<html>
<head>
<base href="https://bugs.webkit.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - [Soup] Deadlock in NetworkProcess"
href="https://bugs.webkit.org/show_bug.cgi?id=167876#c3">Comment # 3</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - [Soup] Deadlock in NetworkProcess"
href="https://bugs.webkit.org/show_bug.cgi?id=167876">bug 167876</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=167876#c2">comment #2</a>)
<span class="quote">> According to myself in <a class="bz_bug_link
bz_status_RESOLVED bz_closed"
title="RESOLVED FIXED - [SOUP] Allow sending URI request data in chunks"
href="show_bug.cgi?id=85880#c9">https://bugs.webkit.org/show_bug.cgi?id=85880#c9</a> the
> mutex is needed, because glib will use a thread to implement read_async, and
> webkitSoupRequestInputStreamAddData can be called from other thread. Before
> r210374, that thread was the message work queue, that's why this never
> happened. Now it's the main thread, so I guess we just need to release the
> lock.</span >
But I was wrong, because we are overriding GInputStreamClass::read_async, claiming we know how to handle the async request, and therefore not using the fallback implementation in GInputStream that runs the thread. It makes sense, indeed, because we are actually a GMemoryInputStream, there's no io in the read function, so we can handle it in the main thread.</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>