|
Posted by Stig Brautaset on August 1, 2004, 2:25 pm
Please log in for more thread options
CPAN-testers reports some failures[0] for a module I've written. I want
to fix these, but I must admit I'm a bit stumped. I can't see why this
should fail.
[0] http://testers.cpan.org/show/Games-AlphaBeta.html#Games-AlphaBeta-0.2.0
Any pointers?
Stig
|
|
Posted by Scott W Gifford on August 4, 2004, 6:09 pm
Please log in for more thread options
> CPAN-testers reports some failures[0] for a module I've written. I want
> to fix these, but I must admit I'm a bit stumped. I can't see why this
> should fail.
>
> [0] http://testers.cpan.org/show/Games-AlphaBeta.html#Games-AlphaBeta-0.2.0
>
> Any pointers?
First, when I asked a similar question on PerlMonks, the consensus was
the CPANPLUS often handled dependencies incorrectly, and that warnings
about dependencies should be ignored:
http://www.perlmonks.org/index.pl?node_id=372922
The error messages for your module don't look like dependency
problems, though.
One useful thing is to see how the environments where it's failing are
different from yours. For example, a different version of Perl, a
different endianness if you're doing the sorts of things that depend
on that, etc. It's straightforward to install a different version of
Perl in a temporary location and see how your module works with it;
finding a machine of a different endianness might be a little
trickier.
Another useful thing is to see what it's dying from (often undefined
or empty variables, it looks like) and add assert-style checks for
that in various places in your program. That way if it dies at least
you'll know where. Something like:
defined($a) or die "$a not defined";
inserted in appropriate places would tell you when a variable first
became undefined, for example.
-----ScottG.
|
|
Posted by Stig Brautaset on August 6, 2004, 7:38 am
Please log in for more thread options >> CPAN-testers reports some failures[0] for a module I've written. I
>> want to fix these, but I must admit I'm a bit stumped. I can't see
>> why this should fail.
>>
>> [0] http://testers.cpan.org/show/Games-AlphaBeta.html#Games-AlphaBeta-0.2.0
>
> First, when I asked a similar question on PerlMonks, the consensus was
> the CPANPLUS often handled dependencies incorrectly, and that warnings
> about dependencies should be ignored:
>
> http://www.perlmonks.org/index.pl?node_id=372922 >
> The error messages for your module don't look like dependency
> problems, though.
I think they might be. The module depends on the latest version of a
different module. If instead an earlier version of this module was used
this problem could well occur, I believe.
> on that, etc. It's straightforward to install a different version of
> Perl in a temporary location and see how your module works with it;
> finding a machine of a different endianness might be a little
> trickier.
I've talked to one of the testers where the tests failed, and he can not
reproduce it. I'm starting to suspect the module is actually ok.
Stig
|
|
Posted by Brian McCauley on August 4, 2004, 8:43 pm
Please log in for more thread options
> CPAN-testers reports some failures[0] for a module I've written. I want
> to fix these, but I must admit I'm a bit stumped. I can't see why this
> should fail.
>
> [0] http://testers.cpan.org/show/Games-AlphaBeta.html#Games-AlphaBeta-0.2.0
>
> Any pointers?
The obvious thing that springs to mind is the possibility that you are
not specifying the correct minimum versions in the prerequisites but
having looked at the source of this module I can't convince myself
that this is the case.
Note: I think module numbers are compared as pure numbers. This means
that 0.2.0 and 0.2.3 are noor distinct.
--
\ ( )
. _\__[oo
.__/ \ /@
. l___\
# ll l\
###LL LL\
|
|
Posted by Stig Brautaset on August 5, 2004, 10:53 am
Please log in for more thread options >> CPAN-testers reports some failures[0] for a module I've written. I want
>> to fix these, but I must admit I'm a bit stumped. I can't see why this
>> should fail.
>>
>> [0] http://testers.cpan.org/show/Games-AlphaBeta.html#Games-AlphaBeta-0.2.0
>>
>> Any pointers?
>
> The obvious thing that springs to mind is the possibility that you are
> not specifying the correct minimum versions in the prerequisites but
> having looked at the source of this module I can't convince myself
> that this is the case.
I've been in contact with one of the testers that had a failure. He said
he now cannot reproduce it, but that the first time he tested it some
dependency problems were reported. The module depends on the latest
version of a different module I've written, and if that was not
installed I can well imagine that this error would occur.
> Note: I think module numbers are compared as pure numbers. This means
> that 0.2.0 and 0.2.3 are noor distinct.
I was of the impression that this would be fine, as long as the version
numbers where quoted (but maybe I've missinterpreted
http://search.cpan.org/~jpeacock/version-0.41/lib/version.pm#Quoted_Versions
?).
Stig
|
| Similar Threads | Posted | | GD make test failure | January 16, 2005, 5:11 pm |
| Sudo.pm test failure | April 4, 2008, 2:37 pm |
| ANNOUNCE: Games::Go::AGATourn and Games::Go::GoPair modules released | January 25, 2005, 9:30 pm |
| problem using Games::OpenGL::Font::2D and Games::Console | February 13, 2007, 8:47 pm |
| perl upgrade, req to understand which modules to install | October 2, 2006, 2:54 pm |
| PERL/CGI Programmer wanted | October 31, 2005, 9:14 pm |
| Name wanted for module in construction | March 15, 2006, 6:21 pm |
| AIX testers wanted for Math::Random | July 11, 2006, 1:55 pm |
| [RFC] Games::Sudoku::General | November 30, 2005, 5:58 pm |
| Peer-reviewed CPAN modules files wanted | August 21, 2006, 3:13 am |
|