot/posix/time/nanosleep.c:
[01:30:06] <CIA-111> Unlike clock_nanosleep(), nanosleep() doesn't return the error directly, but
[01:30:06] <CIA-111> sets errno instead.
[01:45:46] <CIA-111> mmlr * r42000 /haiku/trunk/src/add-ons/kernel/drivers/input/usb_hid/ (Driver.cpp QuirkyDevices.cpp QuirkyDevices.h):
[01:45:46] <CIA-111> * Add device class, subclass and protocol to the quirky device structure.
[01:45:46] <CIA-111> * Add these to the usb_support_descriptor when adding from the quirky devices.
[01:45:46] <CIA-111> * Extend the SIXAXIS comment a bit.
[01:51:25] <CIA-111> mmlr * r42001 /haiku/trunk/src/add-ons/kernel/drivers/input/usb_hid/ (HIDWriter.cpp HIDWriter.h QuirkyDevices.cpp):
[01:51:25] <CIA-111> * Remove the report ID argument of HIDWriter::DefineInput{Data|Paddding} to
[01:51:25] <CIA-111> allow for creating descriptors that don't use report IDs at all.
[01:51:25] <CIA-111> * Adjust sixaxis_build_descriptor and set the global report ID once at the
[01:51:25] <CIA-111> beginning instead.
[02:23:16] <CIA-111> mmlr * r42002 /haiku/trunk/src/add-ons/kernel/drivers/input/usb_hid/ (Driver.cpp HIDDevice.cpp HIDDevice.h):
[02:23:16] <CIA-111> Move the quirky device detection out of the HIDDevice constructor and into the
[02:23:16] <CIA-111> usb_hid_device_added hook so that we can actually detect non-HID devices at all.
[02:32:50] <CIA-111> mmlr * r42003 /haiku/trunk/src/add-ons/kernel/drivers/input/usb_hid/QuirkyDevices.cpp:
[02:32:50] <CIA-111> * Add a quirky device entry for XBOX 360 controllers.
[02:32:50] <CIA-111> * As they don't claim to be HID, they don't provide a report descriptor either,
[02:32:50] <CIA-111> so we build one. The built descriptor is similar to the one found in FreeBSD
[02:32:51] <CIA-111> for the same class of devices. Might fix/help #7649, for lack of hardware I
[02:32:51] <CIA-111> can't test this.
[11:42:19] *** echelog-2 has joined #haiku-jp
[13:58:58] *** echelog-2 has joined #haiku-jp
[14:43:18] <CIA-111> stippi * r42004 /haiku/trunk/ (3 files in 2 dirs): (log message trimmed)
[14:43:18] <CIA-111> Applied patch by Pete Goodeve from ticket #7182. It improves
[14:43:18] <CIA-111> keyboard navigation/tracking of BMenus and BMenuBars, although
[14:43:18] <CIA-111> many issues remain.
[14:43:18] <CIA-111> Should not yet go into alpha, since there is one issue which
[14:43:19] <CIA-111> I am not sure if it's not a regression. The issue is that
[14:43:20] <CIA-111> invoking a menu item with Enter for the first time seems to
[16:36:13] <CIA-111> bonefish * r42005 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/timer_create/10-1.c: Fixed the test: nanosleep() doesn't let the CPU time pass, so spin instead.
[16:39:13] <CIA-111> bonefish * r42006 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/timer_create/11-1.c: Fixed the test: nanosleep() doesn't let the CPU time pass, so spin instead.
[17:03:40] <CIA-111> kallisti5 * r42007 /haiku/trunk/data/artwork/icons/Device_Ramdisk: Device_Harddisk + virtual memory == Device_Ramdisk
[18:17:33] <CIA-111> bonefish * r42008 /haiku/trunk/src/kits/network/libnetapi/NetEndpoint.cpp: Automatic whitespace cleanup.
[18:33:06] <CIA-111> bonefish * r42009 /haiku/trunk/src/kits/network/libnetapi/NetEndpoint.cpp:
[18:33:06] <CIA-111> * Added TODOs where IsDataPending() is called incorrectly. Since it has stupid
[18:33:06] <CIA-111> semantics, I'd recommend introducing an _IsDataPending() with sane semantics
[18:33:06] <CIA-111> and use that.
[18:33:07] <CIA-111> * IsDataPending() Added TODO regarding possible overflow.
[18:33:07] <CIA-111> * SetTimeout(): Added TODO regarding consistency.
[18:34:57] <CIA-111> bonefish * r42010 /haiku/branches/developer/bonefish/signals/src/system/kernel/timer.cpp: Added KDL command "timers" to list the scheduled timers.
[18:49:30] <CIA-111> bonefish * r42011 /haiku/branches/developer/bonefish/signals/src/system/kernel/signal.cpp:
[18:49:30] <CIA-111> sigwait_internal(): Use a SignalHandledCaller instead of a BReference, so that
[18:49:30] <CIA-111> Signal::Handled() is called as required.
[19:34:15] <CIA-111> bonefish * r42012 /haiku/branches/developer/bonefish/signals/src/system/kernel/thread.cpp:
[19:34:15] <CIA-111> * common_snooze_etc(): Fixed comment.
