Source Activity

Syndicate content
Haiku's main repository
Updated: 34 min 13 sec ago

Update gcc5 packages.

Fri, 2016-05-13 04:59
* Add gnutls, libtasn1, nettle x86 packages for x86_gcc2. * Add gnutls, libtasn1, nettle x86 packages for x86_gcc2.
Categories: Development

BluetoothServer.cpp: fix gcc6 build

Thu, 2016-05-12 19:44
Signed-off-by: Augustin Cavalier Signed-off-by: Augustin Cavalier
Categories: Development

Updated WebPositive bookmarks

Thu, 2016-05-12 14:58
Changed the Launcher URLs to the website's user guide and welcome page instead of the trunk at http://cgit.haiku-os.org/, because that'll only show the html source code. Changed http to https in the bookmarks where possible. Changed the Launcher URLs to the website's user guide and welcome page instead of the trunk at http://cgit.haiku-os.org/, because that'll only show the html source code. Changed http to https in the bookmarks where possible.
Categories: Development

Update pci.ids from pciids.sourceforge.net

Thu, 2016-05-12 03:20
Categories: Development

intel_extreme: Shorten graphics family names

Wed, 2016-05-11 17:35
Categories: Development

intel_extreme: Rework PLL calculation

Tue, 2016-05-10 04:12
* More like linux, improved G4x calculations * Reduce un-needed pll limit complexity * Improved pll limits on ports based on type * More like linux, improved G4x calculations * Reduce un-needed pll limit complexity * Improved pll limits on ports based on type
Categories: Development

intel_extreme: Tab cleanup; no functional change.

Mon, 2016-05-09 21:36
* It seems like the Atom editor has been injecting spaces in strange spots. * Clean up spaces. Should help gcc6 as well. * Sorry for the mess! Atom is on probation. * It seems like the Atom editor has been injecting spaces in strange spots. * Clean up spaces. Should help gcc6 as well. * Sorry for the mess! Atom is on probation.
Categories: Development

intel_extreme: Extend DDI port probing to A-E

Sun, 2016-05-08 20:39
* The Linux code made this a bit hard to figure out via complex define functions, however there can be up to 5 DDI ports (A-E) * The Linux code made this a bit hard to figure out via complex define functions, however there can be up to 5 DDI ports (A-E)
Categories: Development

bluetooth_server: Nuke the weird logging system, just use macros...

Sun, 2016-05-08 01:09
... just like everything else on Haiku does. Permanently enabled for the time being. At least I can tell what's going on now. ... just like everything else on Haiku does. Permanently enabled for the time being. At least I can tell what's going on now.
Categories: Development

preferences/bluetooth: More UI creation fixes.

Sun, 2016-05-08 00:18
Hopefully fixes a "pure virtual call" crash I was getting in UI kit layout code in the InquiryPanel. Hopefully fixes a "pure virtual call" crash I was getting in UI kit layout code in the InquiryPanel.
Categories: Development

Nuke the "DUN" (Dial-Up Networking) preflet.

Sat, 2016-05-07 22:13
It literally did nothing except look nice^W^W^W^Wterrible, as it didn't use the Layout Kit and it pre-dates the Haiku coding style (last "real" modification was sometime before 2005). Supposedly it saved its settings to disk but I can't find that. See the "pppconfig" and "ppp_up" applications in src/bin/network for something that actually works (though the modem stuff in the kernel PPP stack hasn't been rewritten, so those are PPPoE only at the moment.) It literally did nothing except look nice^W^W^W^Wterrible, as it didn't use the Layout Kit and it pre-dates the Haiku coding style (last "real" modification was sometime before 2005). Supposedly it saved its settings to disk but I can't find that. See the "pppconfig" and "ppp_up" applications in src/bin/network for something that actually works (though the modem stuff in the kernel PPP stack hasn't been rewritten, so those are PPPoE only at the moment.)
Categories: Development

virtualkeyboard: Delete some of the copied classes and use the originals.

Sat, 2016-05-07 21:25
The keyboard view class has too many modifications to be trivially merged. The keyboard view class has too many modifications to be trivially merged.
Categories: Development

BMediaEventLooper::SetBufferDuration: Clamp if negative

Sat, 2016-05-07 17:21
* For the sake of making the code more backward compatible, and safer eventually. * For the sake of making the code more backward compatible, and safer eventually.
Categories: Development

launch_media_server: Use BLaunchRoster::Start

Sat, 2016-05-07 15:52
* Remove SyncToServices and the related code from BMediaRoster. * Remove SyncToServices and the related code from BMediaRoster.
Categories: Development

readdir_r(): uses the logic of readdir()...

Sat, 2016-05-07 14:35
to retrieve more than one entry. * readdir_r() is now affected by rewinddir(), thus fixing #12755. * extends dirent_test with readdir_r(). to retrieve more than one entry. * readdir_r() is now affected by rewinddir(), thus fixing #12755. * extends dirent_test with readdir_r().
Categories: Development

BMediaEventLooper: Handle events in past differently

Sat, 2016-05-07 11:26
* When the loop didn't handle an event in performance time, we will attempt at reading the port only if there are messages using a null timeout. * When the loop didn't handle an event in performance time, we will attempt at reading the port only if there are messages using a null timeout.
Categories: Development

MediaEventLooper: Calculate lateness from the nearest value

Sat, 2016-05-07 11:17
* When the event is late there are two major situations, the first is when the enqueue time is in past compared to the performance time. The second happen when the performance time is in past compared to the enqueue time. The purpose of this change is to calculate the lateness considering always the nearest of the two values. This way the exceding latency is trimmed out from the calculus depending on what the actual source of delay is. * There is a lot more that could be done by looking at the previous lateness value. * When the event is late there are two major situations, the first is when the enqueue time is in past compared to the performance time. The second happen when the performance time is in past compared to the enqueue time. The purpose of this change is to calculate the lateness considering always the nearest of the two values. This way the exceding latency is trimmed out from the calculus depending on what the actual source of delay is. * There is a lot more that could be done by looking at the previous lateness value.
Categories: Development

Update translations from Pootle

Sat, 2016-05-07 04:49
Categories: Development