Bug 227514
| Summary: | [GLIB] regression(r278304): fast/css/parsing-relative-color-syntax.html fails on GTK and WPE | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Arcady Goldmints-Orlov <crzwdjk> |
| Component: | Layout and Rendering | Assignee: | Nobody <webkit-unassigned> |
| Status: | NEW | ||
| Severity: | Normal | CC: | bfulgham, darin, sam, simon.fraser, zalan |
| Priority: | P2 | ||
| Version: | WebKit Nightly Build | ||
| Hardware: | Unspecified | ||
| OS: | Linux | ||
Arcady Goldmints-Orlov
This test has started failing since r278304, in particular one assertion in the test (added in the aforementioned change) fails as follows:
FAIL computedStyle("background-color", "lch(from peru calc(l * 0.8) c h)") should be lch(49.80138% 54.003296 63.680317). Was lch(49.80138% 54.003284 63.68035).
Which seems like a precision issue somewhere (perhaps the test demanding too much precision here?)
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Sam Weinig
Interesting. I wonder where things could be differing from glib to the cocoa platforms here, as I would kind of expect double / float arithmetic to be the same. Perhaps compiler flags or something.
But yeah, making these tests a little more lenient somehow makes sense to me.