<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - [GTK] Use GTestDBus instead of dbus-launch in WebKitTestBus.cpp"
href="https://bugs.webkit.org/show_bug.cgi?id=161481#c17">Comment # 17</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - [GTK] Use GTestDBus instead of dbus-launch in WebKitTestBus.cpp"
href="https://bugs.webkit.org/show_bug.cgi?id=161481">bug 161481</a>
from <span class="vcard"><a class="email" href="mailto:cgarcia@igalia.com" title="Carlos Garcia Campos <cgarcia@igalia.com>"> <span class="fn">Carlos Garcia Campos</span></a>
</span></b>
<pre>(In reply to <a href="show_bug.cgi?id=161481#c14">comment #14</a>)
<span class="quote">> (In reply to <a href="show_bug.cgi?id=161481#c11">comment #11</a>)
> > (In reply to <a href="show_bug.cgi?id=161481#c9">comment #9</a>)
> > > While the patch from <a class="bz_bug_link
bz_status_NEW "
title="NEW - [GTK] Use dbus-run-session when possible to avoid left-over dbus-daemon instances"
href="show_bug.cgi?id=161135">bug 161135</a> applies along with this one, it does disable
> > > this one which I think is undesirable. The result of having both applied
> > > would be dbus-run-session be used every time it is available. This solution
> > > is a better one, so I'd recommend just going with it, if we can fix the
> > > threading concerns.
> >
> > I'm not sure this patch solves the same problem. The bus started by
> > WebKitTestBus::run() will be leaked if the test crashes, because
> > ~WebKitTestBus() won't be called in that case. So, we can leave this code as
> > fallback when dbus-run-session is not present, or when tests are run
> > manually (without using run-gtk-tests), something that I do a lot. When
> > dbus-run-session is present we simply set m_address, I don't think the
> > patches are mutually exclusive.
>
> I thought GTestDBus handled cleaning up the dbus daemon process even on
> crashes? But if you think it's useful to have both and ok to have
> dbus-run-session used when present I'll go ahead and land it =)</span >
You are right, it uses a global watcher that forks a child that watches the parent and kills spawned processes in that case. So, I agree this solution is better, because it also works when tests are run directly without using run-gtk-tests</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the assignee for the bug.</li>
</ul>
</body>
</html>