Tomas Reporter
	Description • 17 years ago

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a9pre) Gecko/2007092223 Minefield/3.0a9pre
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a9pre) Gecko/2007092223 Minefield/3.0a9pre

Page http://www.one.lt doesn't load with any build of FF3. Well, indeed it does: I can see main page for a few second, but then it redirects my to page, saying, that page can't be loaded. There is no problem using FF2 or any other browser.

Reproducible: Always

Steps to Reproduce:
1. Go to http://www.one.lt
Actual Results:  
Image http://c19.one.lt/res/images/labels/tmp_close2_lt.gif is showed up in spite of main page.

Expected Results:  
To see main page of http://www.one.lt and be able to browse it.

	Tomas Reporter
	Comment 1 • 17 years ago

Just noticed, that same thing happens with page http://www.videogaga.lt. Both of these pages are build be same developers and, indeed, I think, that it's done pretty bad. So, the problem is bad code of these pages, but on the other side, both of them work OK with FF2 and all other browsers.

	Peter van der Woude [:Peter6]
	Comment 2 • 17 years ago

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a9pre) Gecko/2007092216 Minefield/3.0a9pre ID:2007092216

Confirmed, works in IE7 and Opera 9.22

Martijn, can you figure out what's going on here ?
I see a truckload of warnings in the console but none seem to be a reason for the page not properly loading.

	Tomas Reporter
	Comment 3 • 17 years ago

My guess is that problem is connected to CSS of these pages. Firebug shows this errors while loading http://videogaga.lt (while loading http://one.lt I see many more similar errors):

Expected declaration but found '/'. Skipped to next declaration.
[Break on this error] undefined
main.css (line 493)
Error in parsing value for property 'filter'. Declaration dropped.
[Break on this error] undefined
main.css (line 494)
Expected declaration but found '/'. Skipped to next declaration.
[Break on this error] undefined
main.css (line 501)
Error in parsing value for property 'filter'. Declaration dropped.
[Break on this error] undefined
main.css (line 502)
Selector expected. Ruleset ignored due to bad selector.
[Break on this error] undefined
main.css (line 905)

	Steve England [:stevee]
	Comment 4 • 17 years ago

User Agent sniffing?

	Tomas Reporter
	Comment 5 • 17 years ago

Don't think so. Firstly, I don't see point of developers of these sites to block FF3. Especially, not Firefox in general, but just Firefox 3. Secondly, I've tried "User agent switch" (or something like that) extension, to change user agent. It didn't change a thing about loading these sites.

	Martijn Wargers (dead)
	Comment 6 • 17 years ago

I get a regression range between 2007-07-08 and 2007-07-09:
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2007-07-08+04&maxdate=2007-07-09+07&cvsroot=%2Fcvsroot
I guess a regression from bug 373231, bug 382947 or bug 372969.

	Peter van der Woude [:Peter6]
	Comment 7 • 17 years ago

I'll narrow it down later today

	Peter van der Woude [:Peter6]
	Comment 8 • 17 years ago

regressionwindow
works in 20070708_1921_firefox-3.0a7pre.en-US.win32
fails in 20070708_1936_firefox-3.0a7pre.en-US.win32

http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=PhoenixTinderbox&branch=HEAD&branchtype=match&filetype=match&whotype=match&sortby=Date&hours=2&date=explicit&mindate=1183947660&maxdate=1183948559&cvsroot=%2Fcvsroot

-> Bug 382947

	Martijn Wargers (dead)
	Comment 9 • 17 years ago

As a sidenote, see this page of why it's so difficult to get a testcase out of the page:
http://my.opera.com/hallvors/blog/2007/03/27/how-to-say-hello-world-with-javascript

	Boris Zbarsky [:bzbarsky] Assignee
	Comment 10 • 17 years ago

Has anyone contacted the site yet?

	Damon Sicore (:damons)
	Comment 11 • 17 years ago

OK.  Let's get to the bottom of this.  +'ing.  Works fine in Fx 2 and Safari.  Moving to DOM.

	Tomas Reporter
	Comment 12 • 17 years ago

So it's going to be fixed?

	Steve England [:stevee]
	Comment 13 • 17 years ago

Tomas the problem is going to be investigated. That is no guarantee it will be fixed, but one would hope so.

	Boris Zbarsky [:bzbarsky] Assignee
	Comment 14 • 17 years ago

OK, so the issue here is that we used to send:

  Content-Type: text/plain; charset=utf-8

Now we send:

  Content-Type: text/plain; charset=UTF-8

and this seems to make the site blow up.  Looks a lot like bug 398280 indeed.

	Boris Zbarsky [:bzbarsky] Assignee
	Comment 15 • 17 years ago

I think the right fix here is to leave the existing charset if it matches the one we want to set (up to case).  I'll probably roll that change into the fix for bug 393968.

	Boris Zbarsky [:bzbarsky] Assignee
	Comment 16 • 17 years ago

Fixed by checkin for bug 393968.  Checked in a test too.

	Julian Reschke
	Comment 17 • 13 years ago

See tests in <http://lists.w3.org/Archives/Public/public-webapps/2011OctDec/0955.html>. It appears that, as of 2011, only FF tries to preserve the charset casing (where Opera and IE preserve it because they do not fix the charset at all).

	Boris Zbarsky [:bzbarsky] Assignee
	Comment 18 • 13 years ago

Yes, but the upshot is that for pages that set the charset the three have compat.

If course you can try backing this patch out and seeing whether this site breaks, as well as the site in bug 398280.

	Julian Reschke
	Comment 19 • 12 years ago

See related XHR issue: https://www.w3.org/Bugs/Public/show_bug.cgi?id=16302

	Julian Reschke
	Comment 20 • 12 years ago

(In reply to Julian Reschke from comment #19)
> See related XHR issue: https://www.w3.org/Bugs/Public/show_bug.cgi?id=16302

Sorry, wrong link.

https://www.w3.org/Bugs/Public/show_bug.cgi?id=16303