* maintain a list of all BPictures to do so
* BView downloads the BPicture data after recording the picture. This could probably done more efficiently using shared memory in the first place.
* Rename init_sse to init_fpu and handle FPU setup.
* Stop trying to set up FPU before VM init.
We tried to set up the FPU before VM init, then
set it up again after VM init with SSE extensions,
this caused SSE and MMX applications to crash.
* Be more logical in FPU setup by detecting CPU flag prior
to enabling FPU. (it's unlikely Haiku will run on
a processor without a fpu... but lets be consistant)
* SSE2 gcc code now runs (faster even) without GPF
* tqh confirms his previously crashing mmx code now works
* The non-SSE FPU enable after VM init needs tested!
* Move compiled binaries into lib.haiku within Mesa optional
pacakge, this enables us to use Make or scons to build Mesa
* Untangle libGL linking and order library inclusion
* Remove zmuldef hacks
* Compile libgallium into the libGL on Mesa 8.0 / gcc4 images
* Don't Mesa calculate cpu optimizations due to visibility
issues, use Mesa cpu functions for this.
* Don't allow undefined symbols in libGL.so to ensure sanity.
* Move Mesa optional packages to haiku-files
* Aleksas Pantechovskis added into contributors list for his work on
setmime and writembr tools, and many patches provided for Mail
application during completing about 12 GCI tasks;
* Peter Poláčik added into contributors list for his work on adding
localization support into Jamfile engine and documenting
using of makefile/jamfile engines;
* Mariya Pilipchuk added into translators list for her work on Ukrainian
translation;
Thank you for the contribution!
* Append the signature only after the e-mail Reader thread will finish
it's work. This prevent from inserting the signature in wrong place;
* Added extra lines between original e-mail quote and the signature for
more comfortable editing of replies.
* This fix was made during GCI2011;
* This fixes#4590.
Signed-off-by: Siarzhuk Zharski <zharik@gmx.li>
* Fixing gcc2 build by removing exceptions using, that are looking like
not supported by our version of compiler.
* Thanks to Alexander von Gluck for pointing the problem out.
* Implemented writembr tool used to update MBR area with easy boot
loader starting the system from currently active partition;
* This work fixes#5595;
* This is result of GCI2011 task.
Signed-off-by: Siarzhuk Zharski <zharik@gmx.li>
* Introduce the "MAIL:read_pos" attribute of e-mail file node to
store the latest scroll position of the mail text view;
* Fixes#4302 (Mail should store and restore scroll offset);
* This feature was implemented during completing GCI2011 task.
Signed-off-by: Siarzhuk Zharski <zharik@gmx.li>
* Rename _SetSpan to _SetupRenderBuffer to perform more general
render buffer setup and consolidate code
* Remove CMAP8 calls. Mesa no longer supports CMAP8. CMAP8 GL
use cases are non-existent
* We just wait a little longer in this case, like 4.5s.
* Changed the code to only send the messages when needed, before they were
being sent every 1.5s.
* However, this is untested as of yet, as the layout changes didn't allow me
to run the MediaPlayer on my system. Will do so in a minute on another one.
* Call different functions depending on Mesa version
* The --no-warnings is for the GL headers which produce
large numbers of attribute directive ignored warnings
that cannot be disabled on gcc2
* User collected dictionaries and indexes are moved under
~/config/data/spell_check/ folder. Fixes#7887.
* This fix was made during completing GCI 2011 task.
Signed-off-by: Siarzhuk Zharski <zharik@gmx.li>
* Allow displaying of "Read" button only for Incoming e-mails.
Fixes#4773;
* Move to the next message after pressing "Unread" button that
is consistent with corresponding "Unread" button case.
Fixes#4774;
* Those problems were fixed during completing GCI 2011 task.
Signed-off-by: Siarzhuk Zharski <zharik@gmx.li>
* Check for real e-mail status before activating corresponding
"Close and Leave as <Status>" popup menu entries and modifying them
correspondently;
* That was implemented as GCI 2011 task. Fixes#5251.
Signed-off-by: Siarzhuk Zharski <zharik@gmx.li>
* Some of these changes may be pulled back with the NEW_MESA
define so this works across older and newer Mesa versions
* NEW_MESA defines and checks should be removed when gcc2 is
no longer a supported platform
In Playlist, whenever a move of items occured causing the
currently playing song to change its position, so :
1. Importing files (D&D for example) before its position
2. Removings files before it
3. Moving files before it
was causing the currently playing song to restart because
it was thinking a new entry was asked (it had a different
index number suddently).
Also adjusted the behaviour when you delete the currently
playing track.
Should fix ticket #6689.
* add Wcscoll() and Wcsxfrm() ICU locale backend
* provide implementations of wcscoll() and wcsxfrm() that are using
the respective methods of the locale backend
The actual xfreopen function is only relevant if O_BINARY != 0 and all
places invoking it are surrounded by if statements that are always
false if O_BINARY == 0. The compiler therefore stripped all the calls
out at the default -O2, but this obviously breaks down with DEBUG on.
The variable is used in the TRACE(), but that is only enabled for
DEBUG >= 2. This therefore avoids a warning on DEBUG=1 builds.
Inspired by patch in #8252 by Urias McCullough.
1. When in /boot, you couldn't go to the parent (which is fine by itself),
but the menu item wasn't disabled.
2. In Single Window Browsing, doing "Open parent" wasn't respecting that
setting and was opening new windows.
3. In filepanel, when navigating to Desktop folder, the "Go to parent" was
still proposed even though it would do nothing choosing it.
4. Opening a Pose in Single Window Browsing was repopulating the menus and
putting the shortcuts again. That was an issue if Navigator bar was used
since it conflicts with its shortcuts.
Fixing ticket #6851.
In an earlier commit, I removed a memleak, but it was covering more cases
than intended. Thanks to Alex Wilson's insight, it seems that simply removing
the check is better (since that check is essentially redone in FSMoveToTrash()).
Rather than checking if "window" AND "view" are NULL, check if
"window" OR "view" are NULL, since neither could be null and deferenced later
on.
From the Matrox accelerant, it seems that such event occurs when a
"Workspace switch, screen prefs change, or overlay app shutdown occurs"
Might be interesting to revisit tickets related to intel accelerant.
CID 205.
BVolume::BlockSize() might return a negative value (error code) and it
was passed as parameter of CalcItemsAndSize() as a size_t (unsigned).
Thus the check for an error code failed.
CID 4171.
The HIDReportItem reads 32 bit chunks from the report buffer. To avoid
having to check the remaining buffer space on each extraction, we pad
the buffer so it is always valid to read 32 bits at a time. Also add
a comment explaining why we do it that way.
Thanks to Johannes Anderwald for pointing out the potential out of
bound access!
The dynamic_cast was called twice, and the assert was
done on the first attempt. Now it just cast once, assert on this
and pass it further as before.
CID 3122.
