[Biojava-dev] sorry about biojava-live/trunk repository mess; was Re: biojava-legacy 1.8.2 release plan

[Andreas Prlic]

Hi Michael,

I would tend to split as well. Are you and Hannes in agreement in how
to proceed with this or is there anything we should discuss?  Let me
know if I can help somehow in resolving this.

What is the benefit of the Guava library? a quick check suggests it
contains mainly utilities. Unless it is a significant time-saver I
would vote for providing similar utility code as part of biojava and
avoid dependency creep. We have a couple of module specific
dependencies, but they are mostly adding features that are
complementary, like the forester library or Jmol.

Andreas



On Tue, Mar 13, 2012 at 8:56 AM, Michael Heuer <heuermh at gmail.com> wrote:
> All,
>
> In addition to committing changes to biojava-legacy/trunk for the
> 1.8.2 release, I have now merged the updated fastq code from
> biojava-legacy/trunk to biojava-live/trunk for the upcoming 3.0.3
> release.  In the process I have made a mess of the revision history
> for biojava-live/trunk, since I was not doing a good job coordinating
> with Hannes' changes.
>
> (see rev 9567 in particular)
> http://code.open-bio.org/biojava-live/trunk/biojava3-genome/?root=biojava&view=log
> http://code.open-bio.org/biojava-live/trunk/biojava3-sequencing/?root=biojava&view=log
>
> At present there are two copies of the fastq package, one in
> org.biojava3.sequencing.io.fastq in a re-added biojava3-sequencing
> module, and one in org.biojava3.genome.parsers.fastq in
> biojava3-genome.  Before the 3.0.3 we need to resolve this.
>
> I'm more of a splitter than a lumper when it comes to maven modules;
> the pros of keeping biojava3-sequencing as a separate module
>
>  - self-contained package, only dependent on biojava3-core
>  - same API and module structure as biojava-legacy (although different package)
>  - biojava3-sequencing adds a new third party compile scope dependency
> (on Google Guava, version 11.0.1)
>  - would be the home for Phred, ABI, and similar code if/when migrated
> from biojava-legacy
>
> and the cons
>
>  - narrow scope feature-wise for a module
>  - FASTQ parsing is not all that different from FASTA parsing
> conceptually, although no code is shared
>
> Please let me know what you think.
>
> Sorry about creating this mess.  Since I have re-added a directory
> that had been removed in an earlier commit, you may receive a
> complaint from subversion about biojava3-sequencing on update.
>
>   michael
>
>
> On Thu, Mar 8, 2012 at 11:31 PM, Michael Heuer <heuermh at gmail.com> wrote:
>> ok, scratch my previous plan.  I have committed the new fastq code to
>> biojava-legacy for the 1.8.2 release and started this legacy cookbook
>> page
>>
>> http://biojava.org/wiki/BioJava:Cookbook:SeqIO:FASTQ
>>
>> Please let me know what you think.  I have performance benchmark
>> numbers if anyone needs convincing that this API is an improvement
>> over the current one.
>>
>>   michael
>>
>>
>> On Thu, Mar 8, 2012 at 1:33 PM, Michael Heuer <heuermh at gmail.com> wrote:
>>> Andreas Prlic wrote:
>>>
>>>> thanks for volunteering for making a biojava-1.x release! Let's do
>>>> this in sync together with the 3.x series. What about doing the
>>>> releases next Friday, March 16th?
>>>
>>> Sure, that works for me.
>>>
>>>
>>>>> I assume the documentation on cutting a release at
>>>>> http://biojava.org/wiki/BioJava:Make_release
>>>>> is still relevant?
>>>>
>>>> In fact this is not up-to-date any more.
>>>>
>>>> The code is now Mavenized and the key-task for the person making the
>>>> release is to set up ssh-keys to the portal-obf server prior to making
>>>> this release. After that you can make a release as simply as that:
>>>>
>>>> mvn release:clean release:prepare release:perform
>>>
>>> Much better.
>>>
>>>   michael
>
> _______________________________________________
> biojava-dev mailing list
> biojava-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/biojava-dev