480303 - location bar recurrent cyclic incrementation on some flash pages, becomes really slooooooow

Status: RESOLVED DUPLICATE of bug 475178

(Core :: General, defect)

Description

[Emmanuel JARRI]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090224 Minefield/3.2a1pre
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090224 Minefield/3.2a1pre

on some flash pages, when I click on the next page link, the adress gets completed recurrently and whole system freezes for some time.
As we can see in the reported url,
http://surcouf.com/Catalogue/ListeProduit.aspx?TypeAction=goProd&idnoeud=3061#numPage:1#numPage:13numPage:11numPage:9numPage:7numPage:5numPage:3
increment by step 2

Reproducible: Always

Actual Results:  
system freezes for a moment

Expected Results:  
open the next flash page

happens on some other pages too (almost only flash pages I think)

Comment 1

[Ria Klaassen (not reading all bugmail)]

I see only that it hangs like crazy when I load the page. Seems to be caused by Bug 472706.

Comment 2

[(not currently active) Ted Mielczarek]

I suspect that this is a duplicate of bug 475178. See bug 475178 comment 19 where I figured out that location.hash=foo is broken. Looks like the same result here.

Comment 3

[(not currently active) Ted Mielczarek]

Can someone test this build to see if it fixes your issue:
http://mavra.perilith.com/~luser/latest-teds-builds/firefox-3.2a1pre.en-US.win32.zip

If so, then this bug is definitely a duplicate of bug 475178.

Comment 4

[Emmanuel JARRI]

OK, it seemed better in a first run !
(the second page correctly appeared, and quickly ;-)

BUT, on a deeper test, on this page :
http://surcouf.com/Catalogue/ListeProduit.aspx?TypeAction=goProd&idnoeud=3069#numPage:01
It worked well until after multiple pages swap, then I got this url :
http://surcouf.com/Catalogue/ListeProduit.aspx?TypeAction=goProd&idnoeud=3069#numPage:6
http://surcouf.com/Catalogue/ListeProduit.aspx?TypeAction=goProd&idnoeud=3069#numPage:10
Even if there is only 3 pages, and it got slow again.
But the pages correctly appeared, that wasn't the case before.

Sorry if the 2 bugs are related, how can we merge them ?

Comment 5

[(not currently active) Ted Mielczarek]

Thanks for testing! Those URLs look ok, so it definitely seems like the same bug. If it's getting slow, there might be another bug in there, feel free to file a new bug on that.

Don't worry about having filed a duplicate, it wasn't obvious what the root cause was in either of these bugs without a lot of debugging.

Comment 6

[Emmanuel JARRI]

I'm Ok to fill a new bug report (I have another one to fill too), but, I don't see how to explain what happens.
In fact, the pages appear, but the page number continues to be weirdly incremented, and goes (over,beyond,upon ?) the real number of pages
(in this case, there is 3 real pages, but page number increments to 6,7,8, etc. without limit.

Ted, I've another remark regarding your build : 
(I'd prefer waiting for your response about this before filling a new bug)
I ran Acid 3 test, and I managed to reach the 94/100 step, that I didn't managed to reach with the last nightly build (20090226035935), even after 20 tests (Yes, the score isn't always the same).

In fact, I rank 92/100, 93/100 or 94/100 with your version, against 92/100 or 93/100 with the last nightly buid.
Have you done some new code that seems to allow this ? Or is it related to your patch regarding this bug and bug 475178 ?

I have the screenshots of my tests, (including the dialog box with the details of the failed tests).

In fact, there is 2 different things :
_ the nightly build doesn't always score the same (92% or 93%) -> bug report ?
_ your version score 94% -> do you have any idea about why ? can I send you the details ?

Comment 7

[(not currently active) Ted Mielczarek]

I have no idea about the acid3 bits, the only patch I have locally is the bugfix from bug 475178. If you can narrow down to the specific tests that are different from my build and the nightly, we might be able to track down what's happening there. If we're intermittently failing a test, I'd like to know that too.

Comment 8

[Emmanuel JARRI]

I added a commentary on the bug 410460 page, with your build's link.
Can you keep your build for some time ? in order some others test it.

Nevertheless, I would like to thank you all for your work !
Minefield is my everyday loved browser for some months now :-)

Comment 9

[(not currently active) Ted Mielczarek]

I don't tend to delete builds, but I do occasionally overwrite builds there with other test builds. I've copied the build to a different location so I can avoid overwriting it accidentally:
http://mavra.perilith.com/~luser/latest-teds-builds/firefox-3.2a1pre.en-US.win32.bug475178.zip (same exact build, slightly different URL)