cache new pointers to dircache items when dircache goes
offline and comes back onlineagain (during tagcache commit). This
should prevent wrong filenames to appear in playlist.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30047 a1c6a512-1295-4272-9138-f99709370657
Only integer IDs are exposed from dircache with this. This way the cache is isolated from other modules.
This is needed for my buflib gsoc project.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30038 a1c6a512-1295-4272-9138-f99709370657
It's the correct type and should save some memory due to struct padding (on eabi).
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30027 a1c6a512-1295-4272-9138-f99709370657
characters in a tag could cause the parser fail to import a track
statistics correctly. Now line feeds are escaped properly and carriage
returns ignored on import.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30021 a1c6a512-1295-4272-9138-f99709370657
must be evaluated as MP3. Also fix supported formats: hwcodec doesn't support MP1.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30019 a1c6a512-1295-4272-9138-f99709370657
memory. One of these headers (current_tcmh) was not loaded in when
tagcache state was initialized from a hibernated state file (flashed
H1xx targets). That caused internal serial number to start always from
zero, rendering "recently played tracks" query not working as expected.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30016 a1c6a512-1295-4272-9138-f99709370657
Translators need to verify the voice string, dumbly copied from the dest string
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30007 a1c6a512-1295-4272-9138-f99709370657
The problem r26713 addressed was not caused by missing yields but by too high
priority of the background loader thread. Undoing the yield() change and decreasing
the priority fixes the audio drop as well but doesn't cause jerkiness.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@30000 a1c6a512-1295-4272-9138-f99709370657
* There's no need to reset the buffer argument to
check_against_clauses to 0 when it's not needed. We can always
initialize str = buf and only change it on a RAM search hit.
* Do not memset buffer to 0 -- it's sufficient to make sure the
retrieved tag string is zero-terminated.
* Factor out a call to check_virtual_tags from two branches of an if
statement.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29981 a1c6a512-1295-4272-9138-f99709370657
used for nonnumeric tags as well.
Optimization for the nonnumeric case: Only numerics are ever updated,
so there's no need to scan the command queue for updates to nonnumeric
tags.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29979 a1c6a512-1295-4272-9138-f99709370657
Such targets would previously default to using the configuration for targets with small iram which uses an extra memcpy per block.
This saves 2MHz decoding a 128kbps vorbis file on the Gigabeat S and saves a bit of codec buffer.
Patch from FS#11268, also replaces patch from FS#12147.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29976 a1c6a512-1295-4272-9138-f99709370657
Some of these were found with http://www.samba.org/junkcode/#findstatic. Changes of note:
* The old MDCT has been removed.
* Makefile.test files that create test programs for libatrac, libcook, and libffmpegFLAC have been removed, as they don't work. My project will have a replacement that works with all codecs.
* I've tried not to remove anything useful. CLIP_TO_15 was removed from libtremor because there's another copy (also commented) in codeclib.
git-svn-id: svn://svn.rockbox.org/rockbox/trunk@29945 a1c6a512-1295-4272-9138-f99709370657