RT 3.6.0rc2

RT Developers
1

I’m pleased to announce the second release candidate of RT 3.6.0. (The
astute among you may well note that I never sent mail about RC1. We caught
a couple of showstoppers before I managed to get the email written).

RT 3.6.0rc2 includes a provisional fix for the dreaded “multi custom
field search” bug, backing out the second acl check on ticket linking
as well as a few smaller bug fixes.

You know the drill.

Jesse

r13786 (orig r5222): jesse | 2006-05-14 21:37:17 -0400

r13785@hualien: jesse | 2006-05-14 21:34:01 -0400

  • 3.6.0rc2

r13784 (orig r5221): jesse | 2006-05-14 21:31:30 -0400

r13781@hualien: jesse | 2006-05-14 21:08:33 -0400

  • fixed the test count for the cli tests

r13783 (orig r5220): jesse | 2006-05-14 21:31:19 -0400

r13780@hualien: jesse | 2006-05-14 21:08:21 -0400

  • When using the quick ticket creation widget, set the requestor to the current user

r13782 (orig r5219): jesse | 2006-05-14 21:31:10 -0400

r13779@hualien: jesse | 2006-05-14 21:04:13 -0400
By popular demand, adding a ticket link doesn’t check ACLs on the second ticket

r13378 (orig r5202): ruz | 2006-05-11 03:45:48 -0400

  • add test for CF.{cfname} search format
    r13377 (orig r5201): ruz | 2006-05-11 03:34:50 -0400

  • fixes in searches by CFs
    ** CustomFields alias could be undef, we not always join to this table
    ** cache both aliases we use (ObjectCustomFieldValues and CustomFields)
    r13376 (orig r5200): ruz | 2006-05-11 03:31:30 -0400

  • add several CF searches that are failing
    r13375 (orig r5199): ruz | 2006-05-11 03:30:35 -0400

  • drop noisy unint warning
    r13319 (orig r5182): trs | 2006-05-08 22:08:15 -0400

r10382@zot: tom | 2006-05-08 22:07:05 -0400
Some fixes for horizontal scrolling problems (the query builder will still cause horizontal scroll if the window isn’t wide enough, however, as a result of it’s layout)

r13318 (orig r5181): trs | 2006-05-08 21:07:04 -0400

r10380@zot: tom | 2006-05-08 21:06:34 -0400
More robust way of figuring out the user’s group which doesn’t rely on order

r13317 (orig r5180): kevinr | 2006-05-08 19:20:16 -0400

r12580@SAD-GIRL-IN-SNOW: kevinr | 2006-05-08 19:18:46 -0400

  • Spec’d out tests for ‘take’ and ‘steal’ via the CLI
2
> > I'm pleased to announce the second release candidate of RT 3.6.0. (The astute among you may well note that I never sent mail about RC1. We caught a couple of showstoppers before I managed to get the email written). > > RT 3.6.0rc2 includes a provisional fix for the dreaded "multi custom field search" bug, backing out the second acl check on ticket linking as well as a few smaller bug fixes. > > You know the drill. > > Jesse

Hi Jesse,

Are there any screenshots available, or can we do some, of the 3.6 series?

Or is it the same as the current CPAN version?

Thanks,

Gavin.

r13786 (orig r5222): jesse | 2006-05-14 21:37:17 -0400

r13785@hualien: jesse | 2006-05-14 21:34:01 -0400

  • 3.6.0rc2

r13784 (orig r5221): jesse | 2006-05-14 21:31:30 -0400

r13781@hualien: jesse | 2006-05-14 21:08:33 -0400

  • fixed the test count for the cli tests

r13783 (orig r5220): jesse | 2006-05-14 21:31:19 -0400

r13780@hualien: jesse | 2006-05-14 21:08:21 -0400

  • When using the quick ticket creation widget, set the requestor to the
    current user

