1034974 - Page won’t be drawn until CSS finishes loading

Status: NEW

(Core :: DOM: Core & HTML, defect, P5)

Description

[Simonas Kazlauskas [:nagisa/:simukis]]

If the HTML loads fine, but linked CSS does not, the page will not be drawn until linked CSS finishes (or fails) loading.

I noticed this issue on gmane a few days ago where the page itself loaded very quickly, but loads for CSS were blackholed and I had to for wait TCP connection timeout (i.e. for stylesheet load to fail) to view contents of the page.

I attached a simple HTTP server written in python to which pointing firefox will easily reproduce the issue.

Comment 1

[Simonas Kazlauskas [:nagisa/:simukis]]

Attachment: Simple HTTP server which demonstrates the issue

After launching the server, point your firefox at http://localhost:8080/ or equivalent.

Comment 2

[Matthias Versen [:Matti]]

I think that could be a result of a recent networking change (bug 237623) ?

Comment 3

[Daniel Stenberg [:bagder]]

(In reply to Matthias Versen [:Matti] from comment #2)
> I think that could be a result of a recent networking change (bug 237623) ?

I don't think it sounds like that. The code changes done for Bug 237623 only trigger once the connection is closed, so it won't affect slow or stalled transfers.

If however the connection that delivers the CSS ends prematurely, that would result in an error and would cause the CSS to not be used.

Comment 4

[Boris Zbarsky [:bzbarsky]]

> I think that could be a result of a recent networking change (bug 237623) ?

Do you think that because this is a regression or something?

Because this bug as reported sounds like a duplicate of a very old bug that I can't locate right now about timing out the wait on stylesheets and starting layout anyway even if we have pending stylesheet loads.  We normally wait on CSS to load before starting layout to avoid flashes of unstyled content.

Comment 5

[Matthias Versen [:Matti]]

>Do you think that because this is a regression or something?
Yes, i thought that this sounds like a recent regression but i could be wrong - sorry in that case.

Comment 6

[Simonas Kazlauskas [:nagisa/:simukis]]

I could reproduce in Firefox 4, Firefox 2, Chrome, IE 11 and IE 8, so this smells like expected behaviour… even if not user-friendly one.

Comment 7

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1472046

Move all DOM bugs that haven’t been updated in more than 3 years and has no one currently assigned to P5.

If you have questions, please contact :mdaly.