There are several optimizations here:
* We make font families atoms, to allow for quicker comparisons.
* We precalculate an FNV hash of the relevant fields of the font style
structure.
* When obtaining a platform font group, we first check pointer equality
for the font style. If there's no match, we go to the FNV hash. Only
if both caches miss do we construct and cache a font group. Note that
individual fonts are *also* cached; thus there are two layers of
caching here.
15% improvement in total layout thread time for Facebook Timeline.
r? @glennw (since you last worked on `get_layout_font_group_for_style()` IIRC)
`FontContext::get_layout_font_group_for_style()`.
There are several optimizations here:
* We make font families atoms, to allow for quicker comparisons.
* We precalculate an FNV hash of the relevant fields of the font style
structure.
* When obtaining a platform font group, we first check pointer equality
for the font style. If there's no match, we go to the FNV hash. Only
if both caches miss do we construct and cache a font group. Note that
individual fonts are *also* cached; thus there are two layers of
caching here.
15% improvement in total layout thread time for Facebook Timeline.
This is a series of already-reviewed changes by @yodalee from #5025, rebased onto current servo master, with some fixups applied:
* Fixed build errors/warnings from the latest rust upgrade.
* Marked `tests/wpt/web-platform-tests/encoding/single-byte-decoder.html` as expecting CRASH.
I could not verify locally that the new test annotation is correct, since the test appears to hang rather than crash on my Linux box. (Or maybe I just didn't wait long enough.) If this crash isn't consistent, or if it takes a long time, maybe we should skip this test instead?
r? @jdm
I think this should have been changed in #3618 but was missed. r? @pcwalton
I wasn't able to come up with a good test case for this, partly because of other bugs related to floats and formatting contexts.
In my mistake, I close PR for this commit.
I'm sorry for inconvenience.
@jdm 's comment
The test expectations for /tests/wpt/web-platform-tests/html/dom/interfaces.html and /tests/wpt/web-platform-tests/workers/interfaces.workers.js will need to be updated - http://mxr.mozilla.org/servo/search?string=appversion
According to your guide, I revised commit. Is this your intention?
r? @jdm
cc @yichoi
Transition events are not yet supported, and the only animatable
properties are `top`, `right`, `bottom`, and `left`. However, all other
features of transitions are supported. There are no automated tests at
present because I'm not sure how best to test it, but three manual tests
are included.
r? @glennw
Transition events are not yet supported, and the only animatable
properties are `top`, `right`, `bottom`, and `left`. However, all other
features of transitions are supported. There are no automated tests at
present because I'm not sure how best to test it, but three manual tests
are included.
I decided to use the old Netscape broken image link icon (later we may
replace the image asset for something more trendier). The ref test will
expect that a failed load should display the rippy image.
ImageCacheTask users can define if a placeholder image should be loaded
at start up or not. This enables both the new behavior (e.g. always
return an image even for broken urls) as also the previous one.