r13782 (orig r5219): jesse | 2006-05-14 21:31:10 -0400

r13779@hualien: jesse | 2006-05-14 21:04:13 -0400
By popular demand, adding a ticket link doesn’t check ACLs on the second
ticket

r13378 (orig r5202): ruz | 2006-05-11 03:45:48 -0400

  • add test for CF.{cfname} search format

r13377 (orig r5201): ruz | 2006-05-11 03:34:50 -0400

  • fixes in searches by CFs
    ** CustomFields alias could be undef, we not always join to this table
    ** cache both aliases we use (ObjectCustomFieldValues and CustomFields)
    r13376 (orig r5200): ruz | 2006-05-11 03:31:30 -0400

  • add several CF searches that are failing

r13375 (orig r5199): ruz | 2006-05-11 03:30:35 -0400

  • drop noisy unint warning

r13319 (orig r5182): trs | 2006-05-08 22:08:15 -0400

r10382@zot: tom | 2006-05-08 22:07:05 -0400
Some fixes for horizontal scrolling problems (the query builder will
still cause horizontal scroll if the window isn’t wide enough, however,
as a result of it’s layout)

r13318 (orig r5181): trs | 2006-05-08 21:07:04 -0400

r10380@zot: tom | 2006-05-08 21:06:34 -0400
More robust way of figuring out the user’s group which doesn’t rely on
order

r13317 (orig r5180): kevinr | 2006-05-08 19:20:16 -0400

3

Brilliant! This RC2 fixes many of our Problems, great Release Candidate!!!

No errors so far!

4

Hi RT Developer, Hi Jesse,

The RC2 Search by CF is cool, but now i found another small ?? Bug. If you
define a search and set the sort order to a CF, then you get empty results,
or better, RT displays xxx found tickets, but don’t display something.

Torsten

5

Or is it the same as the current CPAN version?

It’s almost identical to the current version running on rt.cpan.org

6
> > >> Or is it the same as the current CPAN version? > > It's almost identical to the current version running on rt.cpan.org

Righto. Thanks.

7

I’m pleased to announce the second release candidate of RT 3.6.0. (The
astute among you may well note that I never sent mail about RC1. We caught
a couple of showstoppers before I managed to get the email written).

r13782 (orig r5219): jesse | 2006-05-14 21:31:10 -0400

r13779@hualien: jesse | 2006-05-14 21:04:13 -0400
By popular demand, adding a ticket link doesn’t check ACLs on the second ticket

Jesse,

I get the impression that some people think checking ACLs on the
second ticket is a new behavior. It isn’t. That’s the way things
have been for a long time. A security change like this is
dangerous. It should be a config file option, not a change in the
default security policy.

What has changed is the creation of the transaction in the
second ticket. If there are problems associated with that
then a different soltion should be found.

-Todd

8

Jesse,

I get the impression that some people think checking ACLs on the
second ticket is a new behavior. It isn’t. That’s the way things
have been for a long time. A security change like this is
dangerous. It should be a config file option, not a change in the
default security policy.

My profound apologies. The nature of the reports confused me. It seemed
fairly clear from the reports that the permission check on the second
ticket was a new behaviour. It’ll be reverted and get a bit more study.

9

Hi Jesse,

FYI.

On a Solaris 9 box, the ./configure results in
…
id: illegal option – g
Usage: id [-ap] [user]

checking the diff between RC1 and RC2 I noticed that
$ diff rt-3.6.0rc2/configure rt-3.6.0rc1/configure
…
2027c2027
< my_group=$(id -gn)

my_group=$(groups|cut -f1 -d’ ')

On my Solaris, “man id” gets

System Administration Commands id(1M)
NAME
id - return user identity

SYNOPSIS
/usr/bin/id [-p] [user]
/usr/bin/id -a [-p] [user]

  /usr/xpg4/bin/id [-p] [user]
  /usr/xpg4/bin/id -G [-n] [user]
  /usr/xpg4/bin/id -g [-nr] [user]
  /usr/xpg4/bin/id -u [-nr] [user]

…

and /usr/bin is before /usr/xpg4/bin on the PATH. So the -g illegal
option results. No biggy.

