| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
Remove unused x_odd parameter from blend_???_exact functions
|
|
|
|
once.
This way, it avoids to calculate the multiplication in the for loops and in the
memset() call.
|
|
by not exporting them.
|
|
|
|
This fixes reported alignment issues on ARM.
(We could require correct alignment on some architectures, but this is easier.)
|
|
- validate palette alpha values in overlay manager
(one check / overlay / palette index) instead of
checking every alpha value twice for every
blended pixel in every frame
- remove unneeded calculations
- approximiate expensive integer divisions with
multiplication and shift
|
|
|
|
|
|
redrawing.
While splitting my original big patch set, I've lost three lines.
This change will make xine-lib compile again.
|
|
The original number of OSD objects was 5 which served xine-lib's
needs. VDR may on it's own allocate up to 16 OSD objects and thus
there may be no more OSD objects left for xine-lib's needs.
The change addresses this by increasing the current number by 5.
|
|
These functions are likely to be removed later when more correct
solutions have been found.
|
|
|
|
supply decoded frames.
There can still be scheduling delays which may let the number of
frames ready for displaying to drop below frame drop limit just
for a short period of time.
Therefore the changes remember that the decoder should have been
asked to drop some frames but do not actually have the decoder to
drop some frames. When the situation has improved at the next time
when the check is performed, the remembered frame drop is canceled
or otherwise (when the number of frames is still below frame drop
limit) executed.
|
|
allocated frames.
The current code uses a hard coded frame drop limit of 3 and
doesn't adhere to it's documentation when testing whether frames
shall be dropped. As a result frame drop limit is actually 4,
which means that the decoder is asked to drop some frames when
the number of frames waiting for displaying is less then 4.
Consider a video out device like xxmc which only supplies 8
frames. For MPEG2 decoding, two frames will be used by the
decoder (for the current frame and the forward reference frame)
and two further frames will be used in the video out loop (the
current and the previous frame) so that at any given time (under
perfect conditions) there will be 4 frames waiting to be displayed.
But when there are delays in scheduling, it might happen that
there are only 3 frames ready for displaying and thus will result
in asking the decoder to drop frames.
The changes therefore determine the maximum frame drop limit in
dependence of the number of allocated frames and make the
detection work like documented. In the above scenario, the maximum
number actually used for frame drop limit will then be 2 which
allows to compensate some scheduling delays without causing the
decoder to drop frames.
|
|
available.
Usually it's a good idea to avoid reallocating frames especially
when a deinterlacer needs a different format than the decoder, as
this would then happen all the time.
But when there is only a limited number of frames available, then
even a single frame which is not scheduled at frame allocation may
let the number of frames ready for displaying drop below frame
drop limit and thus resulting in unnecessary frame drops.
|
|
drops.
When a video out device provides only a little number of video
frames, the video decoder should be scheduled immediately to
provide a decoded frame as soon as possible. Otherwise, the
number of available frames for displaying may go below frame
drop limit and thus resulting in unnecessary frame drops.
|
|
redrawing.
The video out loop sleeps up to 20 ms (and the paused loop 20 ms)
which means that pending OSD events are delayed too from beeing
processed. When an OSD is used for example to scroll through a
list of VDR recordings, this delay may slow down scrolling
unnecessarily.
Especially when the OSD manager is able to render the OSD content
indepently from drawing a frame to screen, this change will allow
the fastest OSD update possible.
|
|
This function shall be used to poll the number of outstanding OSD
events from a certain point in time on until the reported number
is 0. At that point in time, the content on screen is identical to
a certain state of the stream, at which for example, a hardcopy may
be taken.
|
|
elements by the size of the single element.
|
|
memcpy structures.
When array of constant pointers are used for register enum configurations,
this creates more warnings because of pointer mismatches; I'd consider
casting them, but not yet.
In the memcpy_method array, mark the parts that are constant at build time
as const so to try reducing the overhead.
|
|
|
|
used in a few more cases.
|
|
|
|
than in the root of it.
|
|
With this commit, also fix the previous mistakes in the iteration.
|
|
With this change, when rendering a font through FreeType2 (and not using
FontConfig) the directories looked up are the ones defined in XDG_DATA_DIRS,
right after XDG_DATA_HOME; this way the user can decide to have fonts data
in a different place than its home directory.
This also splits up the lookup for fontconfig and non-fontconfig cases in
two different functions, to avoid gotos and labels.
|
|
|
|
the check manually.
|
|
This handle can then be used by all plugins and other parts of the
xine engine without having to duplicate it.
|
|
Directory Specification.
With this change, xine-lib starts abiding to the XDG Base Directory
Specification, allowing the user to define a different path to save
its cache data (by setting XDG_CACHE_HOME environment variable).
|
|
The introduced functions give "frontend like" plugins a chance to
lock and unlock frontend_lock. This protects such threads for
example from beeing blocked when changing the streams speed.
|
|
The introduced function give "frontend like" plugins a chance to
lock and unlock port rewiring. This protects such threads (when
combined with holding the frontend lock) from beeing blocked when
calling functions like xine_get_current_frame().
|
|
|
|
The idea is to allow only a "single" frontend to rewire ports at
a certain point in time. Regarding a stream, frontend_lock is used
for example to allow only a single frontend to change the speed.
Unfortunately, frontend_lock cannot be used as the rewire functions
are not stream related.
Therefore a new port_rewiring_lock was introduced and used at
appropriate locations. When an arbitrary thread now holds the
frontend_lock and the port_rewiring_lock, it is safe that acquiring
a port ticket in functions like xine_get_current_frame() will
never block the thread.
|
|
This function shall be used to poll the number of remaining frames
from a certain point in time on until the reported numbers are all
0. At that point in time, the content on screen is identical to a
certain state of the stream, at which for example, a hardcopy may
be taken.
|
|
The current code has a race condition which can block arbitrary
threads that call for example xine_get_current_frame() until the
stream gets unpaused again. This can happen when the internal
ticket acquiration collides with a ticket revokation for example
when another thread is going to pause the stream.
There are a few situations where a port ticket needs to be
acquired for calling a port function but where it is absolutely
undesireable to get blocked for an undetermined period of time.
Therefore the ticket system should be extended by nonblocking
functions which allow ticket acquiration even when a ticket
revokation is in progress. And in the case where blocking is
not avoidable, it should simply be indicated that no ticket was
acquired. The caller can then choose to repeat the call at a
later point in time.
|
|
The _x_config_init() and _x_config_change_opt() functions should stay internal
and not be exported by the library, so remove XINE_PROTECTED to hide them.
|
|
extend the documentation available before.
|
|
First of all, as we check the presence of the attribute in configure.ac and
we generate the XINE_PACKED definition there, there's no need to do the work
again on attributes.h.
Also remove #pragma pack support for now, as the trick is just not neat enough,
and as it is it's just noise.
Make Wine, libdvdnav and nosefart use XINE_PACKED rather than doing their own
checks, as they are less reliable.
|
|
bswap.h already.
(transplanted from a7820d16324dbc6fd6b35481a78ff532f59ebe71)
--HG--
extra : transplant_source : %A7%82%0D%162M%BCo%D6%B3T%81%A7%8F%F52%F5%9E%BEq
|
|
bswap.h already.
|
|
|
|
|
|
Two of the modified files are headers, but each contains definitions as well as
declarations and is only ever used once.
|
|
|
|
to be video_overlay.
|
