Bug 195072

Summary:

REGRESSION: Crash beneath ResourceLoadObserver::logSubresourceLoading

Product:

WebKit

Reporter:

Brent Fulgham <bfulgham>

Component:

WebKit Misc.

Assignee:

Brent Fulgham <bfulgham>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

achristensen, bfulgham, cdumez, commit-queue, rniwa, wilander

Priority:

Keywords:

InRadar

Version:

WebKit Nightly Build

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
Patch	none
Patch	none

Description Brent Fulgham 2019-02-26 15:23:30 PST

ResourceLoadObserver and WebProcess should exist as long as the process is running, but the m_networkProcessConnection can be closed for a variety of reasons. If the ResourceLoadObserver is being asked to observe an operation in the WebProcess, we should attempt to reconnect to the NetworkProcess as we do for other message send cases.

Comment 1 Brent Fulgham 2019-02-26 15:31:18 PST

<rdar://problem/47773409>

Comment 2 Brent Fulgham 2019-02-26 15:36:09 PST

Created attachment 363029 [details]
Patch

Comment 3 John Wilander 2019-02-26 15:42:19 PST

Comment on attachment 363029 [details]
Patch

Looks good to me.

Nit: In your Changelog entry, "observing" should be "observes" to make the grammar work.

Comment 4 Brent Fulgham 2019-02-26 15:45:27 PST

(In reply to John Wilander from comment #3)
> Comment on attachment 363029 [details]
> Patch
> 
> Looks good to me.
> 
> Nit: In your Changelog entry, "observing" should be "observes" to make the
> grammar work.

Oh! Good catch. Thanks.

Comment 5 Brent Fulgham 2019-02-26 15:51:26 PST

Created attachment 363030 [details]
Patch

Comment 6 WebKit Commit Bot 2019-02-27 16:57:42 PST

Comment on attachment 363030 [details]
Patch

Clearing flags on attachment: 363030

Committed r242183: <https://trac.webkit.org/changeset/242183>

Comment 7 WebKit Commit Bot 2019-02-27 16:57:44 PST

All reviewed patches have been landed.  Closing bug.