Lidio.

10

Thanks. This will be fixed for rc3
Best,
Jesse
This message was sent from my Treo. Please accept my apologies for its brevity and for any typos.From: Lidio Presutti lidio@ic-unix.ic.utoronto.ca
Date: Monday, May 15, 2006 6:02 pm
Subject: [Rt-devel] Re: RT 3.6.0rc2

Hi Jesse,

FYI.

On a Solaris 9 box, the ./configure results in
…
id: illegal option – g
Usage: id [-ap] [user]

checking the diff between RC1 and RC2 I noticed that
$ diff rt-3.6.0rc2/configure rt-3.6.0rc1/configure
…
2027c2027
< my_group=$(id -gn)

my_group=$(groups|cut -f1 -d’ ')

On my Solaris, “man id” gets

System Administration Commands id(1M)
NAME
id - return user identity

SYNOPSIS
/usr/bin/id [-p] [user]
/usr/bin/id -a [-p] [user]

  /usr/xpg4/bin/id [-p] [user]
  /usr/xpg4/bin/id -G [-n] [user]
  /usr/xpg4/bin/id -g [-nr] [user]
  /usr/xpg4/bin/id -u [-nr] [user]

…

and /usr/bin is before /usr/xpg4/bin on the PATH. So the -g illegal option results. No biggy.

Lidio.
List info: The rt-devel Archives

Best Practical is hiring! Come hack Perl for us: Careers — Best Practical Solutions

11

Fixed by r5231On Mon, May 15, 2006 at 05:21:57PM -0400, Lidio Presutti wrote:

Hi Jesse,

FYI.

On a Solaris 9 box, the ./configure results in
…
id: illegal option – g
Usage: id [-ap] [user]

checking the diff between RC1 and RC2 I noticed that
$ diff rt-3.6.0rc2/configure rt-3.6.0rc1/configure
…
2027c2027
< my_group=$(id -gn)

my_group=$(groups|cut -f1 -d’ ')

On my Solaris, “man id” gets

System Administration Commands id(1M)
NAME
id - return user identity

SYNOPSIS
/usr/bin/id [-p] [user]
/usr/bin/id -a [-p] [user]

 /usr/xpg4/bin/id [-p] [user]
 /usr/xpg4/bin/id -G [-n] [user]
 /usr/xpg4/bin/id -g [-nr] [user]
 /usr/xpg4/bin/id -u [-nr] [user]

…

and /usr/bin is before /usr/xpg4/bin on the PATH. So the -g illegal
option results. No biggy.

Lidio.

List info: The rt-devel Archives

Best Practical is hiring! Come hack Perl for us:
Careers — Best Practical Solutions

12

Hi all.

The RC2 Search by CF is cool, but now i found another small
?? Bug. If you define a search and set the sort order to a
CF, then you get empty results, or better, RT displays xxx
found tickets, but don’t display something.

I’ve seen that the Field Metadata can’t match the SortField, so i’ve added
a:

CustomField => [ ‘CUSTOMFIELD’, ]

And CF decipher LoadByName if no queue was found so you can get always get
an cfid back to the sub:

if ($queue) {
…
}
else {
my $cf = RT::CustomField->new( $self->CurrentUser );
$cf->LoadByName( Name => $field );
$cfid = $cf->id if $cf;
}

In the OrderByCols function, if a cfid was found i’ve limited to the
specific CF… I don’t know if it’s useful but the result was sorted to the
only one CF and not to all which are found

unless ($cfid) {
…
}
else {
$self->SUPER::Limit(
ALIAS => $CFvs,
FIELD => ‘CustomField’,
VALUE => $cfid,
OPERATOR => ‘=’,
);
}

Regards,
Marius.