[Bioperl-l] Can we make Clone a required module so that automatic installation works?

[Fields, Christopher J]

Yes (posts to Google should go to the list).

Frankly, we should simply move to google groups, lists.open-bio.org<http://lists.open-bio.org> is on life support at the moment.  Of course we would do that and Google would drop support for it ala Google Code :P

chris

On Mar 17, 2014, at 2:07 PM, Scott Cain <scott at scottcain.net<mailto:scott at scottcain.net>> wrote:

Hi Chris,

The 5 or 6 emails that all came within a few minutes of each other were from me going to the bioperl-l moderation page and allowing them through (a few were from February; guess I haven't done that in a while).  Are they getting posted to the google list but not bioperl-l, and they're supposed to?

Scott



On Mon, Mar 17, 2014 at 2:49 PM, Fields, Christopher J <cjfields at illinois.edu<mailto:cjfields at illinois.edu>> wrote:
Huh, for some reason the bioperl-l google group replies aren’t being bounced back to the list.

Okay, I’ll address each point below...

On Feb 26, 2014, at 11:40 AM, george.hartzell at gmail.com<mailto:george.hartzell at gmail.com> wrote:

> In short, I can't do this:
>
>  cpanm -L FOO BioPerl-1.6.923.tar.gz
>
> which makes me Very Sad.
>
> I've followed up on an existing issue about this here:
>
>  https://redmine.open-bio.org/issues/3447
>
> A hacked version of the tarball that moves Clone from a recommended module
> to a required module (Build.PL, META.{json,yaml}) lets all of our tests
> pass and keeps me from being Very Sad.
>
>
>   - What's the right fix?

Make Clone required (now committed in the 1.6.x branch).  On master branch it makes no sense to do this, as the requirement is at the Bio::Root level (and there is a req for Bio::Root in place now).  So, this will also be added to the Bio-Root repo.

>   - Is our test broken?

Not sure; the failure is an indication that Storable::dclone is problematic for cloning instances.  There is a possible fix (I alluded to it in the reply that didn’t go to the list), but really, it would be much simpler to just require Clone instead.

>   - Is there some reason to not require Clone (particularly if we actually
>   require it to function correctly?

Mentioned above.

>   - Should we prefer Clone::Fast?

Not sure; as I also mentioned (but will reiterate here for posterity) there may be OS-related problems.

Also (and somewhat strangely) Clone::Fast is only on search.cpan.org<http://search.cpan.org/> and not on metacpan, which makes me wonder about it’s status; there is a replacement called Data::Clone that is supposed to be faster and that no longer mentioned Clone::Fast.  Maybe it’s been removed?

That might simplify things for the time being...

>   - If we want to present the Bio::* consumer with a choice, is driving
>   that choice by whatever cloner happens to have been dragged in from God
>   Knows Where the best way to do it?

It should minimally work with something, and work better with something else.  That something should be (minimally) Clone, and if Clone::Fast is present then it can be used instead.

Maybe we can simply default to Clone, and if a clone class is specified (via a global or env variable) use that instead.

> So many questions....
>
> Thoughts, comments?
>
> g.

For some context: the use of ‘clone’ here was meant to address a missing core function, namely a way to simply copy an instance (and possibly modify the copy on the fly if needed).  So far it has worked with only a few hiccups.

chris


_______________________________________________
Bioperl-l mailing list
Bioperl-l at lists.open-bio.org<mailto:Bioperl-l at lists.open-bio.org>
http://lists.open-bio.org/mailman/listinfo/bioperl-l



--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research