Bug 198927 – Web Inspector: REGRESSION(r245498): Timelines: CPU: discontinuities are filled in by the next record

• Component: Web Inspector
• Status: RESOLVED FIXED

[Devin Rousso]

Created attachment 372260 [details]
[Image] Screenshot of Issue

# STEPS TO REPRODUCE:
1. inspect any page
2. start a timeline recording
3. stop the timeline recording
4. wait a few seconds (this creates a discontinuity)
5. start a timeline recording
 => the space between the two "ranges" is filled by the next CPU bar that's added

[Devin Rousso]

Created attachment 372280 [details]
Patch

[Matt Baker]

Comment on attachment 372280 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=372280&action=review

r=me, very nice!

> Source/WebInspectorUI/UserInterface/Models/TimelineRecording.js:197
> +        if (!isNaN(this._discontinuityStartTime)) {

Nit: Above (175) the test is for `this._discontinuityStartTime` being truthy, instead of !isNaN.

> LayoutTests/inspector/unit-tests/set-utilities.html:17
> +            InspectorTest.expectTrue(set.has(key), "Set has `key`.");

I'm sure this is here to make the pre-take state more obvious, but I think it's unnecessary. The key was just added, and we aren't testing `has`.

> LayoutTests/inspector/unit-tests/set-utilities.html:21
> +            InspectorTest.expectFalse(set.take("doesNotExistKey"), "Set cannot take `doesNotExistKey`.");

Nit: `nonExistentKey` sounds better, IMO.

[Devin Rousso]

Created attachment 373287 [details]
Patch

[WebKit Commit Bot]

Comment on attachment 373287 [details]
Patch

Clearing flags on attachment: 373287

Committed r247033: <https://trac.webkit.org/changeset/247033>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/52495442>