Use boost::filesystem::path for all paths, and std::string for all other
strings, converting to/from wxString as close to the actual uses of wx
as possible.
Where possible, replace the uses of non-UI wxWidgets functionality with
the additions to the standard library in C++11, or the equivalents in
boost.
Move the path token management logic to libaegisub (and rewrite it in
the process).
Add a basic thread pool based on asio and std::thread to libaegisub.
This touches nearly every file in the project and a nontrivial amount of
code had to be rewritten entirely, so there's probably a lot of broken
stuff.
100 no-op non-amend commits on a subtitle file with 6689 dialogue lines,
with the undo limit set to 100:
Without flyweight:
No video open:
Initial memory usage: 30.6 MB
Final memory usage: 498.0 MB
Elapsed time: 6.3 seconds
Video open, using libass:
Initial memory usage: 54.3 MB
Final memory usage: 653.3 MB
Elapsed time: 23.7 seconds
With flyweight:
No video open:
Initial memory usage: 26.0 MB
Final memory usage: 104.5 MB
Elapsed time: 3.0 seconds
Video open, using libass:
Initial memory usage: 46.7 MB
Final memory usage: 251.8 MB
Elapsed time: 13.0 seconds
No video open:
Memory usage: -79%
Time: -52%
Video open:
Memory usage: -61.5%
Time: -45%
100 no-op amend commits on a line in the middle of a subtitle file with
6689 dialogue lines, with video open:
Without flyweight:
Initial memory usage: 48.2 MB
Final memory usage: 182.3 MB
Elapsed time: 22.3 seconds
With flyweight:
Initial memory usage: 39.8 MB
Final memory usage: 165.8 MB
Elapsed time: 13.8 seconds
Note: The large jump in memory usage here is due to that the benchmark
is blocking the main thread, so at the end there are ~100 video frames
waiting to be displayed.
Move SMPTE handling to agi::vfr::Framerate to get all of the interesting
logic dealing with timcodes in one place, and to make it testable.
Completely rewrite the SMPTE time conversions as testing them reveals
that they were incorrect in some cases.
Originally committed to SVN as r6631.
All of the subtitle formats also already do charset detection when
needed, so doing it in it AssFile::Load is pointless and results in
performing the auto-detection even when it's pointless (because the
format isn't text-based or always uses a single encoding).
Originally committed to SVN as r6459.
Calling AssFile::Save/Load from multiple threads (even on different
objects) was not safe due to that is uses SubtitleFormat internally,
which was inheriently thread-unsafe. To fix this, change
SubtitleFormat's interface to support immutable implementations, and
make all of the current implementations immutable.
This isn't a perfect solution - making a subtitle format implemented in
lua immutable would be rather difficult - so at some point in the future
SubtitleFormat should probably be changed to a factory which returns new
objects from GetReader/GetWriter.
Originally committed to SVN as r6365.
Document all of the SubtitleFormat methods.
Add default implementations of CanReadFile and CanWriteFile that check
against the appropriate wildcard list.
Clean up and simplify a lot of very odd code.
Throw typed exceptions in all subtitle readers rather than strings.
Originally committed to SVN as r5617.
Kill vfr.h and vfr.cpp and use the libaegisub versions of them instead.
Rather than the globals VFR_Input and VFR_Output, everything related to
frame rate is now part of the video context. Most things which used to
use VFR_Output now call VideoContext::TimeAtFrame etc.; video providers,
rather than modifying VFR_Input directly, now have getters for their
frame rates which VideoContext calls. Read-only public access to
VFR_Input and VFR_Output are still provided (hopefully temporarily) for
a few things which were awkward to do through VideoContext.
The Avisynth provider now might correctly handle VFR MKVs which can be
opened with DirectShowSource but not DSS2.
Rework keyframe handling as well, so that it continues to match the vfr
handling in design and implementation.
Originally committed to SVN as r4662.
Rather than going through a ton of work to ensure that every AssEvent
has a sensible start time that happens to leave it in the right place
after sorting, simply sort only consecutive blocks of AssDialogues and
leave everything else untouched. Note that this results in different
behavior when there are multiple [Events] sections in a file -- rather
than moving lines between sections and keeping the number of lines per
section constant, each section keeps the lines it had.
As it's now easy to do, also add sorting by style name and end time.
Closes#614.
Originally committed to SVN as r4307.
Document all functions, class, struct, union, enum, macro, variable, typedefs. This isn't the actual document in itself but empty documentation using any old documentation if it was there.
This was done using exuberant ctags to get tag info, then a TCL script to parse/remove old comments and convert them into Doxygen-style.
Some notes:
* Anything labeled 'DOCME' needs to be documented, @param and @return have been left blank as it would be annoying to delete the 'DOCME' from every one of those.
* Some multiline comments may have been munged into single line comments
* Leave the /// comments above global variables with a space, if they're harder to read then we'll be less likey to use them.
* Enum comments can go after the enumeration itself '[value] /// comment'
* include/aegisub/*.h haven't been converted yet, this will be done in a later commit
* Some documentation blocks are in the wrong place, in the .h when it should be in the .cpp, or vice versa.
See http://devel.aegisub.org/wiki/Doxygen for some details on Doxygen and a 'style guide'.
Originally committed to SVN as r3312.
* Swap old email + website address with 'Aegisub Project http://www.aegisub.org/'
* Set categories for all files (jfs)
* Add descriptions for each file (jfs)
* Add $Id$ keyword
Originally committed to SVN as r3310.
- 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.
