Bug 194787 – Spew: Unhandled web process message 'VisitedLinkTableController:VisitedLinkStateChanged'

• Component: History
• Status: RESOLVED FIXED

[Joseph Pecoraro]

When navigating in a Debug build I see lot of spew in the Console:

    ERROR: Unhandled web process message 'VisitedLinkTableController:VisitedLinkStateChanged'
    Source/WebKit/WebProcess/WebProcess.cpp(740) : virtual void WebKit::WebProcess::didReceiveMessage(IPC::Connection &, IPC::Decoder &)

Steps to Reproduce:
1. Open about:blank
2. Navigate to cnn.com
  => Spew in console about unhandled message

[Radar WebKit Bug Importer]

<rdar://problem/48175520>

[Chris Dumez]

Likely a PSON regression, I'll investigate as soon as I have spare cycles.

[Chris Dumez]

The WebProcess does not have a VisitedLinkTableController, likely because it does not have any pages any more. Before PSON, we use to terminate such processes. However, with PSON, we may keep processes without pages alive due to:
- Suspended pages
- WebProcess cache
- Process pre-warming

This basically mean we're sending VisitedLinkTableController:VisitedLinkStateChanged IPC to WebProcesses who do not care about such notifications. I do not believe this is actually harmful but the error logging is indeed unfortunate.

Currently, WebProcessProxy::shutDown() is in charge of unregistering the process from the visitedLinkStores. I guess we have to do so when the process no longer has any WebPage.

[Chris Dumez]

Created attachment 365086 [details]
Patch

[Geoffrey Garen]

Comment on attachment 365086 [details]
Patch

r=me

[Chris Dumez]

Comment on attachment 365086 [details]
Patch

Clearing flags on attachment: 365086

Committed r243142: <https://trac.webkit.org/changeset/243142>

[Chris Dumez]

All reviewed patches have been landed.  Closing bug.

[Chris Dumez]

Reverted r243142 for reason:

Caused assertion hits in WK2 Debug

Committed r243154: <https://trac.webkit.org/changeset/243154>

[Chris Dumez]

Comment on attachment 365086 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=365086&action=review

> Source/WebKit/UIProcess/WebProcessProxy.cpp:404
> +    auto users = m_visitedLinkStoresWithUsers.ensure(&visitedLinkStore, [] {

this should be auto& XD

[Chris Dumez]

Created attachment 365201 [details]
Patch

[Chris Dumez]

Committed r243159: <https://trac.webkit.org/changeset/243159>

[Truitt Savell]

It looks like the changes in https://trac.webkit.org/changeset/243159/webkit

broke 16 API tests on Debug.

Build:
https://build.webkit.org/builders/Apple%20Mojave%20Debug%20WK2%20%28Tests%29/builds/2006

I was able to reproduce and regress this down to 243159 locally.

[Chris Dumez]

(In reply to Truitt Savell from comment #12)
> It looks like the changes in https://trac.webkit.org/changeset/243159/webkit
> 
> broke 16 API tests on Debug.
> 
> Build:
> https://build.webkit.org/builders/Apple%20Mojave%20Debug%20WK2%20%28Tests%29/
> builds/2006
> 
> I was able to reproduce and regress this down to 243159 locally.

I can work on fixing these first thing tomorrow morning. If this is not soon enough, feel free to revert my change.

[Chris Dumez]

(In reply to Chris Dumez from comment #13)
> (In reply to Truitt Savell from comment #12)
> > It looks like the changes in https://trac.webkit.org/changeset/243159/webkit
> > 
> > broke 16 API tests on Debug.
> > 
> > Build:
> > https://build.webkit.org/builders/Apple%20Mojave%20Debug%20WK2%20%28Tests%29/
> > builds/2006
> > 
> > I was able to reproduce and regress this down to 243159 locally.
> 
> I can work on fixing these first thing tomorrow morning. If this is not soon
> enough, feel free to revert my change.

Looking at these now.

[Chris Dumez]

(In reply to Chris Dumez from comment #14)
> (In reply to Chris Dumez from comment #13)
> > (In reply to Truitt Savell from comment #12)
> > > It looks like the changes in https://trac.webkit.org/changeset/243159/webkit
> > > 
> > > broke 16 API tests on Debug.
> > > 
> > > Build:
> > > https://build.webkit.org/builders/Apple%20Mojave%20Debug%20WK2%20%28Tests%29/
> > > builds/2006
> > > 
> > > I was able to reproduce and regress this down to 243159 locally.
> > 
> > I can work on fixing these first thing tomorrow morning. If this is not soon
> > enough, feel free to revert my change.
> 
> Looking at these now.

Follow-up fix landed in <https://trac.webkit.org/changeset/243202>.