308633 - Horrible performance on Ubuntu Bugzilla site, involves table reflow

Status: RESOLVED WORKSFORME

(Core :: Layout: Tables, defect)

Description

[IU]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b4) Gecko/20050914 Firefox/1.4
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8b4) Gecko/20050914 Firefox/1.4

When trying to report a bug for Ubuntu, you can click a "Package" link (to see a
full list) or enter the package name manually (packages are displayed as you
type).  If you click the "Package" link, Firefox maxes the CPU, sucks RAM (over
130 MB), and takes *forever* to render the list (I killed it so I don't know how
long it would have eventually taken).  Internet Explorer accomplishes the same
task in about 20 seconds using less than 21 % CPU utilization (hitting ~90+
utilization towards the last 5 seconds) and less than 40 MB of RAM.


Reproducible: Always

Steps to Reproduce:
1. Go to http://bugzilla.ubuntu.com/enter_bug.cgi
2. Click LiveCD on the left
3. Login with your Ubuntu Bugzilla account (or create one if you don't have one,
and then login)
4. Open Windows Task Manager and locate Firefox
5. Back at the Ubuntu Bugzilla page, click "Package" (about middle of the page)
6. Observe the CPU utilization, memory growth, and time to render.

NOTE: My system is a P III 933 MHz w/ 512 MB of PC133 RAM.  YMMV so please don't
be quick to dismiss this bug because you're running a Cray XD1.

Actual Results:  
Firefox maxing out my system.


Expected Results:  
Performance comparable to Internet Explorer would be nice.

Comment 1

[Martijn Wargers (dead)]

Attachment: not minimal testcase

Tables issue.

Comment 2

[Felix Miata]

Isn't this the same as bug 119309 or bug 254185 ? Mandriva's bugzilla has been
dismal forever.

Comment 3

[Martijn Wargers (dead)]

No, it's not a duplicate of those bugs, since the "not minimal testcase" doesn't
have <select> tags in them.
But this could very well (even likely) be a duplicate of one of the table bugs
(haven't really looked, yet).

Comment 4

[Boris Zbarsky [:bzbarsky]]

Nothing terribly surprising in the profile; just the fact that this is O(N^2) in
length of table because every time we add a chunk of content we reflow the whole
thing....

Comment 5

[Bernd]

is there any evidence that the ubuntu bugzilla is slow while mozilla's is faster?

Comment 6

[IU]

> is there any evidence that the ubuntu bugzilla is slow while mozilla's is faster?

What difference does that make?  Compare IE's performance with Firefox, and then
Ubuntu Bugzilla site performance becomes a rabbit trail leading to nowhere.

Comment 7

[Bernd]

>What difference does that make?
The only difference it makes to me is: comment 4 indicates thats a general
reflow architecture bug, so its simply a dupe of bug 54542.  It turns out that
https://bugzilla.mozilla.org/show_bug.cgi?id=54542#c140 was too optimistic so it
might be more https://bugzilla.mozilla.org/show_bug.cgi?id=54542#c141.
On the other hand if the issue is more specific I might be able to fix it before
that.

Comment 8

[Stephanie Daugherty [:sdaugherty]]

The site listed as a test case here has been obsoleted - Ubuntu doesn't use bugzilla anymore. I'm going to close this  as there is not, and will likely never be information necessary to reproduce this.
.

Comment 9

[Boris Zbarsky [:bzbarsky]]

Stephanie, the information needed to reproduce this bug is attached right to the bug.  This is why we attach testcases instead of just pointing to them... it works, it really does.  ;)

Comment 10

[Wayne Mery (:wsmwk)]

WFM Mozilla/5.0 (Windows NT 6.0; rv:2.0b12pre) Gecko/20110214 Firefox/4.0b12pre
loads <2sec  ~25MB memory