Closed Bug 90363 Opened 23 years ago Closed 23 years ago

Mozilla causing corruption of Bugzilla bug data

Categories

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

Product:
Core
Component:
DOM: Core & HTML
Platform:
x86
Windows 2000
Type:
defect
Priority:
Not set
Severity:
blocker

Tracking

()

Status:
VERIFIED FIXED

People

(Reporter: asa, Assigned: jst)

Details

(Keywords: smoketest) 

I've been seeing a few bugs where the component has been changed from something
to ActiveX Wrapper unintentionally. I belive this is a case of Mozilla's form
submission screwing up our data again. We should block recent builds of Mozilla
(the last few days) until we figure out what's going on. Investigating. (see
bug 90294 and 90298 for examples)
Keywords: smoketest
Summary: Mozilla causing corruption of Bugzilla bug data → Mozilla causing corruption of Bugzilla bug data
bug 90298 actually isn't ActiveX Wrapper _yet_ but it shows up as that for me in 
Mozilla build today and I'll bet if I submit a change it will update the 
component.
testing new feature to block today's build. this comment should not be blocked
Using Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.2+) Gecko/20010711, when I 
summited changes to bug 83431, the pages saying changes were made showed but 
the changes weren't made, any data on bug modified.
rod is on sabbatical.  cc:ing some other people who might be interested.
with today's build I cannot report a new bug.  Not seeing any of these problems
in last night's builds.
can't reproduce after a couple of restarts. sorry for all the noise. leaving 
this open while ai investigate today to try to reproduce but we can discontinue 
the blockage of today's browser.
Using Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.2+) Gecko/20010711, I 
created a new bug on http://landfill.tequilarista.org/bz37339/show_bug.cgi?
id=264, and when i summited it, changed the component automatically.
Haven't restarted my pc yet. 
My summiting problem showed intermitently while trying things at landfill.


I just looked at all changes to bug components today and see that 
bernd.mielke@snafu.de also had this problem.

'Component' seems to be the only affected field. Not sure why
the other dropdowns aren't affected.

Bugs affected by this problem are.

|  60919 | 2001-07-11 10:49:22 | HTMLTables            | ActiveX Wrapper      |
|  84307 | 2001-07-11 10:50:53 | HTMLTables            | ActiveX Wrapper      |
|  54187 | 2001-07-11 10:54:57 | HTMLTables            | ActiveX Wrapper      |
|  90294 | 2001-07-11 11:05:06 | Browser-General       | ActiveX Wrapper      |
I've disallowed use of today's build on bugzilla.mozilla.org.
bernd.mielke@snafu.de seems to be having this problem too.
Bernd Changed 3 html tables bugs to the ActiveX Wrapper component.
Also, one of these bugs (54183) had the target milestone changed to ---.

This mainly seems to be affecting the component field although there
are other dropdowns on the bug page although i haven't fully investigated.

I don't see any problems with severity or priority. If there were a problem
with product getting changed, the user would notice it when bugzilla prompted
for a new component. The only questionable target milstone change was by 
Bernd in bug  html tables bugs to the ActiveX Wrapper component.
Also, one of these bugs had the target milestone changed to null.

This mainly seems to be affecting the component field although there
are other dropdowns on the bug page although i haven't fully investigated.

I don't see any problems with severity or priority. If there were a problem
with product getting changed, the user would notice it when bugzilla prompted
for a new component. The only questionable target milstone change was by
Bernd in bug 54183. (By 'questionable', i mean that i don't think Bernd
meant to make this change). 


all changes to operating system so far today


+--------+---------------------+----------+----------+----------------------+
| bug_id | bug_when            | oldvalue | newvalue | login_name           |
+--------+---------------------+----------+----------+----------------------+
|  80479 | 2001-07-11 04:05:49 | PC       | All      | bclary@netscape.com  |
|  90293 | 2001-07-11 05:05:09 | PC       | All      | mpt@mozilla.org.uk   |
|  90300 | 2001-07-11 07:44:17 | PC       | All      | ask@swva.net         |
|  89297 | 2001-07-11 07:46:13 | PC       | All      | _basic@yahoo.com     |
|  90297 | 2001-07-11 07:53:24 | PC       | All      | ask@swva.net         |
|  90312 | 2001-07-11 08:00:08 | Other    | All      | ask@swva.net         |
|  90325 | 2001-07-11 08:40:13 | PC       | All      | axel@pike.org        |
|  90264 | 2001-07-11 09:03:31 | PC       | All      | tguyot@geocities.com |
|  83765 | 2001-07-11 14:01:21 | PC       | All      | ask@swva.net         |
+--------+---------------------+----------+----------+----------------------+


