Daniel Alley
Daniel Alley
The system zlib backend was ~15% faster, but this is measuring my whole application including all the XML parsing I'm doing, so the actual differences are probably bigger.
The test is a little flaky when not run in headless mode -- when I just leave it alone, it times out, when I click on the window that is...
@jdm It passes in headless mode consistently.
Looks there are some unknown failures
I'll hold off on changing anything until after the surfmanup PR merges, because that looks like it changes a bunch of the osmesa context code.
@jdm Updated, hopefully this should fix it.
Why would it fail to resize? Should I just recreate the entire ```WebrenderSurfman``` object then?
I'll try to figure it out when I get some spare time (not as much of it at the moment as 2 weeks ago)
When I try to run the tests I get a bunch of ConnectionRefused errors and assertion failures that look like this ``` E AssertionError: assert 'Content-Type' in {'cache-control': 'no-cache', 'content-length':...
@asajeffrey yes, it looks like at some point during the test it is trying to resize to 0x0 ``` 0:05.10 pid:469525 [ports/glutin/headless_window.rs:84] self.inner_size.get() = 800x600 0:05.10 pid:469525 [ports/glutin/headless_window.rs:85] new_size =...