368909 - Reloading a loading XML page causes a well-formedness error to appear briefly

Status: REOPENED

(Core :: XML, defect, P3)

Description

[Jesse Ruderman]

Steps to reproduce:
1. Start loading http://www.hixie.ch/tests/evil/page-loading/incremental/001.cgi?mime=text%2Fxml&delay=1&repeats=10
2. Press Cmd+R (reload) or click the home button.

Result: "no element found" XML error

Expected: cancel the load peacefully, like when I press Esc (stop).

Comment 1

[Mikko Rantalainen]

Seeing this issue with
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a7pre) Gecko/2007070804 Minefield/3.0a7pre
(Hit CTRL+R while loading the url in the bug, well-formedness error is displayed for one second - that is, before the server sends start of the content).

I agree with Jesse, the display should look like the same as hitting the Esc.

Hardware/OS should be set to ALL/ALL.

Comment 2

[AndrewM]

I tried several times to reproduce this, but couldn't. Is it still an issue?

Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9a9pre) Gecko/2007102705 Minefield/3.0a9pre ID:2007102705

Comment 3

[Jesse Ruderman]

The URL isn't displaying incrementally at all for me now :(

Comment 4

[Johnny Stenback (:jst)]

Is this a dupe of bug 401613? If not, can you please have a look at this one, Peter?

Comment 6

[Jesse Ruderman]

The testcase doesn't work at all now, probably because of a problem on Hixie's server, so this is currently impossible to test.

Sicking and jst, why do you think this is a dup of bug 401613?

Comment 7

[Jesse Ruderman]

Apparently Hixie's server has decided to gzip the output of his script :/

Comment 9

[Boris Zbarsky [:bzbarsky]]

This isn't a duplicate.  The issue here is that we stop the old load and start the new one.  But stopping the old load cancels the parser partway, so the XML content sink never gets the closing tags it's expecting and shows the yellow screen of death.  All this happens before the new load gets a response, so we get a brief flash of YSOD for the page we're _leaving_.

Not sure whether the XML sink can usefully detect this situation.

Comment 10

[Mikko Rantalainen]

(In reply to comment #9)
> The issue here is that we stop the old load and start
> the new one.  But stopping the old load cancels the parser partway, so the XML
> content sink never gets the closing tags it's expecting and shows the yellow
> screen of death.

How is hitting the stop button handled? As far as I know, no yellow screen of death appears if I hit the stop button during the loading (perhaps I've been just lucky?). Doesn't the parser report (to XML sink) that it got killed partway?

Comment 11

[Damon Sicore (:damons)]

Wouldn't hold the release for this.  Marking wanted1.9.0.x.  If people feel otherwise, please re-nom.  The XML still loads, albeit it is still ugly for a second.

Comment 12

[Boris Zbarsky [:bzbarsky]]

> How is hitting the stop button handled?

Same exact way.  Same exact error code.  Of course, I can't reproduce this bug as originally filed, nor with a large XML file loaded from a file:// URI.  Can you?

Comment 13

[Johnny Stenback (:jst)]

If someone can still reproduce this, please tell us how and renominate accordingly.

Comment 14

[Vlad Alexander]

>If someone can still reproduce this
Yes, please see this screen shot:

http://misc.xstandard.com/mozilla/parsing-error.gif

... it was taken using build:

Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9pre) Gecko/2008052206 Minefield/3.0pre

> renominate accordingly
Sorry, don't know how to do this.

> please tell us how
This happens over a slow Internet connection. Since we don't have access to a slow Internet connection any longer, I reproduced the problem using a proxy server set to simulate modem speed. Here are the steps:

1. Onto your local computer, install proxy server called Fiddler2 from:
http://fiddler2.com/dl/Fiddler2Setup.exe

2. Run Fiddler2. From the Rules menu, select:

Performance > Simulate Modem speeds

and

Performance > Disable Caching

3. In Firefox, set the manual proxy connection to localhost on port 8888.

4. Go to xhtml.com or xstandard.com and click around.

Comment 15

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Renominating for dot-release

Comment 16

[Vlad Alexander]

RE: Comment #14, last point should read:

4. Go to xstandard.com and click around.

Comment 17

[Daniel Veditz [:dveditz]]

nominating for 3.1 in case Jonas still cares about it. Clearing "wanted1.9.0.x" because we don't care enough to push for a fix, but if it does get fixed on trunk we'll look at a small, safe patch.

Comment 20

[Stefan]

(In reply to Jesse Ruderman from comment #6)
> The testcase doesn't work at all now, probably because of a problem on
> Hixie's server, so this is currently impossible to test.

Find a new testcase here: http://www.vogtner.de/mozilla/801416
Surprisingly the YSsOD in most cases contain no text.

Comment 21

[BugBot [:suhaib / :marco/ :calixte]]

The severity field for this bug is relatively low, S3. However, the bug has 3 duplicates.
:peterv, could you consider increasing the bug severity?

For more information, please visit auto_nag documentation.

Comment 22

[BugBot (nomail) [:suhaib / :marco/ :calixte]]

The last needinfo from me was triggered in error by recent activity on the bug. I'm clearing the needinfo since this is a very old bug and I don't know if it's still relevant.