Skip Menu |

Preferred bug tracker

Please visit the preferred bug tracker to report your issue.

This queue is for tickets about the Crypt-Rijndael CPAN distribution.

Report information
The Basics
Id: 1444
Status: resolved
Priority: 0/
Queue: Crypt-Rijndael

People
Owner: Nobody in particular
Requestors: stratus [...] acm.org
Cc:
AdminCc:

Bug Information
Severity: Critical
Broken in: (no value)
Fixed in: (no value)



Subject: "make test" fails on alpha.
I'm the maintainer of Crypt::Rijndael at Debian (www.debian.org) according with Andreas U. Trottmann the 'make test' fails on alpha. For full information, you can see: bugs.debian.org/156779 Show quoted text
> -- System Information > Debian Release: 3.0 > Kernel Version: Linux clockwork 2.2.21 #1 Fri May 24 18:18:47 CEST 2002 alpha unknown > > Versions of the packages libcrypt-rijndael-perl depends on: > ii perl 5.6.1-7 Larry Wall's Practical Extraction and Report > ii perl-base 5.6.1-7 The Pathologically Eclectic Rubbish Lister. > ^^^ (Provides virtual package perlapi-5.6.1)
[guest - Thu Aug 15 09:49:39 2002]: In the meantime I had time to look further into the problem. It *seems* that it is a very "easy" problem to fix: The file rijndael.h in the source contains the line typedef unsigned long UINT32; This is wrong on alpha; it causes everything that is declared UINT32 in _rijndael.c to be of a 64 bit type. Changing it to typedef unsigned int UINT32; and recompiling the module makes it pass "make test", and also further encrypting/decrypting tests seem to be fine. Note that I'm no cryptoanalyst, and I do not know the Rijndael algorithm at all - but the fix looks "easy" and correct enough. Of course, to be really portable, you would need something different, which would define UINT32 depending on the architecture - but I think that at least all the architectures supported by Debian have an "unsigned int" type of 32 bits. -- Andreas Trottmann <andreas.trottmann@werft22.com>
v1.02 now supports 64 bit machines