headers
Chris Fields | 2 Jan 16:27
Favicon
Gravatar

Re: BioPerl 1.6 RC1


On Dec 26, 2008, at 4:07 PM, Sendu Bala wrote:

> Alex Lancaster wrote:
>> Some other oddness: when the Build.PL script checks for the
>> dependencies, it notes a few optional CPAN packages which aren't yet
>> packaged for Fedora such as 'Algorithm::Munkres' and 'Math::Random',
>> are missing.  Since they are optional and missing dependencies, I
>> would have expected that the relevant bioperl modules such
>> Bio::PhyloNetwork should also not be installed, but it appears that
>> the .pm files are installed, see lines like:
>> Copying Bio/PhyloNetwork/Factory.pm -> blib/lib/Bio/PhyloNetwork/ 
>> Factory.pm
>> from the full build.log:
>> https://koji.fedoraproject.org/koji/getfile?taskID=1022511&name=build.log
>> This causes problems because RPM then generates Requires such as:
>> Requires: perl(Algorithm::Munkres)
>> (see bottom of the build.log referred to above) which can't (yet) be
>> fullfilled by any other package in the Fedora package set, leading to
>> broken dependencies.
>> Is this the expected behaviour upon missing optional dependencies?
>
> Yes, though I can see how it's broken from your point of view. It is  
> set up to install everything, but simply not everything will work  
> without the optionals. I hadn't considered not installing the  
> modules that won't work.
>
> Something to think about, but it's probably not going to happen for  
> 1.6?
(Continue reading)

Permalink | Reply | 0 comments |
headers
Chris Fields | 2 Jan 16:40
Favicon
Gravatar

Re: BioPerl 1.6 RC1

Hilmar,

Bio::PhyloNetwork is now published, so I would assume the API has  
stabilized enough (not that publications are any metric):

http://www.ncbi.nlm.nih.gov/pubmed/18371228
http://www.ncbi.nlm.nih.gov/pubmed/18477576

I'm planning on releasing RC2 sometime in the next few days.  Does  
anyone think we should leave this out?

chris

On Dec 26, 2008, at 6:15 PM, Hilmar Lapp wrote:

>
> On Dec 26, 2008, at 4:39 PM, Alex Lancaster wrote:
>
>> it notes a few optional CPAN packages which aren't yet
>> packaged for Fedora such as 'Algorithm::Munkres' and 'Math::Random',
>> are missing.  Since they are optional and missing dependencies, I
>> would have expected that the relevant bioperl modules such
>> Bio::PhyloNetwork should also not be installed
>
>
> I may have failed to chime in here earlier at the appropriate  
> opportunity, but I would be very hesitant about putting  
> Bio::PhyloNetwork into the stable release.
>
> These modules have never been announced or discussed on the mailing
(Continue reading)

Permalink | Reply | 6 comments |
headers
Chris Fields | 2 Jan 21:15
Favicon
Gravatar

Re: BioPerl 1.6 RC1

I asked Lincoln about this but hadn't received a reply; oddly, I  
removed scripts/biographics and examples/biographics from trunk but  
the merge didn't remove them from the branch.  They won't be in RC2  
(Sun-Mon).

chris

On Dec 30, 2008, at 12:48 PM, Scott Cain wrote:

> Hi,
>
> For the scripts currently in BioPerl core that use BioGraphics, do we
> think that they should no longer be distributed with BioPerl but
> should instead be moved to the BioGraphics distribution?  I imagine
> someone trying to use bp_embl2picture.pl and being surprised that it
> doesn't work.
>
> Thoughts?
>
> Scott
>
>
> On Tue, Dec 30, 2008 at 11:16 AM, Christopher Fields
> <cjfields <at> illinois.edu> wrote:
>> All,
>>
>> I can't respond adequately until I return from vacation; I'm   
>> responding from a dial-up line in Texas (?!?) so responding to each  
>> message in kind will take a year or two (I did mention that I would  
>> be away from Dec 26-31, but it looks like that will be until Jan 1).
(Continue reading)

Permalink | Reply | 2 comments |
headers
Chris Fields | 2 Jan 21:34
Favicon
Gravatar

Re: BioPerl 1.6 RC1


On Dec 30, 2008, at 7:59 PM, Sendu Bala wrote:

> Christopher Fields wrote:
>>> Chris: I note that extraneous files like 'test.txt' and others made
>>> it into the RC1 .tar.gz you uploaded. What I always did was a clean
>>> export of the tag and built from there. BTW, the dist action also
>>> warns you about modules with their own version:
>>> Bio::DB::GFF::Aggregator::orf in this case. You might want to
>>> investigate that.
>> I noticed that it's packaging up everything in the local directory,
>> yes (that was after the upload unfortunately).  That'll be fixed for
>> RC2; I'll look for a more amenable fix when I get back (a packlist of
>> files would work around this, but I'm not sure how well that will
>> work with a large distro like BioPerl).
>
> No, no, it wouldn't work at all. The dist action is what /generates/  
> the 'packlist' (MANIFEST) - we don't want to have to manually  
> specify anything because it would be too easy to forget to add new  
> files and there would be no obvious indication anything was left out  
> of a distribution. It's very important this remain automatic and  
> fool proof.
>
> It should package up everything under version control. Only if there  
> are things under version control that we don't want packaged should / 
> exceptions/ manually be added (MANIFEST.SKIP, ne Build.PL).
>
> Again, all you have to and should do is build from a fresh export of  
> the tag, not from your normal working directory. You should never do  
> that in any case, because you might have been working on a module
(Continue reading)

Permalink | Reply | 0 comments |
headers
Alex Lancaster | 2 Jan 21:53
Picon
Gravatar

Re: BioPerl 1.6 RC1

>>>>> "CF" == Christopher Fields  writes:

[...]

