* Fix .talk clips on hwcodec. Voice does have the entire audio buffer available there.
* Get rid of the separate TALK_PROGRESSIVE_LOAD in favour of the more advanced
TALK_PARTIAL_LOAD i.e. use the latter on the Ondios as well. This gets rid of quite
some ifdefing, and has the advantage that the voice file can be larger than the buffer
(at a slight binsize cost).
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30916 a1c6a512-1295-4272-9138-f99709370657
if the voice is bigger than the audiobuffer. NOTE: This is the case on the sim
so voice doesn't appear to work currently on hwcodec. Someone needs to verify
on a real target.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30910 a1c6a512-1295-4272-9138-f99709370657
- move some stuff out of manual-prep and improve dependencies.
- always update version information, version.sh is run anyway.
- use $(SILENT) in more places to allow showing the commands.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30902 a1c6a512-1295-4272-9138-f99709370657
As discussed on IRC the cyrillic characters can cause various problems and not
all fonts include them so stick with extended latin.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30901 a1c6a512-1295-4272-9138-f99709370657
Very frequent start-stop cycles (as caused by frequent core_alloc() calls)
of audio makes the codecs lose the resume position, and this causes playback
from the beginning.
To work around, use queue_post() instead of queue_send() to delay the resume
so that it only resumes once per core_alloc() set.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30900 a1c6a512-1295-4272-9138-f99709370657
audio playback. If it goes below 256K new buflib allocations fail.
This prevents buffer underruns as the new buffer size wasn't actually
checked at all.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30893 a1c6a512-1295-4272-9138-f99709370657
- When trying to retrieve the human readable device name on target mismatch
detection it is not possible to figure the "extra" variant internally used
from the information in rockbox-info.txt. Since the h10_5gb didn't have a
non-extra version resolving the name would fail. Remove the extra part for
the UMS version so that will get used. It's better to display a slightly
wrong name instead of none.
- Add some comments describing the sections and their values.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30887 a1c6a512-1295-4272-9138-f99709370657
Instead of showing the target name as stored in rockbox-info.txt retrieve the full descriptive name.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30886 a1c6a512-1295-4272-9138-f99709370657
The buildserver_modelname is still in use, and the commit missed some parts on
removal, causing segfaults on start.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30884 a1c6a512-1295-4272-9138-f99709370657
Since changing the manual URL creation to be template based this value isn't
used anymore. It's also pretty much the same as configure_modelname, so remove
it completely.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30869 a1c6a512-1295-4272-9138-f99709370657
When building the linked libraries pass the machine flag as well. For now add
them to the compiler variable directly, should go via CFLAGS later.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30868 a1c6a512-1295-4272-9138-f99709370657
The current implementation fails silently if Windows Media Player is version
10. Add a check and inform the user if the version installed is too old to work
properly with beastpatcher.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30864 a1c6a512-1295-4272-9138-f99709370657
