Commit Graph

11 Commits

Author SHA1 Message Date
Thiago Marcos P. Santos
b7e26d1ff2 [metrics] Delete test that doesn't make sense
This test is ignored because we expect it to "ignore but fail".

But it will generate a fatal error because it doesn't have the
metrics file, thus should have been skipped instead.
2019-12-16 22:52:01 +02:00
Thiago Marcos P. Santos
8869d09b25 [build] Collect and publish new baselines
Also fail the bot in case there are new baselines.
2019-12-16 22:52:01 +02:00
Thiago Marcos P. Santos
970cabe0eb [build] Reorganize the render tests
Move everything that might collect metrics to the metrics/ directory,
the only test manifests inside the render-test/ directory are the
unit tests of the probes.
2019-12-16 22:52:01 +02:00
Thiago Marcos P. Santos
71054c8b60 [glfw] Add static render test creator
Press F1 to create a render test for the current view.

The test will be created at the base folder (current if nothing is
provided) + /NNNN (the next free entry).
2019-12-09 16:09:16 +02:00
Alexander Shalamov
1ad1358c8e [metrics] Rebaseline linux-gcc8 mbgl-offline filesize metric 2019-12-02 17:11:49 +02:00
Alexander Shalamov
154d95a626 [metrics] Rebaseline filesize metric for linux-clang8 2019-12-02 17:11:49 +02:00
Thiago Marcos P. Santos
af0d46c367 [android] Use icf=all
Chromium is using it for a while on ARM.

https://codereview.chromium.org/1153143004/patch/20001/30002
2019-11-29 13:42:18 +02:00
Thiago Marcos P. Santos
e3d1daaf66 [metrics] Measure compressed size for Android
Because what metters is the size we add to the APK.
2019-11-28 19:18:17 +02:00
Thiago Marcos P. Santos
316f1d0ad1 [metrics] Add new baselines. 2019-11-28 12:34:22 +02:00
Thiago Marcos P. Santos
d13562252d [metrics] Isolate binary size metrics
Will give us a better view of what changed and by how much. Otherwise
the test stops at the first failure.
2019-11-28 12:34:22 +02:00
Thiago Marcos P. Santos
8d163c159f [metrics] Add binary size metrics
This test will check the size of the tools and shared libraries
built from previous bots in a single run.

We will have +/- 1% acceptable variance before the test fails.
2019-11-27 18:32:58 +02:00