Encoding error

RT Users
1

Hello,

With RT 3.8.7, we are getting relatively often (about 1 out of 5
attachments) the following encoding error:

Can’t call method “add_suspects” on unblessed reference at
/usr/lib/perl5/5.8.8/i386-linux-thread-multi/Encode/Guess.pm line 36.

The intriguing part is that resubmitting the attachment works most of the time.

Any suggestions on how to address that?

We are going to upgrade to 3.8.10, hoping that this will solve this issue.

Thanks,
Thierry Thelliez

2

With RT 3.8.7, we are getting relatively often (about 1 out of 5
attachments) the following encoding error:

Can’t call method “add_suspects” on unblessed reference at
/usr/lib/perl5/5.8.8/i386-linux-thread-multi/Encode/Guess.pm line 36.

The intriguing part is that resubmitting the attachment works most of the time.

Any suggestions on how to address that?

We are going to upgrade to 3.8.10, hoping that this will solve this issue.

rt$ git checkout rt-3.8.7
rt$ grep Encode sbin/rt-test-dependencies.in
Encode 2.21
perl$ git checkout perl-5.8.8
perl$ grep VERSION ext/Encode/Encode.pm
our $VERSION = sprintf “%d.%02d”, q$Revision: 2.12 $ =~ /(\d+)/g;
XSLoader::load(PACKAGE, $VERSION);

I suspect you need to install a newer version of Encode.

Dominic Hargreaves, Systems Development and Support Team
Computing Services, University of Oxford

signature.asc (198 Bytes)

3

Dominic,

Thank for your answer.

We check our current version and we are at 2.43…

grep VERSION Encode.pm

our $VERSION = sprintf “%d.%02d”, q$Revision: 2.43 $ =~ /(\d+)/g;
XSLoader::load( PACKAGE, $VERSION );

There is a 2.44 version but the release notes (
http://cpansearch.perl.org/src/DANKOGAI/Encode-2.44/Changes) does not
appear to have addressed this issue. In fact I do not know if the
problem is with Encode, Perl, or with RT…

I looked at the current bug list
(Active bugs for Encode), but
nothing similar has been reported recently.

Thierry

4

Just to follow up, we migrated from 3.8.7 to 3.8.10 and that fixed our issue.

The Encode library stayed at the same version number (2.43). I am
guessing that the problem was in some RT code.

Anyway, we are not seeing the error anymore.

Cheers,
Thierry