* Added support for Radeon X1050, which is essentially an RV370.
I cannot confirm this works, but I assume the creator of the patch can. :-)
Thanks a lot, fixes#3435.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34182 a95241bf-73f2-0310-859d-f6bbb57e9c96
PThreads didn't have their array for the TLS values initialized.
From what we can tell, this would have been a problem for any program using
pthreads, but since all threads are pthreads now, it was much more likely to
be encountered. Like in Beam as reported in #4949 (which via libbind seems
to use some pthread stuff).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34161 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Properly register playlist mimetype in MediaPlayer. Double clicking them
will now open them in MediaPlayer.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34143 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Resolved TODO: Use the existing "FileChanged" notification in the main
window to adopt the UI to the currently playing item.
* When opening a file takes some time, the window is now free to start hidden
and setup a message runner to unhide it after 150 msecs.
* When launching MediaPlayer with files, the 150 msecs delay is used to
start the window hidden and pop up right at the correct location for audio.
If opening the first audio file takes less than 150 msecs, the window will
already show as soon as possible.
* While opening a file in the Controller, the window will say so in the
disabled seek slider instead of showing the message "Drop media files here.".
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34132 a95241bf-73f2-0310-859d-f6bbb57e9c96
the views in the FrameResized() hook which does the layout had some
unanticipated results. We need to know relative to the views themselves. This
fixes the video view being visible despite audio-only content.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34129 a95241bf-73f2-0310-859d-f6bbb57e9c96
again. Bute this commit seems to make the old bug that some parts of the GUI
are black much more likely. Will look into it tomorrow.
* The calculations to move the window wholy into the screen frame were buggy.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34120 a95241bf-73f2-0310-859d-f6bbb57e9c96
* The target volume block size is passed to the CalcItemsAndSize() function
to get a more accurate picture of whether items will fit. Before, it would
assume hard coded 1024 bytes block size.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34107 a95241bf-73f2-0310-859d-f6bbb57e9c96
HasStatus() and UpdateStatus() were not run in the same lock, so the job
view could be gone anyway. And secondly, both methods iterate over all status
items to find the one with the correct thread. Doh, if we don't have enough
work to do, lets make some! Of course this was not critical in any case...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34104 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Simplified the handling of the first window and removed FirstWindow(),
we can tell by fPlayerCount.
* Both the above would fix a crash when requesting the about alert with the
first window already gone. Respectively another player instance opening
if fFirstWindow was reset to NULL after some recent revision.
* Implemented restoring the current playlist, index and position in the file.
* Devised a more robust way to solve asynchronous seeking. The Controller is
now notified that a seek request has been handled with a dedicated hook.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34079 a95241bf-73f2-0310-859d-f6bbb57e9c96
and empty windows only):
* New windows use the width from the window that was last resized by the user.
* The first window uses the stored window position and size.
* When switching from video content to audio content, the last window width
from previous audio-only content is restored.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@34078 a95241bf-73f2-0310-859d-f6bbb57e9c96
first I tried adding support for changing the format during running the node
connection, but later found out that this is not implemented in the system
mixer (it has this weird setting to allow input format changes, but if you
do this, the media_server will just crash, since the backend does not support
it yet). Also, the Media Kit documentation is extremely lacking in this regard.
I ended up re-establishing the node connection when the audio format is
supposed to change, just like it is already done for video. This means that
audio files now play with their native channel count and frame rate. But it
isn't so well tested yet, if 48 kHz for example introduce some clicks. The
channel count should not be a problem though, I've also tested that with some
movies and it works fine.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33678 a95241bf-73f2-0310-859d-f6bbb57e9c96
a button or within the bar background within the timeout for the repeat-thread,
the thread would exit and there would be no repeating. I think there is still
a race condition somewhere since I managed to see it stall once again, probably
because it locks the looper twice in the loop...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33666 a95241bf-73f2-0310-859d-f6bbb57e9c96
manager and inform it that the video target, ie the view will be gone. The
Controller needs to use the SetTo() method in order to actually close down
the nodes properly in the destructor. Otherwise asynchronous threads will use
the BMediaFile and tracks after it has been deleted (until the NodeManager
destructor is executed).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33662 a95241bf-73f2-0310-859d-f6bbb57e9c96
Quad are all reported as "Core 2", because the CPU count is already reported
separately.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33646 a95241bf-73f2-0310-859d-f6bbb57e9c96
that. But I didn't think of any better solution either... Hopefully these are
all, but I will find out once I built a complete image.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33641 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Rewrote Devices application. -> Thanks a bunch!! Nice work!
