Thanks Chris,

About PGPLOT issue, I really don't use it, I just installed it because was
in the recommended part of http://pdl.perl.org/?page=install.


On Wed, Dec 31, 2014 at 3:48 PM, Chris Marshall <[email protected]>
wrote:

> On Mon, Dec 22, 2014 at 1:06 AM, Pablo marin-garcia <
> [email protected]> wrote:
>
>> Hello Chris,
>>
>> I have take a look at the the failure matrix  of Module::Compile[1]  in
>> the automatic testing and it is a bit worrying for having this module as a
>> dependency. Probably the poor score is because the undefined dependency on
>> Module::Install but anyway too much RED in there. It would be nice to add a
>> note in the PDL install page about it, just in case.
>>
>> [1] http://matrix.cpantesters.org/?dist=Module-Compile%200.34
>>
>
> This problem was reported and has been addressed in the latest version of
> Module::Compile.  Thanks mohawk and ingy.
>
>
>> One of the reason I use R instead PDL in production software that I need
>> to distribute, even for simple things that would be more natural to use
>> PDL, is because is easier to install R and bioconductor as not root than
>> PDL. Nowadays things has been improved a lot, and I think that for next
>> release we can help to keep improving it if we make a 'install hackaton'
>> and see how much success we have with a simple 'cpanm PDL' in different
>> environments.  Many of the issues are not with PDL itself but its
>> dependencies, so having a list of common pitfalls and its workarounds would
>> be helpful.
>>
>
> I'm sorry to hear about your difficulties.  PDL builds out of the box on
> the platforms I have access to but the external dependencies are still a
> work in progress since different platforms have different requirements.  If
> you have a specific set of problems with distributing PDL, please do work
> with the perldl lists to resolve the problem and open a ticket for the
> issue on our bug tracker at http://sourceforge.net/p/pdl/bugs/
>
>
>>
>> For example a recent 'issue' I have was having perl module for PGPLOT
>> installed:
>> http://stackoverflow.com/questions/27585591/how-to-avoid-cpanm-stopping-installation-waiting-for-a-prompt-that-is-shown-only
>>
>
> I can't help with PGPLOT as I am unable to build it on cygwin.  There are
> other plotting options for PDL which may simpler to provide in a
> distribution:  PDL::Graphics::Gnuplot, PDL::Graphics::Prima come to mind.
> The latest version of PLplot now builds out of the box on cygwin so I plan
> to update PDL::Graphics::PLplot as time permits.
>
> Regards,
> Chris
>
>
>>
>>
>> On Mon, Dec 22, 2014 at 1:44 AM, Chris Marshall <[email protected]>
>> wrote:
>>
>>> Thanks for the feedback, Pablo.  The warning about another PDL in your
>>> path is because you can get test failures if the new install has a
>>> different configuration than the previous install.  In my experience so
>>> far, that has meant false failures in that the new install is good for its
>>> part.  At some point it might make sense to "hide" the previous install for
>>> testing but that hasn't happened yet.  The warning is an intermediate step.
>>>
>>> I see you've found bug ticket #365 regarding the recent Module::Compile
>>> regression.  This is definitely one of the things to be fixed for the
>>> PDL-2.008 release this January.
>>>
>>> --Chris
>>>
>>>
>>> On Sun, Dec 21, 2014 at 6:49 PM, Pablo marin-garcia <
>>> [email protected]> wrote:
>>>
>>>> Hello Chris and PDLers
>>>>
>>>> Just as you are talking about a new release and we would need to
>>>> update/install it, I would like to talk about a recent issue I had
>>>> installing PDL.
>>>>
>>>>  I had 'issues' with some perl dependencies with installing cpanm
>>>> PGPLOT and cpanm Module::Compile but I will I report it in a blog entry
>>>> soon so I could help to improve the documentation.
>>>>
>>>> Here I would like to ask a very specific question:  If you have already
>>>> a PDL installed and you install a new version a warning appears:
>>>>
>>>> --------------------------------------------------
>>>> Unpacking PDL-2.007.tar.gz
>>>> Entering PDL-2.007
>>>> Checking configure dependencies from META.yml
>>>> Checking if you have Devel::CheckLib 1.01 ... Yes (1.02)
>>>> Configuring PDL-2.007
>>>> Running Makefile.PL
>>>>
>>>> +------- WARNING ------------------------------------------------
>>>> | PDL version 2.4.11 was detected in your @INC
