Closed
Bug 1310676
Opened 8 years ago
Closed 8 years ago
another rgb->hsl mistranslation
Categories
(DevTools :: Framework, enhancement)
DevTools
Framework
Tracking
(Not tracked)
RESOLVED
WONTFIX
People
(Reporter: tromey, Assigned: tromey)
References
Details
https://bugzilla.mozilla.org/show_bug.cgi?id=1302787#c28 notes that our JS code doesn't convert some hsl value (unfortunately you'll have to dig into the tests to see which one) to rgb differently than platform. Not sure how serious this is but perhaps it bears investigation.
Updated•8 years ago
|
Flags: qe-verify-
Priority: -- → P3
Whiteboard: [devtools-html] → [reserve-html]
Assignee | ||
Comment 1•8 years ago
|
||
This shows the failure: diff --git a/devtools/client/shared/test/unit/test_cssColor-03.js b/devtools/client/shared/test/unit/test_cssColor-03.js index 0657f74..1b002c3 100644 --- a/devtools/client/shared/test/unit/test_cssColor-03.js +++ b/devtools/client/shared/test/unit/test_cssColor-03.js @@ -24,6 +24,7 @@ const OLD_STYLE_TESTS = [ "hsl(120, 100%, 40%)", "hsla(120, 100%, 40%, 0.25)", "hSlA(240, 100%, 50%, 0.25)", + "hsl(500, 5%, 5%)", ]; const CSS_COLOR_4_TESTS = [ 0:00.55 TEST_STATUS: Thread-3 run_test FAIL [run_test : 48] color hsl(500, 5%, 5%) matches DOMUtils - {"r":12,"g":13,"b":13,"a":1} deepEqual {"a":1,"b":12,"g":13,"r":12}
Assignee | ||
Updated•8 years ago
|
Assignee: nobody → ttromey
Status: NEW → ASSIGNED
Updated•8 years ago
|
Iteration: --- → 52.3 - Nov 7
Priority: P3 → P1
Assignee | ||
Comment 2•8 years ago
|
||
I've looked into this a bit. I can fix this particular bug by changing _hslValue to round a bit differently -- using roundTo(..., 6) worked for me. However, that then regresses test_cssColor-02.js, because it breaks round-tripping from rgb->hsl->rgb. I don't think this is a regression from any earlier state (and I think regressions are what we're primarily concerned with here); because CSS defines hsl->rgb but not the reverse. Maybe this round-tripping could be fixed by preserving more digits in the rgb->hsl translation, but then we're getting into "ugly display" territory. Also, in an earlier bug we discussed but rejected the idea of having the CssColor object maintain state, on the basis that while this would avoid this kind of problem, it would introduce a new user-visible problem, namely incorrect round-tripping when moving between elements. So on this basis I'm going to WONTFIX this.
Status: ASSIGNED → RESOLVED
Closed: 8 years ago
Resolution: --- → WONTFIX
Updated•8 years ago
|
No longer blocks: devtools-html-3
Iteration: 52.3 - Nov 7 → ---
Priority: P1 → --
Whiteboard: [reserve-html]
Updated•6 years ago
|
Product: Firefox → DevTools
You need to log in
before you can comment on or make changes to this bug.
Description
•