>>> If the Bio::Graphics is truly not needed, for the moment it is
>>> possible to also override and filter out the bogus Requires until
>>> such time as these scripts are moved to the appropriate place.
>> 
>> Great, go ahead and do that if you like.

CF> I'm not quite sure why we are attempting to RPM package up a
CF> release candidate unless it's strictly for the purposes of testing
CF> things out.  I anticipate the final 1.6 will be out in short
CF> period of time (within a few weeks).  Maybe this has already been
CF> answered, just haven't had time to read back along the thread yet.

The Fedora package for BioPerl is strictly being updated on the
development branch (aka "rawhide") of Fedora and is not being added to
any officially released "stable" versions of Fedora (i.e. F-9 and F-10
are staying at 1.5.2 until 1.6.x final is released and gets some
testing in rawhide).

In Fedora, it's common practice to package up release candidates on
the development branch as it shakes out problems with packaging and
can even (as it has in this case) identify issues in the upstream
tarballs.  In turn, that means that those problems can be fixed
upstream before the final release, rather than having to fix things in
the package by patching or otherwise modifying the upstream source
etc.
(Continue reading)

Permalink | Reply | 1 comment |
headers
Sendu Bala | 2 Jan 22:19
Picon
Favicon

Re: BioPerl 1.6 RC1

Alex Lancaster wrote:
> I understand that this is a RC, not a final release, which is
> why I'm reporting them now so hopefully things will work smoothly when
> packaging BioPerl as an RPM in the final release.

Yeah, thanks Alex, we appreciate it.
Permalink | Reply | 0 comments |
headers
Hilmar Lapp | 2 Jan 23:42
Picon
Gravatar

Re: BioPerl 1.6 RC1


On Jan 2, 2009, at 10:40 AM, Chris Fields wrote:

> Bio::PhyloNetwork is now published, so I would assume the API has  
> stabilized enough (not that publications are any metric):

Right. I was referring to its integration into BioPerl, not the  
publication.

	-hilmar

--

-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:-  hlapp at gmx dot net :
===========================================================
Permalink | Reply | 0 comments |
headers
Hilmar Lapp | 2 Jan 23:52
Picon
Gravatar

Re: BioPerl 1.6 RC1


On Jan 2, 2009, at 10:40 AM, Chris Fields wrote:

> Bio::PhyloNetwork [...]  Does anyone think we should leave this out?

I would rephrase the question. I think it's a very valuable addition  
to BioPerl, and the above may be understood as a vote on that, which  
AFAIAC is not a vote we need to have.

Instead, I would ask the following. Generally, i) are there any  
opinions on whether the Bio::XXX root namespace should be permissively  
expanded, and ii) should new modules that have not been reviewed yet  
by core devs be included in a stable release. Specifically with  
respect to Bio::PhyloNetwork, are there opinions on i) moving or not  
moving this to the Bio::Phylo::Network namespace, and on ii)  
harmonizing or not the API as much as possible with the Bio::Tree APIs.

(Chris - you would probably agree that the publication neither answers  
the above questions, nor guarantees for the API's stability.)

	-hilmar
--

-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:-  hlapp at gmx dot net :
===========================================================
Permalink | Reply | 4 comments |
headers
Lincoln Stein | 3 Jan 16:51
Picon
Gravatar

Re: BioPerl 1.6 RC1

Sorry, what's the question? Anything having to do with biographics should be
removed as it now has its own separately installable CPAN module.

Lincoln

On Fri, Jan 2, 2009 at 3:15 PM, Chris Fields <cjfields <at> illinois.edu> wrote:

> I asked Lincoln about this but hadn't received a reply; oddly, I removed
> scripts/biographics and examples/biographics from trunk but the merge didn't
> remove them from the branch.  They won't be in RC2 (Sun-Mon).
>
> chris
>
>
> On Dec 30, 2008, at 12:48 PM, Scott Cain wrote:
>
>  Hi,
>>
>> For the scripts currently in BioPerl core that use BioGraphics, do we
>> think that they should no longer be distributed with BioPerl but
>> should instead be moved to the BioGraphics distribution?  I imagine
>> someone trying to use bp_embl2picture.pl and being surprised that it
>> doesn't work.
>>
>> Thoughts?
>>
>> Scott
>>
>>
>> On Tue, Dec 30, 2008 at 11:16 AM, Christopher Fields
(Continue reading)

Permalink | Reply | 1 comment |
headers
Chris Fields | 3 Jan 18:39
Favicon
Gravatar

Re: BioPerl 1.6 RC1

Lincoln,

There were several scripts and examples in bioperl-live which have  
been removed but somehow persisted in the branch and were in 1.6 RC1  
(a test also remained which was also removed, t/Graphics/ 
Pictogram.t).  I didn't know if you wanted these moved to Sourceforge;  
I saw there were several examples already in the Bio::Graphics  
repository.

There were a few other modules in Bio::Graphics namespace moved over  
to Sourceforge that I wasn't sure whether you wanted to maintain, such  
as DrawTransmembrane.pm and Pictogram.pm.  If needed we can resurrect  
them in trunk, under either Bio::Graphics or a different namespace  
(latter is probably better, any suggestions?).  They don't appear to  
rely on other Bio::Graphics modules directly.

chris

On Jan 3, 2009, at 9:51 AM, Lincoln Stein wrote:

> Sorry, what's the question? Anything having to do with biographics  
> should be
> removed as it now has its own separately installable CPAN module.
>
> Lincoln
>
> On Fri, Jan 2, 2009 at 3:15 PM, Chris Fields <cjfields <at> illinois.edu>  
> wrote:
>
>> I asked Lincoln about this but hadn't received a reply; oddly, I
(Continue reading)

Permalink | Reply | 0 comments |

Gmane