Bug 193309

Summary: Safari Crashing in Version 12.0.1 (14606.2.104.1.1) WebCore::GraphicsLayerCA::updateBackdropFilters
Product: WebKit Reporter: Dean Jackson <dino>
Component: New BugsAssignee: Dean Jackson <dino>
Status: RESOLVED FIXED    
Severity: Normal CC: ews-watchlist, graouts, simon.fraser, webkit-bug-importer
Priority: P2 Keywords: InRadar
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   
Attachments:
Description Flags
Patch graouts: review+

Description Dean Jackson 2019-01-09 16:33:55 PST
Safari Crashing in Version 12.0.1 (14606.2.104.1.1) WebCore::GraphicsLayerCA::updateBackdropFilters
Comment 1 Dean Jackson 2019-01-09 16:34:20 PST
<rdar://problem/45279224>
Comment 2 Dean Jackson 2019-01-09 16:40:42 PST
Created attachment 358765 [details]
Patch
Comment 3 Antoine Quint 2019-01-10 02:31:52 PST
Comment on attachment 358765 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=358765&action=review

> Source/WebCore/platform/graphics/ca/GraphicsLayerCA.cpp:2157
> +        if (backdropFilterRect.width() > 0 && backdropFilterRect.height() > 0) {

Should `canHaveBackdropFilters` be set to true in case either the width or height is 0? Also, do we need the `> 0` check? I assume these values can't be negative.
Comment 4 Dean Jackson 2019-01-10 09:26:45 PST
Comment on attachment 358765 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=358765&action=review

>> Source/WebCore/platform/graphics/ca/GraphicsLayerCA.cpp:2157
>> +        if (backdropFilterRect.width() > 0 && backdropFilterRect.height() > 0) {
> 
> Should `canHaveBackdropFilters` be set to true in case either the width or height is 0? Also, do we need the `> 0` check? I assume these values can't be negative.

If either the width or height is 0, then there is no point having a backdrop layer.

Regarding negative, I don't think they ever should be < 0, but it is technically possible... and one of the ways the crash could have happened. I'm hoping to get a reproduction from an originator to see what really went wrong.
Comment 5 Dean Jackson 2019-01-10 10:09:29 PST
Committed r239833: <https://trac.webkit.org/changeset/239833>