240093 - Unable to confirm other people's bugs, even with the canconfirm permission

Status: RESOLVED FIXED

(Bugzilla :: User Accounts, defect)

Description

[Petey]

I have the canconfirm permission, but I can't confirm.  I am only having
problems confirming bugs from other users; I /am/ able to register my own bugs
as 'New'.

Comment 1

[Will Jenkins]

Yes I have the same problem.

Comment 2

[Dave Miller [:justdave]]

Is this on mozilla.org?  How long ago did this start happening?

Comment 3

[Will Jenkins]

Yes this is on bugzilla.mozilla.org. At first I guessed it must be because I was
given confirm on the mozilla browser and that I was now trying to confirm bugs
on Camino and Firefox, but I can see no mention of privs being restricted to
products.

Comment 4

[Chris Ilias [:cilias]]

Same thing happened to me on bug 224166.
I can select "Confirm bug (change status to NEW)", add a comment, and add to CC
list. In the bug processed screen, my comment is added, my address is added to
the CC list, but the status stays at UNCONFIRMED.

Comment 5

[Will Jenkins]

Chris, I tried confirming to new on Bug 224166 but had the same problem. This is
very annoying.

Comment 6

[Gervase Markham [:gerv]]

Dave: several people are seeing this. It's somewhat scary, and would be good to
nail down for 2.18. What can we do to diagnose it, do you think? Get one of
these users to lend us their account?

Gerv

Comment 7

[Petey]

I volunteer my account, if you can use it.

Comment 8

[Dave Miller [:justdave]]

Yeah, probably good to nail it, if we can.  No idea how to trace it...  we
should be able to clone an account instead of stealing one...

Comment 9

[Gervase Markham [:gerv]]

Attachment: Patch v.1

Seems like canconfirm just plain didn't work :-( 

The code does a CheckCanChangeField() call for a fake field called
"canconfirm". CheckCanChangeField() wasn't allowing people in the canconfirm
group to change any field apart from specific values of bug_status, and so the
request was failing, and "everconfirmed = 1" was not being added to the SQL. So
the bug stayed UNCONFIRMED.

You could still confirm your own bugs because the reporter can change anything.
But no-one with canconfirm can confirm other people's bugs at the moment. This
has obviously been true since the last b.m.o. update, and as far as I can see,
since February 2003, when I "fixed" bug 186994.
http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&root=/cvsroot&subdir=mozilla/webtools/bugzilla&command=DIFF_FRAMESET&file=process_bug.cgi&rev2=1.178&rev1=1.177


Shows how little it's actually used without editbugs...

Gerv

Comment 10

[Joel Peshkin]

Comment on attachment 153278 [details] [diff] [review]
Patch v.1

r=joel
By inspection, not tested. But the logic makes total sense

Comment 11

[Gervase Markham [:gerv]]

Fixed on both trunk and branch.

Checking in process_bug.cgi;
/cvsroot/mozilla/webtools/bugzilla/process_bug.cgi,v  <--  process_bug.cgi
new revision: 1.206; previous revision: 1.205
done

Checking in process_bug.cgi;
/cvsroot/mozilla/webtools/bugzilla/process_bug.cgi,v  <--  process_bug.cgi
new revision: 1.205.2.1; previous revision: 1.205
done

Gerv

Comment 12

[Will Jenkins]

I don't mean to moan, but when will this actually take effect on this site?

Comment 13

[Gervase Markham [:gerv]]

When we next update. I don't know if we have an ETA for this at the moment; the
last ETA was "after Firefox 1.0".

Gerv