* Put Devices onto the image.
Changes by myself:
* Merge some code from the old "preflet" which the new app still relied on.
* Fixed some coding style issues. I didn't review all of the code yet, though.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33640 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Tracker forwards auto mounter related messages to the mount_server.
* Rewrote AutoMounterSettings to not know AutoMounter and use
the layout-management.
* Moved the "Eject When Unmounting" setting into the Mount Settings.
* Launch the mount_server during boot, but delay the script until all
previously mounted volumes have been mounted. This solves some annoying
timing bugs during boot. For example when you have desktop backgrounds
on other volumes and some servers don't deal well with the situation
of links to add-ons on other volumes becoming valid with a delay...
* src/kits/tracker/Commands.h includes the private headers/private/
mount/MountServer.h header, which made adjustments to the DiskUsage
Jamfile necessary.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33555 a95241bf-73f2-0310-859d-f6bbb57e9c96
the team with the signature becomes a valid target for BMessages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33553 a95241bf-73f2-0310-859d-f6bbb57e9c96
anymore).
* Added interface to get the current settings.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33486 a95241bf-73f2-0310-859d-f6bbb57e9c96
a dedicated mount server. This is pretty much a straight copy from the
AutoMounter code from Tracker, except
* the eject on unmount setting has been added (in Tracker, it's part of the
general settings, not the mount specific ones),
* scripting features have been added, such that it becomes possible to trigger
mounting the previoulsy mounted volumes from the outside, and most
importantly block until the operation is done (waiting for the reply).
TODO:
* Change Tracker to not run it's own AutoMounter, but send messages to the
new server.
* Move the eject when unmounting setting to the mount settings window.
* Enable the mount_server in the Bootscript.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33484 a95241bf-73f2-0310-859d-f6bbb57e9c96
enhanced it to also say Core 2 Quad and Core 2 Duo again, instead of just
Core 2. Thanks!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33470 a95241bf-73f2-0310-859d-f6bbb57e9c96
I.e. you need to hold down exactly Ctrl+Alt for the server side window
modifiers.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33459 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Trigger server side window modifiers only if exactly those modifiers are
pressed.
* Add the new window flags defined in Window.h to what the server considers
valid flags. Otherwise the new flags important to the server (only
B_NO_SERVER_SIDE_WINDOW_MODIFIERS) get filtered out.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33458 a95241bf-73f2-0310-859d-f6bbb57e9c96
Otherwise libnetwork.so needs to be dragged in, which has some issues.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33454 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Declaring the structs doesn't work so well for the libbe_test target.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33452 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Remove Fonts from the app_server test environment build, since liblocal.so
is not yet part of that target platform.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33450 a95241bf-73f2-0310-859d-f6bbb57e9c96
#4707, this is the BeOS behavior. I have not checked myself and the BeBook is
very brief on this, but similar methods do trim indices to be in range.
Another alternative would be returning an index indicating an error (-1).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33439 a95241bf-73f2-0310-859d-f6bbb57e9c96
The audio decoding in AVDecoder needs this to work at all.
* Set the infoBuffer and infoSize correctly in GetStreamInfo(). At least this
is what I extract from what the AVDecoder expects.
* Use a slightly more precise timeStamp calculation in the Seek() method.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33358 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Added TODO about a probable mix-up to use the media_format meta data
because the FFmpeg Reader plug-in forgot to set the info-buffer correctly
from GetStreamInfo().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33355 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Some key-sequences cause \0 chars in the "bytes" data which is supposed to
be passed to BView::KeyDown() and BView::KeyUp(). Therefor, one cannot use
string methods for adding/extracting the data to/from the events. For
example, Control-Space now works in the Terminal.
