301240 - editvalues.cgi should not update bugs.delta_ts

Status: RESOLVED FIXED

(Bugzilla :: Administration, task)

Description

[Albert Ting]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.5) Gecko/20041223 Firefox/1.0
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.5) Gecko/20041223 Firefox/1.0

I'm not sure what's supposed to happen, only that after I reran checksetup.pl,
all my "last modified" fields got set to today (mysql's NOW()).  I ended up
having to write a script that reset delta_ts based on the max bug_when in
longdescs and bugs_activity. 

It also seems bugs.delta_ts is supposed to hold the date, not a delta date.  Or
at least, the "last modified" text for each bug report looks correct if I set it
to the right date.

Granted, I have a couple custom patches, but I believe this is occurring with
the standalone 2.20rc1 copy. 



Reproducible: Always

Comment 1

[Albert Ting]

Ah, it's not checksetup, but editvalues.cgi.  Whenever you change a field value,
all bugs that had that field modified gets the delta_ts field updated accordingly.
In my case, that impacted nearly all of my bugs.

I'm not sure I like that, for if you click on a specific bug, there is no
relevant info stating why the last_modified field got changed.  

Comment 2

[Frédéric Buclin]

editvalues.cgi, line 358:

        $dbh->do("UPDATE bugs SET $field = ?, delta_ts = NOW()
                   WHERE $field = ?", undef, $value, $valueold);

So we explicitly update delta_ts here. But I agree, this doesn't make sense!
This would make sense only if users were getting an email about this change,
which is not the case here.

mkanat, can you justify this decision to update delta_ts?

Comment 3

[Frédéric Buclin]

Attachment: patch, v1

The actual code breaks all queries based on the "last changed time". That's
bad!

Comment 4

[Albert Ting]

(In reply to comment #3)
> The actual code breaks all queries based on the "last changed time". That's
> bad!

Yes, this is how I originally found out about it and I personally don't like it.  

Comment 5

[Frédéric Buclin]

Albert, we all agree :)

(02:37:19) LpSolit: myk, could you have a look at my last comment in bug 301240?
I would like your opinion
(02:38:45) myk: LpSolit: it doesn't make sense to me either
(02:41:38) LpSolit: myk: ok, I will fix it now. The actual code breaks all
queries based on the "last changed time"
(02:42:29) myk: LpSolit: yeah, that sucks

Comment 6

[Albert Ting]

(In reply to comment #5)
> Albert, we all agree :)

Great, can it be a 2.20 blocker?  It's a small change. 

Comment 7

[Max Kanat-Alexander]

Yeah, I don't know why I have that code there, either. :-)

Comment 8

[Marc Schumann [:Wurblzap]]

Comment on attachment 189730 [details] [diff] [review]
patch, v1

Tested; works.

Comment 9

[Marc Schumann [:Wurblzap]]

I don't have a 2.20 branch ready, so I didn't check whether it applies.

Comment 10

[Frédéric Buclin]

(In reply to comment #9)
> I don't have a 2.20 branch ready, so I didn't check whether it applies.

The patch applies cleanly for 2.20 too. This part of editvalues.cgi is still
marked as v1.1 (original file). ;)

Comment 11

[Frédéric Buclin]

tip:

Checking in editvalues.cgi;
/cvsroot/mozilla/webtools/bugzilla/editvalues.cgi,v  <--  editvalues.cgi
new revision: 1.4; previous revision: 1.3
done


2.20rc1:

Checking in editvalues.cgi;
/cvsroot/mozilla/webtools/bugzilla/editvalues.cgi,v  <--  editvalues.cgi
new revision: 1.3.2.1; previous revision: 1.3
done

Comment 12

[Bernd Kuemmerlen]

(In reply to comment #0)
> I'm not sure what's supposed to happen, only that after I reran checksetup.pl,
> all my "last modified" fields got set to today (mysql's NOW()).  I ended up
> having to write a script that reset delta_ts based on the max bug_when in
> longdescs and bugs_activity. 

Would you care to share your script? We have a similar problem where adding a new group control to a bug product set all bugs as modified.

Comment 13

[Albert Ting]

Attachment: fix delta_ts

Bernd, this is a rather old script.  You'll need to tweak it for the newer versions of Bugzilla. 

However, if adding a new group control to a bug product set all bugs as modified (for the latest bugzilla), then this in itself is probably a bug and should post a new ticket.

Comment 14

[Bernd Kuemmerlen]

Thanks, your script did the trick. I had to tweak it a bit though, because adding the group permissions is noted in bugs_activity, so I had to filter the query from bugs_activity to bugs where 'added' contained the group I added. Worked like a charm!

Comment 15

[Bernd Kuemmerlen]

BTW, I have filed Bug 380085 for my problem, but now (since I had a closer look at the DB), I am not sure if it is a bug after all... It is still annoying.