Bug 190202 – REGRESSION(r236154) Page loads are not drawn when loaded in shipped Safari11/12

• Component: Page Loading
• Status: RESOLVED FIXED

[youenn fablet]

After this revision, page loads are hanging on previous Safari version.
This was reported for Safari high Sierra and Safari Mojave.
I bissected the issue with WebKit Nightlies to  http://trac.webkit.org/r236153

[Rob Buis]

I have this problem on High Sierra and I am seeing:
./Tools/Scripts/run-safari --no-saved-state
Starting SafariForWebKitDevelopment with DYLD_FRAMEWORK_PATH set to point to built WebKit in /Users/rwlbuis/WebKit/WebKitBuild/Release.
Failed to acquire exclusive access to AutoFill corrections SQLite store at AutoFillCorrections.db.
Failed to acquire exclusive access to AutoFill corrections SQLite store at CloudAutoFillCorrections.db.
Failed to acquire exclusive access to AutoFill corrections SQLite store at AutoFillCorrections.db.
Failed to acquire exclusive access to AutoFill corrections SQLite store at CloudAutoFillCorrections.db.
2018-10-02 17:55:14.088 SafariForWebKitDevelopment[50142:289352] It's not legal to call -layoutSubtreeIfNeeded on a view which is already being laid out.  If you are implementing the view's -layout method, you can call -[super layout] instead. Break on void _NSDetectedLayoutRecursion(void) to debug.  This will be logged only once.  This may break in the future.
Failed to acquire exclusive access to AutoFill corrections SQLite store at AutoFillCorrections.db.
Failed to acquire exclusive access to AutoFill corrections SQLite store at CloudAutoFillCorrections.db.
Failed to acquire exclusive access to AutoFill corrections SQLite store at AutoFillCorrections.db.
Failed to acquire exclusive access to AutoFill corrections SQLite store at CloudAutoFillCorrections.db.
2018-10-02 17:55:14.509 SafariForWebKitDevelopment[50142:289430] AOSKit ERROR: XPC CLIENT: Unauthorized client error received

[Rob Buis]

Also on my system are stock Safari Version 11.1.2 (13605.3.8) and Xcode Version 9.4.1 (9F2000).

[Alex Christensen]

Interesting.  Can this reproduce with the latest?

[youenn fablet]

(In reply to Alex Christensen from comment #3)
> Interesting.  Can this reproduce with the latest?

It was reported in Mojave latest Safari.

[Alex Christensen]

I mean with the latest WebKit.
Also, can it reproduce with local builds of WebKit?

[Chris Dumez]

(In reply to Alex Christensen from comment #5)
> I mean with the latest WebKit.
> Also, can it reproduce with local builds of WebKit?

This is the bug, it happens if you use a recent locally built WebKit with the System's Safari.

[Chris Dumez]

Does it require building WebKit with an external SDK?

[Rob Buis]

(In reply to Alex Christensen from comment #5)
> I mean with the latest WebKit.
> Also, can it reproduce with local builds of WebKit?

Yes, I built r236716 today for my test.

[Alex Christensen]

Created attachment 351698 [details]
Patch

[youenn fablet]

Comment on attachment 351698 [details]
Patch

Are we missing some API tests to catch these kind of regressions?

[Alex Christensen]

(In reply to youenn fablet from comment #10)
> Comment on attachment 351698 [details]
> Patch
> 
> Are we missing some API tests to catch these kind of regressions?
Technically yes, but we are in the process of removing all this SPI.

http://trac.webkit.org/r236886

[Radar WebKit Bug Importer]

<rdar://problem/45055789>

[Chris Dumez]

Comment on attachment 351698 [details]
Patch

Seems like we could have very easily written an API test for this.