326217 - Change Several Bugs at Once goes into an endless loop

Status: RESOLVED DUPLICATE of bug 325579

(Bugzilla :: Creating/Changing Bugs, defect)

Description

[Ståle Undheim]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8) Gecko/20051224 Debian/1.5.dfsg-3 Firefox/1.5
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8) Gecko/20051224 Debian/1.5.dfsg-3 Firefox/1.5

I do a search and modify several bugs at once. This results in an seemingly infinite loop beeing executed when I add a comment, causing the same comment to be added over 200 times to each of the bugs I change.

The steps I go through to reproduce this:
1. Search for bugs
2. Click change several bugs at once
3. Change something and add a comment
4. Check all bugs
5. Submit

The page will then load, showing bugs changed, then just continue to load more and more of the same output, writing change to the same bug several times.

Reproducible: Always

Steps to Reproduce:
1. Search for bugs
2. Click change several bugs at once
3. Change something and add a comment
4. Check all bugs
5. Submit
Actual Results:  
Each bug is modified, and the comment I entered is added multiplle times for each bug, usually over 100 times.

Expected Results:  
Each bug modified and only 1 comment added.

Installation is from FreeBSD 4.10-RELEASE ports system, version 2.20.

Output of checksetup.pl:
Checking perl modules ...
Checking for       AppConfig (v1.52)   ok: found v1.56
Checking for             CGI (v2.93)   ok: found v3.15
Checking for    Data::Dumper (any)     ok: found v2.121
Checking for    Date::Format (v2.21)   ok: found v2.22
Checking for             DBI (v1.38)   ok: found v1.50
Checking for      File::Spec (v0.84)   ok: found v3.16
Checking for      File::Temp (any)     ok: found v0.16
Checking for        Template (v2.08)   ok: found v2.14
Checking for      Text::Wrap (v2001.0131) ok: found v2001.0929
Checking for    Mail::Mailer (v1.65)   ok: found v1.73
Checking for        Storable (any)     ok: found v2.15

The following Perl modules are optional:
Checking for              GD (v1.20)   ok: found v2.30
Checking for     Chart::Base (v1.0)    ok: found v2.3
Checking for     XML::Parser (any)     ok: found v2.34
Checking for       GD::Graph (any)     ok: found v1.4305
Checking for GD::Text::Align (any)     ok: found v1.18
Checking for     PatchReader (v0.9.4)  ok: found v0.9.5

Checking user setup ...
Removing existing compiled templates ...
Precompiling templates ...
Checking for      DBD::mysql (v2.9003) ok: found v3.0002
Checking for           MySQL (v3.23.41) ok: found v4.0.21-standard
Checking for        GraphViz (any)     ok: found

Comment 1

[Marc Schumann [:Wurblzap]]

Does the loop stop by itself after some time, or do you need to stop it (by pressing your browser's "stop" button or similar)?

Comment 2

[Marc Schumann [:Wurblzap]]

I see you're on the CC list of bug 325579 already, which is about changes on single bugs, but seems very related...

I have currently no idea what might be causing this, so I'll ask into the blue: do you have the mod_perl Apache module active, by chance?

Comment 3

[Ståle Undheim]

mod_perl is not installed.

I do stop the webpage, but I usually have to run in to our system administrator asking him to kill all the perl processes.

It seems to spawn quite alot of perl processes.

Comment 4

[Ståle Undheim]

I created a new product specifically for testing this bug, created 2 bugs, and changed both at once.

I did not get the same result her.

Difference on this product compared to the others:
No QA email set for the component
This product and component was created in bugzilla 2.20, while the others was created in bugzilla 2.16, which we have just upgraded to 2.20.

Bug #325579 does not manifest itself either here.

Comment 5

[Marc Schumann [:Wurblzap]]

Hmmm.... so I'm all the more stumped... Does your Apache error log offer some insight on this?

Comment 6

[Ståle Undheim]

Attachment: Output of changing 2 bugs

So I modified 2 bugs, and here is the output.

Bug 1252 got 1 new comment (as expected)
Bug 1967 got 4 new comments (3 more than expected)

I am thinking this is related to QA contact, so I will try again with that disabled.

Comment 7

[Ståle Undheim]

Nothing in error log that I can see related to this.

I do however think I have found the cause.

I had parameters set to smtp mailing, and changed it back to sendmail now, and it works as expected again, even with QA contact enabled.

It did seem to produce 1 comment pr. mail multiplied by number of bugs when I changed several bugs at once.

It also seemed to spawn 1 page duplicate for each mail when I just submitted a bug.

I only got no errors with the smtp setup when I modified bugs assigned to me with no QA contact (meaning I got excluded from the mailing list).

Comment 8

[Marc Schumann [:Wurblzap]]

Ok... If you don't mind, let's merge this bug with bug 325579 so that we have discussion and findings in one place. The two bugs seem to me to be a single one after all.

*** This bug has been marked as a duplicate of 325579 ***