>>>> | PDL may not build/test correctly with another PDL in @INC!
>>>> +----------------------------------------------------------------
>>>>
>>>> ------------------------------------
>>>>
>>>> This mean that I need manually to remove my previous PDL installtion
>>>> first?
>>>> How do I remove my old vesion of @INC?, you mean I need to rm -rf it?.
>>>> Could this message be improved and hints how to avoid/fix it be added?
>>>>
>>>> Anyway, as this is a warning seems not to impede the the installation:
>>>>
>>>> --------------
>>>> Successfully installed PDL-2.007 (upgraded from 2.4.11)
>>>> Installing
>>>> ~/perl5/perlbrew/perls/perl-5.16.0/lib/site_perl/5.16.0/x86_64-linux/.meta/PDL-2.007/install.json
>>>> Installing
>>>> ~/perl5/perlbrew/perls/perl-5.16.0/lib/site_perl/5.16.0/x86_64-linux/.meta/PDL-2.007/MYMETA.json
>>>> 1 distribution installed
>>>> -----------------
>>>>
>>>> Note: a dependency of PDL is Module::Compile, but this does not get
>>>> installed with cpanm because one of its dependencies it is not stated in
>>>> the package dependencies: Module::Install. This is a known bug
>>>> https://github.com/ingydotnet/module-compile-pm/issues/1 but in my
>>>> case was more cryptic than the ones reported:
>>>>
>>>> | PERL_DL_NONLAZY=1 ~/perl5/perlbrew/perls/perl-5.16.0/bin/perl
>>>> "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')"
>>>> t/*.t
>>>> | t/000-compile-modules.t .. 1/?
>>>> | #   Failed test 'use Module::Compile;'
>>>> | #   at t/000-compile-modules.t line 11.
>>>> | #     Tried to use 'Module::Compile'.
>>>> | #     Error:  Can't open (eval 6) for input:   ###<<<<<<=======
>>>> | # No such file or directory at
>>>> ~/.cpanm/work/1419122233.15164/Module-Compile-0.34/blib/lib/Module/Compile.pm
>>>> line 201.
>>>> | # BEGIN failed--compilation aborted at (eval 6) line 2.
>>>> |
>>>> | #   Failed test 'use Module::Optimize;'
>>>> | #   at t/000-compile-modules.t line 11.
>>>> | #     Tried to use 'Module::Optimize'.
>>>> | #     Error:  Can't open (eval 7) for input:
>>>> | # No such file or directory at
>>>> ~/.cpanm/work/1419122233.15164/Module-Compile-0.34/blib/lib/Module/Compile.pm
>>>> line 201.
>>>> | # BEGIN failed--compilation aborted at (eval 7) line 2.
>>>> | # Looks like you failed 2 tests of 5.
>>>>
>>>> In my errors it was not mention what was the problem, looking at bugs
>>>> opened I saw that the a missed dependency was the culprit Module::Install.
>>>> I installed it and all went fine. I mention this here just in case you
>>>> want to add this as a NOTE in the web page for PDL installation.
>>>>
>>>> Pablo
>>>>
>>>>
>>>>
>>>>
>>>> On Sat, Dec 20, 2014 at 9:41 PM, Chris Marshall <[email protected]
>>>> > wrote:
>>>>
>>>>> All-
>>>>>
>>>>> I've just reviewed the status of the bug and feature request trackers
>>>>> at the PDL site on sf.net:
>>>>>
>>>>>   http://sourceforge.net/p/pdl/bugs/
>>>>>   http://sourceforge.net/p/pdl/feature-requests/
>>>>>
>>>>> I found a number of bug tickets that had already been fixed and
>>>>> determined that there are a fair number of tickets that are either low
>>>>> hanging fruit or already fixed but need to be verified to close.
>>>>>
>>>>> I had hoped that the 64bit support could have been finished by now but
>>>>> lack of developers to debug the final issue and time for me to get back to
>>>>> it, has it still in limbo.  We have a number of nice improvements in the
>>>>> hopper already and it has been a little more than a year since PDL-2.007.
>>>>>
>>>>> I propose we make a bug fix/clean up/... release of PDL-2.008 for
>>>>> January 2015---regardless of whether the remaining 64bit index support is
>>>>> finished.  I encourage all users and developers to take this opportunity 
>>>>> to
>>>>> take a look at the outstanding bugs and close them.  There are a number of
>>>>> feature requests that relate to some items such as improved PDL docs
>>>>> handling that could also be addressed for this release.
>>>>>
>>>>> And, of course, I would love some assistance to finish the
>>>>> implementation debugging for the PDL-2.x 64bit index support.  :-)
>>>>>
>>>>> As always, thoughts and suggestions welcome!
>>>>> Chris
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Perldl mailing list
>>>>> [email protected]
>>>>> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
>>>>>
>>>>>
>>>>
>>>
>>
>
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to