* In karaoke mode, instead of scrolling start to end of active line into view, scroll start of selection to end of line into view. This gives sensible results when horizontal zoom is great enough that the entire line isn't visible at once and the syllable is too far from the start of the line. This fixes#677.
* If the line is too long to fit in the display, always left-align it instead of sometimes left-aligning and sometimes centering. (I think the old behaviour depended on the actual scroll position.)
Originally committed to SVN as r2928.
- Changed the SubtitleFormat::AskForFPS dialog box; removed the "PAL/NTSC only" choice and added a "show SMPTE dropframe" parameter instead. Also added 50fps as a choice.
- While I was at it, reworked the TranStation export filter so it actually looks ahead to see if the next line will overlap with the current, and if so, move the end time of the current line backwards one frame, which fixes#767
Originally committed to SVN as r2920.
- Variable loops are controlled using the tenv.maxloop(newmax) function to set the loop counter max. This is a wrapper around setting tenv.maxj.
- Add tenv.relayer(newlayer) function for changing the Layer of an output line.
- Add tenv.restyle(newstyle) function for changing the Style of an output line, despite this usually breaking positioning/sizing information.
- tenv.maxloops is an alias for tenv.maxloop.
- tenv.loopctl(newj, newmaxj) controls both loop variables if anyone wants to do that.
This closes#753 but documentation still needs to be written.
Originally committed to SVN as r2913.
- When default buttons were present, pushing OK caused no button code to be returned at all.
- A load of controls not handling default/initial values properly, closes#775 and #776
Originally committed to SVN as r2912.
libass: Fix cache lookup problem causing memory bloat
The cache code did hash lookups by storing key values in struct fields
and then hashing and comparing the struct as a single memory block. In
at least one case such a struct contained uninitialized padding bytes
which prevented the complete memory area of the struct from matching
even though the fields did. As a result the code failed to find
existing objects in the cache and stored new versions of them, causing
gigabytes of memory use in some circumstances. Initializing the struct
memory to zero before writing the fields avoided such memory use in
tests but is not guaranteed to work if I interpret the C standard
correctly (the compiler is allowed to write garbage over padding bytes
when changing struct member values).
Change the code to use struct-specific hashing and comparison
functions that work field by field to guarantee correct behavior.
Create these by replacing the struct definition with a template that
lists the fields and can be used the generate each of struct
definition, hash function and compare function with some preprocessor
magic (otherwise every field would need to be listed separately in all
three).
Originally committed to SVN as r2895.
Fix an old bug that would drop the first or second video frame
Greatly improved audio decoding by adding a small internal cache
Originally committed to SVN as r2894.
By default the cache is kept below 42MB and 20 files; to change the defaults change "FFmpegSource max cache size" and "FFmpegSource max cache files" in config.dat. Either or both may be set to 0, but the cleaner will not delete the last index file in the directory, regardless of cache settings and of the file's size.
Originally committed to SVN as r2893.
