864412 - Random zeroes (0) appearing in html content

Status: VERIFIED FIXED

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

Description

[Jordan Arentsen]

Attachment: zerosincontent.png

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:23.0) Gecko/20130422 Firefox/23.0
Build ID: 20130422030937

Steps to reproduce:

Loading a page with DOM insertions via jQuery (1.7.2). Not sure if related, but SVG is rendered on the page as well.


Actual results:

Page renders with seemingly random zeros (the 0 character) between inserted nodes. 


Expected results:

No zeroes should be randomly inserted.

Comment 1

[Jordan Arentsen]

After more testing, it seems to happen intermittently. This makes it hard to pinpoint what code is causing it.

Comment 2

[Liz Henry (:lizzard)]

Hi Jordan, Do you have a url I could take a look at to see if I can duplicate the issue? Or a reduced test case that you could attach?  Thanks!

Comment 3

[Jordan Arentsen]

Hi Liz, unfortunately it's not a publicly available application. I will attempt to make a reduced test case that has this issue.

Comment 4

[Jordan Arentsen]

Liz, I think I found a public site where this is happening. On Mugasha at the end of a set it will display a popup with next suggested sets. Oftentimes (but not always) this pop-up will contain the zeros that I'm seeing in my own project. 

Example link: http://mugasha.com/international-departures/007#/track/51/Tomcraft-vs.%20The%20Doppler%20Effect/Loneliness%20%28Myon%20&%20Shane%2054%20Summer%20Of%20Dub%20Mix%29%20vs.%20Beauty%20Hides%20In%20The%20Deep%20%28Acapella%29

Comment 5

[Bob Clary [:bc] (inactive)]

I see this in Nightly / Linux on phpmyadmin and also at least a local version of phonedash

You might try: http://mrcote.info/phonedash or http://demo.phpmyadmin.net/

Comment 6

[Mark Côté [:mcote]]

Also in Nightly on OS X 10.8.3.

Comment 7

[Bob Clary [:bc] (inactive)]

To reproduce: go to http://mrcote.info/phonedash then click the cached checkbox, wait for repaint, repeat until a 0 appears below the legend on the left side of the page.

I first noticed this with yesterday's Nightly. Not reproducible on Aurora.

Comment 8

[Josh Matthews [:jdm]]

I see this on Mozilla's yammer instance with today's nightly on Fedora 17.

Comment 9

[Reed Loden [:reed]]

I'm seeing this all the time on GroupMe and when upgrading a GitHub instance.

Comment 10

[Félix Saparelli [:passcod]]

I'm seeing this all the time on https://web.tweetdeck.com and games of http://triton.ironhelmet.com. It appears mostly on insertion of content / DOM nodes, afaics.

Comment 11

[u279076]

If you are able to reproduce this please use mozregression to find a regression range:
http://mozilla.github.io/mozregression/

Thank you

Comment 12

[Bob Clary [:bc] (inactive)]

using phpmyadmin
Last good nightly: 2013-04-21
First bad nightly: 2013-04-22

Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=0d50cb959c46&tochange=50d25e083421

Comment 13

[u279076]

(In reply to Bob Clary [:bc:] from comment #12)
> using phpmyadmin
> Last good nightly: 2013-04-21
> First bad nightly: 2013-04-22
> 
> Pushlog:
> http://hg.mozilla.org/mozilla-central/
> pushloghtml?fromchange=0d50cb959c46&tochange=50d25e083421

Thanks Bob. Can you regress this any further with tinderbox builds?

Comment 14

[Bob Clary [:bc] (inactive)]

I'm doing manual builds and bisecting atm. I'll let you know later this evening.

Comment 15

[Alice0775 White]

STR
1. Open http://mrcote.info/phonedash
2. Toggle "Cached" checkbox
3. Repeat Step2.

Actual Results:
  Rundom numerical number such as 0, 9.346043567475603e-307 appears at end of Legent as follows:
	c8_aa_21_ac_0c_b5_droid-pro 00000


Regression window(m-i)
Good:
http://hg.mozilla.org/integration/mozilla-inbound/rev/123a65dba090
Mozilla/5.0 (X11; Linux i686; rv:23.0) Gecko/20130420 Firefox/23.0 ID:20130420072406
Bad:
http://hg.mozilla.org/integration/mozilla-inbound/rev/503a5fb6d530
Mozilla/5.0 (X11; Linux i686; rv:23.0) Gecko/20130420 Firefox/23.0 ID:20130420073304
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=123a65dba090&tochange=503a5fb6d530

Triggered by
503a5fb6d530	Kannan Vijayan — Bug 861596 - Add optimized ArgumentsObject stubs to Ion ICs. r=h4writer


It is better to double-check it about the regression range.

Comment 17

[Alice0775 White]

Sorry, typo in comment #15
s/Legent/Legend/

Comment 18

[Andrew McCreight [:mccr8]]

I've seen this once on the Mozilla Yammer page.  dougt said he's seen it there, too.

Comment 19

[Sean Stangl [:sstangl]]

Grooveshark's Retro interface reproduces the issue consistently, and is public:

http://retro.grooveshark.com/#!/search?q=test

Comment 20

[u279076]

(In reply to Sean Stangl [:sstangl] from comment #19)
> Grooveshark's Retro interface reproduces the issue consistently, and is
> public:
> 
> http://retro.grooveshark.com/#!/search?q=test

Thanks, adding that to the URL field.

Comment 21

[Kannan Vijayan [:djvj]]

This is utterly bizarre.  I can't fathom how an Ion IC change causes this.  Will take a look.

Comment 22

[Ryan VanderMeulen [:RyanVM]]

Happens on TBPL too in the list of trees.

Comment 23

[Cork]

Nice i've seen this one too. Been stuck trying to make a testcase.

Comment 25

[Bob Clary [:bc] (inactive)]

Confirmed Alice's bisection via manual builds and hg bisect.
The first bad revision is:
changeset:   129418:503a5fb6d530
user:        Kannan Vijayan <[e-mail]>
date:        Sat Apr 20 10:32:04 2013 -0400
summary:     Bug 861596 - Add optimized ArgumentsObject stubs to Ion ICs. r=h4writer

Comment 26

[[:Aleksej]]

The reporter of bug 866075 (about non-zero numbers like in Alice's bug 866013) reported not being able to reproduce it with 2013-04-28.

Comment 27

[Hannes Verschore [:h4writer]]

(In reply to Aleksej [:Aleksej] from comment #26)
> The reporter of bug 866075 (about non-zero numbers like in Alice's bug
> 866013) reported not being able to reproduce it with 2013-04-28.

Makes sense. The revision that introduced this issue has been reverted/backed out on 2013-04-28. That is the revision mentioned in comment 25 .

Comment 29

[Kannan Vijayan [:djvj]]

This bug was fixed in the re-push of patch for bug 861596.

Comment 30

[u279076]

(In reply to Kannan Vijayan [:djvj] from comment #29)
> This bug was fixed in the re-push of patch for bug 861596.

Thanks Kannan.

Anyone on this bug who was seeing this issue, please download or update to today's Nightly and see if this is now resolved for you.

Thank you.

Comment 31

[Cork]

I can't reproduce the problem in http://hg.mozilla.org/mozilla-central/rev/b842d26dd5f0 that I saw in http://hg.mozilla.org/mozilla-central/rev/503a5fb6d530

Comment 32

[u279076]

Thank you Cork.