Bug 218213 – REGRESSION: MobileMiniBrowser frequently fails to repaint

• Status: NEW
• Component: Layout and Rendering

[Kenneth Russell]

After syncing my WebKit iOS Simulator Debug build to top-of-tree (it had previously been ~1 month out of date), loading pages in WK1 mode in the MobileMiniBrowser frequently results in the browser failing to repaint.

Clicking the "reload" button makes the page render correctly.

Scrolling doesn't fix the problem; the last page's content (blank, upon the initial launch and load of https://webkit.org/ ) is scrolled.

Control-clicking to try to select text usually provokes a repaint.

This mode (WK1 in the iOS simulator) is the easiest mode for contributors to test. Is there any chance of help tracking down the cause of this regression?

[Alexey Proskuryakov]

Which version of iOS are you using? 

We switched from iOS 13 to iOS 14 over this month, so iOS 13 is technically unsupported in trunk (although I don't think that we made any breaking changes already). But what this means is that we don't have continuous archives for bisection.

[Simon Fraser (smfr)]

Could you test around r268075?

[Radar WebKit Bug Importer]

<rdar://problem/70733416>

[Kenneth Russell]

iOS 13.5 in the iPad Pro 9.7" simulator.

smfr@ thanks for the pointer. A checkout at r268074 doesn't have the problem. I'll test at r268075 next.

[Kenneth Russell]

MobileMiniBrowser seems to repaint reliably at r268075 on the iPad Pro (9.7-inch), iOS 13.5 simulator.

Building WebKit to git bisect is quite painful on my development machine (a laptop), so appreciate any help tracking down the cause. Thanks.

[Alexey Proskuryakov]

The bots already switched to iOS 14 at this point. 

You can download built archives using URLs like <https://s3-us-west-2.amazonaws.com/minified-archives.webkit.org/ios-simulator-14-x86_64-release/268075.zip> if you would like to further bisect.