we waste space, but we also incorrectly fail to update the size for all writes
that do not go beyond the previous file size.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32825 a95241bf-73f2-0310-859d-f6bbb57e9c96
the parent directory in those cases. Also, when updating the parent directory
we need to update the ".." dirent if we already populated it with a different
value before.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32823 a95241bf-73f2-0310-859d-f6bbb57e9c96
fixes#4358, there changing the Tracker preferences causes some default
directories to be created, like "home" and "config", which were previously
simply killing off the real ones.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32822 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Got rid of <dirent_private.h> -- the __DIR structure is private to dirent.c,
now. The attribute directory, index directory, and query functions use the
the public POSIX API, so does the kernel module code. Those components were
not initializing the structure correctly anymore since the introduction of
telldir()/seekdir().
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32819 a95241bf-73f2-0310-859d-f6bbb57e9c96
ensure that these nodes are never put. As they keep all changes in memory,
putting them and consequently freeing them destroys the modifications without
a way to get them back from the underlaying filesystem. In case a created node
was put but still had its entry in the parent directory this could also lead
to the situation where an invalid node was requested from the underlaying
filesystem, leading to #4347.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32815 a95241bf-73f2-0310-859d-f6bbb57e9c96
to " &" as really the whole rest of the file uses that style.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32813 a95241bf-73f2-0310-859d-f6bbb57e9c96
in / is created with the partition content name, pointing to /boot, where the
volume is actually mounted for real. The code is supposed to rename the link
in this case, not the actual folder. Added a note about the special case when
the boot volume is actually named "boot", in which case there will not be a
separate link. Then renaming /boot is prevented in the kernel now.
This commit completes the fix for ticket #4365.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32807 a95241bf-73f2-0310-859d-f6bbb57e9c96
switch /boot to a different volume in two operations, unless you have first
linked /system/lib into /bin. This patch assumes that / will always have the
ID 1. Don't know if that is proper. Note that I also thought about solving
this in the VFS, since perhaps it isn't the job of root-fs to know about
/boot, but that would of course introduce another check for every rename
operation, which I decided against.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32806 a95241bf-73f2-0310-859d-f6bbb57e9c96
codec tags in FFmpeg have changed. However, with the FFmpeg AC-3 decoder, I
am no longer getting sound. I suspect it's something to do with the channel
count.
P.S. I should really fix the endianess mess of the codec tags...
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32803 a95241bf-73f2-0310-859d-f6bbb57e9c96
last FFmpeg update. This let's WMV videos play again. Tested on GCC4 and
GGC2 builds.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32802 a95241bf-73f2-0310-859d-f6bbb57e9c96
- Removed bitmap icons.
- Updated rdef copyrights.
- Added them to the image.
I also tested them.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32799 a95241bf-73f2-0310-859d-f6bbb57e9c96
- fix error handling in CamDevice.cpp,
- implemented PowerOnSensor() for Sonix,
- add a ProbeByIICSignature method to CamSensor, that takes a list of regs and values to match, as some Linux drivers do,
- add (unfinished, mostly copied from TAS5110C1B) support for TAS5130D1B sensor,
- add (unfinished) support for PAS106 sensor, probe by IIC signature as the Linux driver does,
- add probing by IIC signature for the HV7131E1 sensor as the Linux driver does,
- use -lusb for BeOS (actually ZETA) builds, don't try to use the Haiku USB Kit,
- for the TAS5110C1B sensor, use remove code dup and call SetVideoFrame(), tweaked the values to match those sniffed from the XP driver,
- disabled gain control for now on sonix, doesn't seem to work on all models.
This gets another sonix-based webcam working, partially at least, so if you want to get one to try until UVC works, try this one:
http://www.macally.com/EN/product/ArticleShow.asp?ArticleID=119
We should at least have one webcam supported in the alpha :D
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32795 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Desktop::{Move|Resize}WindowBy() could be called with zeros in which case it
doesn't have to do anything.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32775 a95241bf-73f2-0310-859d-f6bbb57e9c96
BDirectWindow::SetFullScreen() before the window is shown.
The app_server didn't like this, since in that case, Window::Screen() is
NULL. I added a check to ServerWindow::_ResizeToFullScreen() to handle
that case. +alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32774 a95241bf-73f2-0310-859d-f6bbb57e9c96
but this fixes the GCC4 build again, without breaking the function.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32772 a95241bf-73f2-0310-859d-f6bbb57e9c96
unflattening if they included any target or reply info. Changing the reader
usage like this makes the first argument a pointer to a buffer. It would
therefore write to random locations, because it would use the literal value of
those header fields as pointers, crashing the using application.
Will check for GCC4 compliance next.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32771 a95241bf-73f2-0310-859d-f6bbb57e9c96
If you get an "error waiting for interrupt" during boot, try to disabled IDE DMA.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32769 a95241bf-73f2-0310-859d-f6bbb57e9c96
of calloc, since the memory is memcpy()'d in the next line. Added a TODO
comment. +alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32767 a95241bf-73f2-0310-859d-f6bbb57e9c96
Seems like we didn't remember to change it back, even
with the #warning. :-)
I tested ftp and it works fine. This also closes ticket #4176.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32766 a95241bf-73f2-0310-859d-f6bbb57e9c96
BGLView header. This lets us get rid of ugly casting in the implementation.
Also changed a calloc() to malloc(), since the memory is memcpy()'d right
afterwards.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32763 a95241bf-73f2-0310-859d-f6bbb57e9c96
workspace activation message was asynchronous, whether or not the
BWindowScreen stopped drawing in time was pure luck (this also caused crashes
with the VESA driver, as that one unmaps its frame buffer during mode switch).
Introduced a new AS_DIRECT_SCREEN_LOCK protocol for this.
* In the long term, we should let BWindowScreen use the same mechanism as
BDirectWindows, though.
* Removed superfluous locking in BDirectWindow::_InitData().
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32746 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Fixed copyrights (puncuation and capitalization, removed authors from
headers)
* Updated indentation style
* Unified pointer/reference style
* Re-ordered some methods for better grouping where it could be done
(abd adopted source accordingly)
* Small coding style fixes here and there
No functional change intended.
+alphabranch
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32745 a95241bf-73f2-0310-859d-f6bbb57e9c96
longer hold the window lock. There is now a lock that guards screen changes
in particular. This fixes the deadlocks seen in apps using BDirectWindow.
* All direct window handling now sits in the Desktop class -
ServerWindow::HandleDirectConnection() is never called from anywhere else
anymore. Furthermore, it's now only called when actually needed.
* Resize/move actions now always send a B_CLIPPING_MODIFIED flag, too.
* When the screen changed, the driver state is supposed to be B_MODE_CHANGED,
not B_SCREEN_CHANGED (which is a message constant).
* Direct windows are no longer suspended too late on screen changes.
* Removed unused members of DirectWindowData, and cleaned it up a bit.
* Made MultiLocker's default, and copy constructors private - I accidently
used them, causing the ASSERT_MULTI_*LOCKED() macros to fail.
* Added Unlock() to AutoWriteLocker as well.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32742 a95241bf-73f2-0310-859d-f6bbb57e9c96
