Re: How to submit software for addition to BioLinux?
Thanks Tim & Andreas,
Your words are very encouraging :-) When I checked in detail I found the
dependency situation is not as bad as I thought. I think these are the
minimal dependencies that are missing or not sufficient:
groovy 2.3.4 : exists already but is only version 2.0.0~beta2
gpars 1.2.1 : exists already, but not sufficient version
extra166y : not in debian, is in fedora
smack 3.2.2 : not in debian, is in fedora
The good news is that only two libraries are not existing at all, and
those are already in Fedora. Is it helpful if there is a src rpm or is it
just as easy to start from scratch?
Many thanks for all your advice and help -
On 18/02/2015 5:37 am, "Tim Booth" <firstname.lastname@example.org> wrote:
>As you've probably guessed, the correct answer is (A) but it might not
>be as bad as you think because:
>1) Most small Java libraries are actually very easy to package
>2) Most updates to small Java libraries are trivial to push through
>3) You are not alone - both Debian Med and Debian Java people will help
>with the workload
>4) You can do your plan B as a beginning, so that you do actually work
>on the Bpipe package first. Don't worry about compiling from source
>here, just round up the binary JARs and dump them directly in the source
>package, then once Bpipe is packaged up and working we worry about
>splitting out the dependencies.
>What could be a showstopper is if you rely on a JAR for which there is
>no source or the source is under a discriminatory license. If any part
>of your software is non-free then the whole thing is non-free.
>Also some Java software depends on a very particular older version of a
>library because it's essentially bug-compatible. Debian frowns on this.
>(Actually, anyone with any sense should frown on this). Compatibility
>should be determined by regression tests, rather than version numbers,
>so that library packages can be safely upgraded.
>So, first job is to make a list of what those libraries are and the
>versions required so we can at least triage them all - similar to what
>they did with OpenClinica. Do you have that list?
>On Tue, 2015-02-17 at 09:32 +0000, Simon Sadedin wrote:
>> Hi all,
>> I guess I have a ³simple² question but it might be a bit of a
>> - Bpipe relies on a number of java libraries that are fetched in binary
>> form from online repositories at build time. If I understand Debian
>> correctly, that is not allowable. I checked through the packages and a
>> number of them seem to be either unavailable in Debian or they are
>> available but not with recent enough versions.
>> If I understand the situation right, my options would be -
>> - somehow get all these third party libraries submitted / updated into
>> - package the source for all these libraries inside the Bpipe package
>> have the package build all the dependent libraries it needs as part of
>> own build.
>> The first I am afraid would be far beyond the time I have to commit. The
>> second is probably also quite a lot of work, but its more feasible.
>> Are these my only options?
>> NB: this seems to be a similar situation to OpenClinica a few years ago.
>> Thread here:
>> Cheers / thanks!
>> On 6/02/2015 7:51 am, "Andreas Tille" <email@example.com> wrote:
>> >Hi Simon,
>> >I'm writing you as a member of the Debian Med team.
>> >On Thu, Feb 05, 2015 at 06:08:49PM +0000, Tim Booth wrote:
>> >> It's great to hear that you are using Bio-Linux and want to
>> >> your Bpipe package. We also have a need for an effective lightweight
>> >> pipeline system that supports clusters and doesn't rely on a web
>> >> (Galaxy) or a mountain of Java dependencies and remote services
>> >> (Taverna).
>> >> What we can do do is to submit the package to Debian Med
>> >> (https://wiki.debian.org/DebianMed) which is what I do with all my
>> >> packages where possible. Putting the package into Debian makes it
>> >> appear in Debian, Ubuntu, Mint and of course Bio-Linux so it's a big
>> >> win.
>> >> To have a Debian package one needs first to make the initial package
>> >> then to maintain it as new versions come out or bugs are reported.
>> >> first bit is hard for you but easy for us. The second bit is where
>> >> can make it much easier for us by being the primary maintainer in
>> >> and handling necessary updates. (Don't worry that you are using
>> >> Bio-Linux rather than an Debian system - for Java apps it makes no
>> >> difference.)
>> >> To this end I've copied this to the Debian Med list. Someone on the
>> >> list might fancy picking up the package - if not I'll have a look at
>> >> myself but it will be a couple of weeks before I have time as I'm
>> >> on various other jobs. Either way, we'll take you through the
>> >> packaging then make sure you are set up to be the primary maintainer.
>> >> It's a bit of learning for you but you won't have to swallow the
>> >> 106 page Debian packagers' policy manual ;-) Does that sound an OK
>> >> plan?
>> >I could also offer to help you getting started in a so called Mentoring
>> >of the Month project:
>> > https://wiki.debian.org/DebianMed/MoM
>> >In short we would be really happy if you could help you to get the
>> >software you need into Debian and its derivatives.
>> >Kind regards
>> > Andreas.
>> >This email has been scanned by the Symantec Email Security.cloud
>> >For more information please visit http://www.symanteccloud.com
>> >If you have any question, please contact MCRI IT Helpdesk for further
>> This email has been scanned by the Symantec Email Security.cloud
>> For more information please visit http://www.symanteccloud.com
>Tim Booth <firstname.lastname@example.org>
>NERC Environmental Bioinformatics Centre
>Centre for Ecology and Hydrology
>Maclean Bldg, Benson Lane
>+44 1491 69 2297
>This email has been scanned by the Symantec Email Security.cloud service.
>For more information please visit http://www.symanteccloud.com
>If you have any question, please contact MCRI IT Helpdesk for further
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com