[19:34:15] <CIA-111> * _user_snooze_etc(): The wrong size was passed to the user_memcpy() copying
[19:34:16] <CIA-111> the remaining time value back to userland.
[19:46:46] <CIA-111> bonefish * r42013 /haiku/branches/developer/bonefish/signals/src/system/kernel/UserTimer.cpp: Made fOverrunCount computation overflow-safe.
[19:47:14] <CIA-111> bonefish * r42014 /haiku/branches/developer/bonefish/signals/headers/private/system/user_timer_defs.h: Bumped MAX_USER_TIMER_OVERRUN_COUNT to INT_MAX.
[19:49:29] <CIA-111> leavengood * r42015 /haiku/trunk/build/jam/ (FileRules ReleaseBuildProfiles):
[19:49:29] <CIA-111> Apply patch from mmadia to add a Release Notes file to the alpha image desktop.
[19:49:29] <CIA-111> Allow wget to output to files with spaces in the DownloadFile rule.
[19:49:29] <CIA-111> Fixes #7609.
[20:04:41] <CIA-111> kallisti5 * r42016 /haiku/trunk/src/add-ons/kernel/file_systems/ramfs/ (5 files): change format of fs_volume to match current kernel filesystem interface
[20:05:52] <CIA-111> leavengood * r42017 /haiku/trunk/src/kits/tracker/ContainerWindow.cpp:
[20:05:53] <CIA-111> Revert the old change r36762 which added a small offset to tracker right click
[20:05:53] <CIA-111> menus since it is no longer needed after r40132/r40142. Thanks diver for the
[20:05:53] <CIA-111> reminder.
[20:05:53] <CIA-111> #5809 can finally be closed again.
[20:06:29] <CIA-111> kallisti5 * r42018 /haiku/trunk/src/add-ons/kernel/file_systems/ramfs/cpp.cpp: undo accidental change to cpp.cpp in r42016
[20:14:25] <CIA-111> mmlr * r42019 /haiku/trunk/src/add-ons/kernel/drivers/input/usb_hid/HIDDevice.cpp: (log message trimmed)
[20:14:25] <CIA-111> Get the HID descriptor from the already retrieved configuration info instead of
[20:14:25] <CIA-111> always calling the device and requesting it. If no HID descriptor is found in
[20:14:25] <CIA-111> the configuration we fall back to retrieving it manually and if that fails we
[20:14:25] <CIA-111> fall back to a dummy report descriptor length. In the most usual case this will
[20:14:26] <CIA-111> save one unnecessary USB control request.
[20:14:26] <CIA-111> Incidentally this also fixes the absolute pointing device under VirtualBox. VBox
[21:21:34] <CIA-111> bonefish * r42020 /haiku/branches/developer/bonefish/signals/ (2 files in 2 dirs): (log message trimmed)
[21:21:34] <CIA-111> * SystemTimeUserTimer, RealTimeUserTimer: Moved the kernel timer scheduling to
[21:21:34] <CIA-111> new method ScheduleKernelTimer().
[21:21:34] <CIA-111> * Changed all UserTimer implementations to not use periodic kernel timers
[21:21:34] <CIA-111> anymore. Instead one-shot kernel timers are used and rescheduled after the
[21:21:35] <CIA-111> event was fired. This is necessary to maintain a precise overrun count.
[21:21:36] <CIA-111> * Introduced a minimum interval by which the kernel timer start time is advanced
[21:35:19] <CIA-111> bonefish * r42021 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/clock_gettime/ (8-2.c coverage.txt):
[21:35:20] <CIA-111> Removed bogus test. It is not specified which clock IDs are negative, so
[21:35:20] <CIA-111> testing arbitrary values isn't useful.
[21:36:11] <CIA-111> leavengood * r42022 /haiku/trunk/src/apps/installer/InstallerWindow.cpp:
[21:36:11] <CIA-111> Make the Installer status view a little wider by default and also resize it's
[21:36:11] <CIA-111> height if needed. This should prevent any text from being cut off (though to be
[21:36:11] <CIA-111> fair it is possible to resize the window horizontally to see any cut off text.)
[21:36:11] <CIA-111> Fixes #5882.
[21:36:12] <CIA-111> +alpha3
[21:39:59] <CIA-111> bonefish * r42023 /haiku/branches/developer/bonefish/signals/src/tests/system/libroot/posix/posixtestsuite/conformance/interfaces/clock_settime/ (17-1.c coverage.txt): Removed another bogus test trying some arbitrary clock ID.
[21:58:09] <CIA-111> scottmc * r42024 /haiku/trunk/build/jam/OptionalPackages: Updated vision OptionalPackage to include a default VisionSettings file. This fixes #6040.
[22:38:13] <CIA-111> yourpalal * r42025 /haiku/trunk/ (7 files in 2 dirs):
[22:38:13] <CIA-111> * Add new spacing/inset constants for the layout API. For now, leave B_USE_DEFAULT_SPACING, but this will go soon.
[22:38:13] <CIA-111> * Rename BControlLook::ComposeItemSpacing() to ComposeSpacing(), and extend it to handle the new constants.
[22:38:13] <CIA-111> * Adjust users of BControlLook::ComposeItemSpacing() accordingly.
[22:38:13] <CIA-111> * part of #7447





top