Bug 210283

Summary:

Reset view navigation gesture state between tests

Product:

WebKit

Reporter:

Simon Fraser (smfr) <simon.fraser>

Component:

New Bugs

Assignee:

Simon Fraser (smfr) <simon.fraser>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

commit-queue, ddkilzer, simon.fraser, thorton, webkit-bug-importer

Priority:

Keywords:

InRadar

Version:

WebKit Nightly Build

Hardware:

Unspecified

OS:

Unspecified

See Also:

https://bugs.webkit.org/show_bug.cgi?id=210433

Bug Depends on:

210330

Bug Blocks:

Attachments:

Description	Flags
Patch	thorton: review+
Patch	none
Patch	none

Description Simon Fraser (smfr) 2020-04-09 09:39:52 PDT

Reset view navigation gesture state between tests

Comment 1 Simon Fraser (smfr) 2020-04-09 09:43:12 PDT

Created attachment 395966 [details]
Patch

Comment 2 Tim Horton 2020-04-09 10:20:21 PDT

Comment on attachment 395966 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=395966&action=review

> Source/WebKit/UIProcess/API/C/WKPage.cpp:2728
> +{

If you put this on WKWebView instead you can avoid the pageclient hits AND avoid expanding the C API.

Comment 3 Simon Fraser (smfr) 2020-04-09 10:35:04 PDT

Created attachment 395975 [details]
Patch

Comment 4 Radar WebKit Bug Importer 2020-04-09 10:50:01 PDT

<rdar://problem/61526278>

Comment 5 Simon Fraser (smfr) 2020-04-09 10:50:18 PDT

(In reply to Tim Horton from comment #2)
> Comment on attachment 395966 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=395966&action=review
> 
> > Source/WebKit/UIProcess/API/C/WKPage.cpp:2728
> > +{
> 
> If you put this on WKWebView instead you can avoid the pageclient hits AND
> avoid expanding the C API.

Yeah I wasn't sure about that. Will do.

Comment 6 Simon Fraser (smfr) 2020-04-09 11:17:48 PDT

Created attachment 395979 [details]
Patch

Comment 7 EWS 2020-04-09 13:11:29 PDT

Committed r259818: <https://trac.webkit.org/changeset/259818>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 395979 [details].

Comment 8 David Kilzer (:ddkilzer) 2020-04-10 03:10:22 PDT

This is causing crashes on three ios-wk2 tests in EWS and on build.webkit.org:

swipe/main-frame-pinning-requirement.html
swipe/pushState-programmatic-back-while-swiping-crash.html
webgl/1.0.3/conformance/ogles/GL/lessThan/lessThan_001_to_008.html

Example test run:  <https://build.webkit.org/results/Apple%20iOS%2013%20Simulator%20Release%20WK2%20(Tests)/r259848%20(3630)/results.html>

Comment 9 David Kilzer (:ddkilzer) 2020-04-10 03:15:58 PDT

See also results:

<https://results.webkit.org/?suite=layout-tests&suite=layout-tests&suite=layout-tests&test=swipe%2FpushState-programmatic-back-while-swiping-crash.html&test=webgl%2F1.0.3%2Fconformance%2Fogles%2FGL%2Ffaceforward%2Ffaceforward_001_to_006.html&test=swipe%2FpushState-programmatic-back-while-swiping-crash.html&platform=ios>

Comment 10 WebKit Commit Bot 2020-04-10 03:16:36 PDT

Re-opened since this is blocked by bug 210330

Comment 11 David Kilzer (:ddkilzer) 2020-04-10 08:36:11 PDT

(In reply to David Kilzer (:ddkilzer) from comment #8)
> This is causing crashes on three ios-wk2 tests in EWS and on
> build.webkit.org:
> 
> swipe/main-frame-pinning-requirement.html
> swipe/pushState-programmatic-back-while-swiping-crash.html
> webgl/1.0.3/conformance/ogles/GL/lessThan/lessThan_001_to_008.html
> 
> Example test run: 
> <https://build.webkit.org/results/
> Apple%20iOS%2013%20Simulator%20Release%20WK2%20(Tests)/r259848%20(3630)/
> results.html>

BTW, this looks kind of like a use after free based on the crash logs.

Comment 12 Simon Fraser (smfr) 2020-04-10 10:11:36 PDT

iOS WK2 EWS is so slow that I can't wait for it.

Comment 13 Simon Fraser (smfr) 2020-04-10 11:18:21 PDT

https://trac.webkit.org/changeset/259883/webkit