FS#67448 - [mlt] compress lumas.
Attached to Project:
Arch Linux
Opened by bartus (bartus) - Friday, 31 July 2020, 11:51 GMT
Last edited by Antonio Rojas (arojas) - Friday, 31 July 2020, 14:29 GMT
Opened by bartus (bartus) - Friday, 31 July 2020, 11:51 GMT
Last edited by Antonio Rojas (arojas) - Friday, 31 July 2020, 14:29 GMT
|
Details
After switching to cmake `mlt` balloons out substantially
(from `3MiB` to `260MiB`) by generating uncompressed lumas
for for `luma_transition_filter.
[There's no flag in `cmake` to turn off lumas build or to build compressed lumas](https://github.com/mltframework/mlt/commit/b9732158e5c8446d5083adcc5586c6938f8b75a1#diff-cb48c84f4672eeab94c4f8908b424727) (those options exists only in `configure/make` build) I propose using [a bit of code of code form `/mlt/src/module/lumas/create_lumas.sh` to compress lumas from `pgm`s to `png`s](https://github.com/mltframework/mlt/blob/master/src/modules/lumas/create_lumas#L68-L73). This way we go down with `mlt` size form `~260MiB` to `8.7MiB` btw. I've putted this as PR on GitHub, but looks like those don't count. https://github.com/archlinux/svntogit-packages/pull/2 |
This task depends upon
Closed by Antonio Rojas (arojas)
Friday, 31 July 2020, 14:29 GMT
Reason for closing: Fixed
Additional comments about closing: mlt 6.22.0-2
Friday, 31 July 2020, 14:29 GMT
Reason for closing: Fixed
Additional comments about closing: mlt 6.22.0-2