Thanks a lot for the original patch, Joshua!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33338 a95241bf-73f2-0310-859d-f6bbb57e9c96
rebooting via pressing Ctr-Alt-Del twice did not work. (But I didn't test)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33335 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Fixed upward scrolling of a partial screen which is anchored at line 0.
Thanks a lot!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33318 a95241bf-73f2-0310-859d-f6bbb57e9c96
* In _InsertGap(), be sure not to set a line's length longer than the space
allocated for it.
Nice catch, thanks a bunch!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33317 a95241bf-73f2-0310-859d-f6bbb57e9c96
Fix fgetln() in libbsd to actually return the length of the string.
Thanks a bunch!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33315 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Use the correct interface index for the control transfer that sets a Wacom
tablet into tablet mode.
* Check if the mode is indeed setup correctly.
* Retry switching the mode up to five times, as done in the Linux driver.
Thanks a bunch! As Michael has proposed in ticket #3744, the whole Wacom
driver should be merged into the existing HID driver framework (eventually).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33301 a95241bf-73f2-0310-859d-f6bbb57e9c96
This is important when setting up the TARGET_LIB* constants later, in case
the target platform is not "haiku". (i.e. libbe_test).
* Added TODO about HOST_GCC_BASE_FLAGS being wrong for GCC4. I couldn't find
the proper fix, but it is important whan compiling libbe_test on Haiku.
I worked around this by having the constant defined correctly for the
problematic file. See the TODO comment.
* Small cleanups.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33246 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Make sure settings are saved rightaway whenever they change. (Solves a
ticket which I cannot seem to find right now.)
* Some improvements to the NamePanel layout/size (when setting descriptions).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33167 a95241bf-73f2-0310-859d-f6bbb57e9c96
on the target image instead of extracting them, but only if the target image
is a CD image. This then speeds up installing a great lot, at the expense of
certain applications not being available in LiveCD mode. This affects
* the revision control tools and their dependency libraries,
* autotools
* texinfo
* perl and python
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32996 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Changed the Bootscript and the InstallerFinishScript to only create and
cleanup a link on the Desktop.
* Applied patch by idefix to fix removing the Installer link from volumes
that contain spaces in the name. (Thanks!)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32995 a95241bf-73f2-0310-859d-f6bbb57e9c96
Decided it would be easiest there, since it already needed to know about them.
Tested with a whole bunch of our packages and it appends them correctly to the
AboutSystem COPYRIGHTS attribute.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32986 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Integrated ProgressReporter into CopyEngine and UnzipEngine.
* Integrated unzipping to main installer engine. Since there are currently no
.zip files in the _packages_ folder (and not even a _packages_ folder
itself), this is completely invisible, but it works.
TODO: Adjust the build system to optionally copy packages into _packages_
instead of extracting them. Like for the CD build profile mode.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32967 a95241bf-73f2-0310-859d-f6bbb57e9c96
simplified the code somewhat (maintaining of state).
* UnzipEngine does a more complete parsing of the .zip contents and maintains
a hashmap of item names to byte size entries. During unzipping later on, this
allows to know how many bytes were unzipped for the given item.
* Test whether folders from the .zip already exists, and exclude them from the
item count, since unzip will not report creating folders that already
existed.
* Added the unzipping of packages in the _package_ folder to WorkerThread, but
it is currently commented out to keep Installer in a working state. It would
work, but the problem is that the progress bar would reset itself at the
moment (for every package).
* Started a ProgressReporter class that will later be used by all CopyEngine and
UnzipEngine instances. This is work in progress and not yet part of the build.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32959 a95241bf-73f2-0310-859d-f6bbb57e9c96
that will fix the problem fixed in r32926 for all derived classes.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32957 a95241bf-73f2-0310-859d-f6bbb57e9c96
reporting is incomplete and it is not yet used for anything, it's not even
part of the build. I am testing it in a selfcontained little test.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32905 a95241bf-73f2-0310-859d-f6bbb57e9c96
Updated to use all the new alpha1 packages. I have not yet done throughrough
testing myself. I figured that trunk should also use the newest packages.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32885 a95241bf-73f2-0310-859d-f6bbb57e9c96
Stephan Aßmus