The bug report has a WAVE file with zero-sized 'data' chunk, which causes
TagLib to iterate over the file, 8 bytes in each iteration. The new code
adds a check for the chunk name, which forces it to mark the file as
invalid if the chunk name doesn't contain ASCII characters.
https://bugs.kde.org/show_bug.cgi?id=283412
Addded fromDict() function to ID3v2Tag. Added fromDict() and
toDict() functions to the TagUnion class (uses the first non-empty tag).
Added fromDict() and toDict() functions for the generic Tag class, only
handling common tags without duplicates. Addded preliminary mp3 test
case. Python3 bindings now available on my github site.
IT: reading was off starting with global volume because of wrong read size.
S3M+IT: correctly read the number of used patterns.
IT: fixed reading of message tag and implemented writing of message tag
(not tested yet).
I also added S3M+IT test files. TODO: Unit tests using them.
If the last chunk had an odd size, the new chunk would have been written at
odd position, which is incorrect.
This is based on the patch by Jens Dyffort, but I ended up changing the
implementation to correctly handle subsequential updates to the file.
The whole RIFF code really needs to be rewritten in a different way...
BUG:243954
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1220223 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
When the WavPack's total_samples header fiels contains -1, try to find
the final block and get the number of samples from there as
block_index + block_samples.
BUG:258016
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1201476 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
Patch by Alex Novichkov, slightly modified by me (code formatting + tests).
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1145554 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
We already do this for TDAT. Using both parts, the year and the date,
would be better we the code currently doesn't have enough context to
do that.
BUG:228968
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1110552 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
Prevously TagLib saved tags to the first block, but read them from the
last one. Having multiple VC blocks is a non-standard situation, but
this is the best we can do (libFLAC also uses the first block in the
case of multiple VC blocks).
BUG:211089
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1043985 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
This is based on patches by Marc Halbruegge, but those only deal with
read-only cases. The code now also correctly adds padding to RIFF
chunks, and calculates offsets in chunkData taking the padding into
account.
BUG:171957
BUG:175781
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1003745 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
We still can't handle actual 64-bit atoms, but we can handle 32-bit sizes
stored in 64 bits.
CCBUG:198730
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@1001897 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
Scott: If you think this is really a bad idea, please revert.
git-svn-id: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/taglib@883108 283d02a7-25f6-0310-bc7c-ecb5cbfe19da