all changes to OS so far today


| bug_id | bug_when            | oldvalue     | newvalue | login_name          
         |
+--------+---------------------+--------------+----------+-------------------------------+
|  90271 | 2001-07-11 00:21:49 | Windows 95   | All      | t.bugz@lament.cjb.net
        |
|  80479 | 2001-07-11 04:05:49 | Linux        | All      | bclary@netscape.com 
         |
|  90312 | 2001-07-11 06:38:17 | Windows 98   | All      |
jarrod.k.gray@rose-hulman.edu |
|  83661 | 2001-07-11 06:42:44 | Windows 2000 | All      | kbh7@cornell.edu    
         |
|  90300 | 2001-07-11 07:44:17 | Linux        | All      | ask@swva.net        
         |
|  89297 | 2001-07-11 07:46:13 | Windows 2000 | All      | _basic@yahoo.com    
         |
|  90297 | 2001-07-11 07:53:24 | Linux        | All      | ask@swva.net        
         |
|  89113 | 2001-07-11 08:12:38 | Windows 98   | All      |
nobandwidth@bigpond.com       |
|  90325 | 2001-07-11 08:40:13 | Windows 2000 | All      | axel@pike.org       
         |
|  90264 | 2001-07-11 09:03:31 | Windows 2000 | All      | tguyot@geocities.com
         |
|  90337 | 2001-07-11 09:54:40 | Windows NT   | Linux    | harishd@netscape.com
         |
|  90299 | 2001-07-11 09:57:59 | Linux        | All      | alexey@ihug.com.au  
         |
|  90378 | 2001-07-11 13:37:20 | Solaris      | All      | fenella@netscape.com
         |
Whiteboard: 3 html tables bugs to the ActiveX Wrapper component.Also, one of these bugs had the target milestone changed to null.
Hi Rod, how is this coming? It's getting pretty late in the afternoon and we
haven't seen any progress from you yet on this smoketest blocker. It'd be great
if you could tell us where you are so we can get an estimate as to when the tree
is going to open. Thanks!
As I said above, Rod is on sabbatical.  Asa said at 12:23 (above) that we
shouldn't hold today for this, since it's only happening for 2 or 3 people.  Or
should we be holding for it?  If so, does anybody have any ideas about how to
reproduce it or are any of the people seeing it using a build in which they can
back out changes?
Now that I'm out of meetings, downgrading this one to critical to get off the
blocker list for today.
Severity: blocker → critical
If we downgrade this from blocker status then we need to unbock bugzilla so you
can use bugzilla with todays build. Not being able to use bugzilla is a 
smoketest blocker. Dawn, can you unblock it for us?
I'm not sure if this is pertinent or not, but some combobox reflow code was
changed recently. If you *can* repro this bug it might be worth backing out the
changes to nsComboboxControlFrame.cpp ver 1.180  It is a shot-in-the-dark, but
stranger things have happened in layout before.
In looking at the values changed by fenella@netscape.com and 
harishd@netscape.com, it appears that the change was intentional / related to 
their comments in the bug report.


"It only happened to two people" is a pretty weak argument. The problem was 
detected early and we caught the corruption and blocked other possible
corruption before it caused much of a problem. 


Asa can't reproduce the bug now but this morning he was in a state where loading
the show_bug page consistantly changed the component to the first on the list. 
This is a real problem. I'm not going to remove the block on bugzilla until i
can be confident that the browser isn't going to ruin our data.
Whiteboard: 3 html tables bugs to the ActiveX Wrapper component.Also, one of these bugs had the target milestone changed to null.
Here's what I saw this morning.  I downloaded and installed the early 
talkback.zip build into a clean directory. I started the browser and mail. after 
looking at a handful of bugs I noticed that every bug I was viewing had the 
Component set to the first item on the list (ActiveX Wrapper for browser and 
Accout Manager for mailnews). Looking over the bugs I had made changes to I 
noticed that one of them had the Component changed to ActiveX.  Blake just 
confirmed this on linux looking at several of his bugs and all the Components 
wers ActiveX
this is reproduceable. Do any query then visit a bug on the list ans see the 
component reset. This is a blocker.
Severity: critical → blocker
Asa says that if he opens a bug by clicking on a link in the query result's
bug list, and doing 'Open Link in New Window', then the bug loads properly.
Clicking the link directly and loading the show_bug.cgi page in the current
window changes the component to the one at the beginning of the list.

cc'ing jst because blake suspects him.

It looks like all the changes to OS that I listed earlier are valid. Bug 89297
is a possible exception but I don't think so because the suspect change was made
before the morning builds showed up on the ftp site this morning.
Taking this since Rod is on sabatical. I have not been able to reproduce it as
described, however (WinNT installer build from this morning), so any more
specifics from people who can reproduce it would be appreciated.
Assignee: rods → pollmann
Here's how I was able to reproduce this with a linux cvs build from this
morning:
1. in a fresh browser window, go to http://bugzilla.mozilla.org/
2. click "View bugs already reported today"
3. click the first bug, and notice the component is correct.
4. Go to http://bugzilla.mozilla.org/
5. click the query page link
6. go to the bottom and in the action bug # put 90363 and press find
7. Notice the component is now "ActiveX Wrapper"

Extra feature: hit "view bug activity" and notice a javascript error:
Error: f.product has no properties
Source File: http://bugzilla.mozilla.org/query.cgi  Line: 813
(This page isn't even loaded.  I suspect either the onLoad or onChange in
query.cgi is being fired again, and that's causing at least part of the problem)

Once it starts to act up in a window, it seems to continue to act up.  If you
use open link in new window such that a show_bug.cgi page hasn't been loaded in
the same window with query.cgi, everything seems fine.
eric, blake says this is jst.  he backed out jst from this morning in his tree 
and fixed the problem. maybe reassign?
Thanks, handing over to Johnny to investigate.
Assignee: pollmann → jst
I have a fix for this, checking in soon...
Status: NEW → ASSIGNED
Which bug is jst's bug which may have caused this?
I need to see if this is only affecting the trunk.
Lisa, this is trunk only. Bug #83433.
jst to checkin a fix around 8:15pm tonight, after dinner
Fixed.
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
verifying
Status: RESOLVED → VERIFIED
Component: HTML: Form Submission → DOM: Core & HTML
You need to log in before you can comment on or make changes to this bug.