Opened 9 months ago

Last modified 8 months ago

#13377 new Bug Report - General

Enable libass subtitle support in android

Reported by: mspieth Owned by: Mark Spieth
Priority: minor Milestone: needs_triage
Component: Ports - Android Version: Unspecified
Severity: medium Keywords:
Cc: Ticket locked: no

Description

Enable libass subtitle support in android

This requires some extra libraries and build script updates.

Attachments (3)

libass.patch (13.2 KB) - added by mspieth 9 months ago.
Initial build patch for libass support
libass.20190119a.patch (14.7 KB) - added by mspieth 9 months ago.
fontconfig-20190126a.patch (6.2 KB) - added by mspieth 9 months ago.

Download all attachments as: .zip

Change History (6)

Changed 9 months ago by mspieth

Attachment: libass.patch added

Initial build patch for libass support

Changed 9 months ago by mspieth

Attachment: libass.20190119a.patch added

comment:1 Changed 9 months ago by mspieth

I will commit this latest patch after merging to master and after a clean rebuild works. Took me a while to get all the lib deps correct but subtitles now work.

Last edited 9 months ago by mspieth (previous) (diff)

Changed 9 months ago by mspieth

Attachment: fontconfig-20190126a.patch added

comment:2 Changed 9 months ago by mspieth

fast way to update fontconfig patch

./makelibs.sh fontconfig
./mythbuild.sh

Havent tried this with David's temp fontconfig patch in subtitlescreen.cpp. This may need to be disabled. Not sure.

Beware, on the first subtitle, there is a 10 second pause while fontconfig initialises (guess)

But subtitles then work for the duration of the video with the correct font(s).

As always YMMV

comment:3 Changed 8 months ago by mspieth

Committed fontconfig update. Now will only wait the first time subtitles are displayed where it takes 10-20sec to build the fontconfig cache. After this every new use is fast.

To update without a full rebuild,

rm -rf libs64/fontconfig* 
./makelibs.sh fontconfig
./mythbuild.sh

Note The workaround has not yet been removed but I suspect is innoccuous. Not tested though.

Note: See TracTickets for help on using tickets.