Bug 247823 – macOS Ventura: SafariForWebKitDevelopment will launch but nothing works (network,history,etc)

• Component: WebKit2
• Status: RESOLVED FIXED

[@TimGMichaud]

Created attachment 463495 [details]
A screenshot of the Safari start page, with google.com typed into the address bar, but the site is not loading

Current HEAD is 10178633f995c1b1d12184d2b73592838f0d18d7 (From Webkit's github clone)

I have tried building and running Webkit after building with --release, --release --asan, and --debug via the ./Tools/Scripts/build-webkit script. When I run, I also use the ./Tools/Scripts/run-safari script

The main browser will open, and will look like the attached screenshot. 

I suspect this is a permissions issue, but my Console app isn't showing much beyond being denied access to the accountsd daemon, bookmarks sync agent is missing and entitlement, and what I think might be a pointer to the root cause: "Could not enable Mach bootstrap, errno = 1" (This is spammed 5-6 times when I try to go to google.com)

[Sam Sneddon [:gsnedders]]

What hardware are you running this on? This could be a duplicate of bug 236829.

[@TimGMichaud]

MBP w/ an intel i7 chip 

* OS version: Ventura 13.1 beta (22c5033e)
* Kernel version: 22.2.0
System firmware version: 1916.40.8.0.0

[Alexey Proskuryakov]

Hmm, I don't even get this far. 

- With SIP enabled, SafariForWebKitDevelopment uses system frameworks (known issue).
- With SIP disabled, it crashes when creating the first WebContent process (need to look into why).

I used an archive uploaded from CI, but those should not be any different from locally built ones. This is on Apple Silicon with macOX Ventura 13.0.1.

[Alexey Proskuryakov]

Looks like the crash is happening with minified archives, full archives have the same problem as originally reported here.

[Alexey Proskuryakov]

I have a local build on a customer OS now, and tried to figure out what's wrong.

Basic process infrastructure appears to work fine - processes are launched, something is happening in them, judging by the amount of seemingly unimportant `log stream` output. Haven't been able to pinpoint where is breaks apart yet.

[Alexey Proskuryakov]

FWIW, *history* appears to work for me. Loading pages doesn't, whether from address bar of from start page, or from history.

[Radar WebKit Bug Importer]

<rdar://problem/102339404>

[Alex Christensen]

*** Bug 247995 has been marked as a duplicate of this bug. ***

[Alex Christensen]

This was fixed in https://bugs.webkit.org/show_bug.cgi?id=247995

[@TimGMichaud]

Thanks ya'll :)! 

Confirming it works on my end too

[Alex Christensen]

Good to hear! Thanks for confirming!