gecko-dev/gfx/wr/wrench
Dzmitry Malyshau 5e158608bc Bug 1528674 - WR improve picture texture cache allocaton r=gw
The change contains a number of incremental improvements with the main goal of:
  - allocating exactly as many tile as required by the app
  - respecting the picture caching option

Differential Revision: https://phabricator.services.mozilla.com/D24740

--HG--
extra : moz-landing-system : lando
2019-03-26 21:18:09 +00:00
..
benchmarks Bug 1528157 - Optimize GPU time for clip mask generation. r=kvark 2019-02-19 20:52:27 +00:00
examples Bug 1525741 - Support animating yaml files in wrench. r=kvark 2019-02-07 18:03:59 +00:00
reftests Bug 1538626 - Support fast path for clip-out rectangles in WR. r=emilio 2019-03-25 08:35:31 +00:00
res
script Bug 1533282 - Restore the debugging options in wrench/script/headless.py. r=kvark,kats 2019-03-08 11:15:32 +01:00
src Bug 1528674 - WR improve picture texture cache allocaton r=gw 2019-03-26 21:18:09 +00:00
.gitignore
Cargo.toml Bug 1527325 - Don't re-export contents of WR API units r=gw 2019-03-14 01:44:05 +00:00
README.md
android.txt Bug 1531524 - Support running wrench on Android. r=kvark 2019-03-04 18:49:52 +00:00
build.rs

README.md

wrench

wrench is a tool for debugging webrender outside of a browser engine.

headless

wrench has an optional headless mode for use in continuous integration. To run in headless mode, instead of using cargo run -- args, use ./headless.py args.

replay and show

Binary recordings can be generated by webrender and replayed with wrench replay. Enable binary recording in RendererOptions.

RendererOptions {
    ...
    recorder: Some(Box::new(BinaryRecorder::new("wr-frame.bin"))),
    ...
}

If you are working on gecko integration you can enable recording in webrender_bindings/src/bindings.rs by setting

static ENABLE_RECORDING: bool = true;

wrench replay --save yaml will convert the recording into frames described in yaml. Frames can then be replayed with wrench show.

reftest

Wrench also has a reftest system for catching regressions.

  • To run all reftests, run script/headless.py reftest
  • To run specific reftests, run script/headless.py reftest path/to/test/or/dir
  • To examine test failures, use the reftest analyzer
  • To add a new reftest, create an example frame and a reference frame in reftests/ and then add an entry to reftests/reftest.list