[Webkit-unassigned] [Bug 116672] [WPE][GTK] WebKit2.WebContext register_uri_scheme does segfault

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=116672

--- Comment #5 from Philip Chimento <philip.chimento at gmail.com> ---
(In reply to Michael Catanzaro from comment #4)
> Ideally we would try to find some way to ensure that pygobject and
> gjs shut down last of all. Normally the way to do that is to register your
> own exit handler first. pygobject and gjs should be able to win and run
> first, because they execute code before WebKit ever does.

It's not an option for GJS to do that unless we allow WebKit's bad practices to leak into GJS, and it would also break the expectations of GJS library users that they should be able to destroy their JS context at a time of their choosing.

I understand the problem, but I think should be libraries' responsibility to at least make it possible to opt out of running weird code in static destructors or atexit() handlers.

How about API to unregister the URI scheme handlers early?

Alternatively we could advise PyGObject and GJS users to use their own WebContext rather than the default one. I seem to remember I tested that at the time and it crashed, but maybe that has been fixed in the meantime.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20181113/b9e78be2/attachment.html>