[Webkit-unassigned] [Bug 191858] [GTK] Unexpected User-Agent on redirect

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=191858

--- Comment #25 from Carlos Garcia Campos <cgarcia at igalia.com> ---
(In reply to youenn fablet from comment #22)
> > > What is the impact of clearing user agent for subresource loads?
> > 
> > In the case of subresource loads the FrameLoader computes the user agent
> > again if the header is not present, like for the main resource.
> 
> In general, navigation loads are not handling redirections so it is fine to
> hope that a new user agent will be created.

You mean because of PSON?

> For subresource loads, redirections are expected to be followed and the
> current model is to keep the same user agent in that case (would need to be
> tested though if we do not have such tests).

Testing quirks is not easy, but at least now we can test that user-agent is preserved after redirects.

> It would be interesting to see what would happen for ping loads/fetch keep
> alive for instance with your change.

hmm, I think we could lose the user agent for ping loads.

-- 
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/20200902/3fb0e6a0/attachment.htm>