Re: Tophat 2.0.6 is ready for upload

2012-11-03 Thread Andreas Tille
Hi Carlos,

could you please mention whether one of those bugs below could be
considered as "Release Critical"?  If yes, we should file an apropriate
bug to the Debian BTS.   Otherwise we should probably consider an upload
to experimental due to the rules in the times of freeze.

Kind regards and thanks for your preparations

 Andreas.

On Fri, Nov 02, 2012 at 10:38:39PM -0400, Carlos Borroto wrote:
> Hi,
> 
> New version Tophat 2.0.6 is ready for upload. I used
> debian/get-orig-source to make sure SeqAn was removed correctly.
> 
> This is a bug fix realease. Upstream changelog:
> TopHat 2.0.6 release 11/02/2012
> Version 2.0.6 is a maintenance release addressing some issues found in
> the 2.0.5 release:
> - corrected the indel finding algorithm that caused segmentation fault
> in certain cases (long_spanning_reads and tophat_reports)
> - fixed the Bowtie version checking code, adding support for newer,
> non-beta Bowtie2 versions
> - several minor fixes in the fusion alignment algorithm
> - fixed an incompatibility issue with Python versions older than 2.6
> (restoring Python 2.4 compatibility)
> - fixed and improved the resuming option (-R/--resume) to better
> handle various failure/resume situations
> - added a warning about Bowtie1 and Bowtie2 index files in the same
> directory (causing trouble if they were built for different genomic
> sequences)
> 
> Thanks,
> Carlos
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
> Archive: 
> http://lists.debian.org/cabgghblnplnktv23q5mkxwauxf9ilpawasmzrkggbsza6zw...@mail.gmail.com
> 
> 

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121103071154.ga7...@an3as.eu



Re: Packaging Ray for Debian Med

2012-11-03 Thread Andreas Tille
Hi Sébastien,

many thanks for you preparations.

On Fri, Nov 02, 2012 at 05:09:10PM -0400, Sébastien Boisvert wrote:
> The Ray Debian package is now at
> http://anonscm.debian.org/gitweb/?p=debian-med/ray.git;a=summary

I cloned this and think the packaging is fine so far.  However, I do not
see any upstream and pristine-tar branch.  Could you please post what a
`git branch` at your site says?
 
> I fixed everything that lintian reported except these:
> 
> W: ray-extra: unusual-interpreter 
> ./usr/share/ray/scripts/plot-color-distributions.R #!/usr/bin/Rscript
> W: ray-extra: unusual-interpreter 
> ./usr/share/ray/scripts/plot-coverage-distribution.R #!/usr/bin/Rscript
> W: ray-extra: unusual-interpreter 
> ./usr/share/ray/scripts/plot-library-distribution.R #!/usr/bin/Rscript
> 
> r-base-core installs /usr/bin/Rscript so it is legit.

Yes.  I would consider this a lintian bug.
 
> Maybe /usr/bin/Rscript should be added in
> 
> /usr/share/lintian/checks/scripts (package lintian)
> 
> 
> Using /usr/bin/env Rscript also throws unusual-interpreter.

I do not see any reason to try tricks to circumvent lintian problems.
If I were you I would use `reportbug lintian` to report this problem.

If you are bored by these lintian warnings you could meanwhile add
the lines above in a file

debian/ray.lintian-overrides

and lintian will ignore the warnings.

> >In the Debian Med team we are using the team mailing list as maintainer
> >so everybody gets informed about uploads, bugs etc.  It is easy to add
> >other uploaders (people feeling responsible and adding code to the
> >packaging).
> >
> 
> Done.

Great.

> >and install these using a file ray.manpages containing just this string
> >(see dh_installman).  A patch would be fine as well but editing a plain
> >file is way more comfortable than handling a patch and it is easier to
> >point upstream to a plain file when asking upstream to include this file.
> >
> 
> I added debian/Ray.1 that is simplier than a patch.

Fine.
 
> >Keep on your good work and feel free to keep on asking if something else
> >might remain unclear
> >
> 
> My question: what's next ?

I think the package deserves a watch file to enable uscan detecting new
versions.  Unfortunately it is not that obvious to me how to download a
versioned source tarball - otherwise I would have just injected such a
watch file.
 
> I think debian/2.1.0-4 should do it, unless you have other suggestions, in 
> which case
> I will gladly implement them.

I just have one remark for the versioning:  While it is a good idea to
have differently versioned packages if there are changes it is kind of
usual to uploading a *new* package starting with version 2.1.0-1.  The
rationale is that changes done before the first upload to the official
Debian mirror are not really relevant.  While this is rather some
"usual" thing to do there is no real need I would like you to consider
this.  However, if there was some non-Debian upload (I have noticed that
2.1.0-2 was targeting at precise) you should definitely leave the
versioning as it is now.

What needs to be done before the upload is to file an so called ITP
(=Intend To Package) bug using

 reportbug wnpp

and following the menu.  It makes sense to add a final remark that the
package will be maintained in the Debian Med team and also give the
Vcs-Git location.

Kind regards and thanks for your work on this

  Andreas.

-- 
http://fam-tille.de


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121103172609.ga18...@an3as.eu



Re: Packaging Ray for Debian Med

2012-11-03 Thread Sébastien Boisvert

Hi,

On 11/03/2012 01:26 PM, Andreas Tille wrote:

Hi Sébastien,

many thanks for you preparations.
 
On Fri, Nov 02, 2012 at 05:09:10PM -0400, Sébastien Boisvert wrote:

The Ray Debian package is now at
http://anonscm.debian.org/gitweb/?p=debian-med/ray.git;a=summary


I cloned this and think the packaging is fine so far.  However, I do not
see any upstream and pristine-tar branch.  Could you please post what a
`git branch` at your site says?



You are right, branches upstream and pristine-tar were missing.

I added them:

$ git branch
* master
  pristine-tar
  upstream

$ git tag
pristine-tar/2.1.0
upstream/2.1.0

The branch upstream contains commits from git://github.com/sebhtml/ray.git

The tag upstream/2.1.0 is named v2.1.0 on github.

The pristine-tar branch contains the content of Ray-v2.1.0.tar.bz2 (tag 
pristine-tar/2.1.0)

As I understand, the package will be built using tag pristine-tar/2.1.0 to
generate the orig.tar.gz file and using tag debian/2.1.0-1 to generate the 
debian.tar.gz
file.

Is that right ?


I fixed everything that lintian reported except these:

W: ray-extra: unusual-interpreter 
./usr/share/ray/scripts/plot-color-distributions.R #!/usr/bin/Rscript
W: ray-extra: unusual-interpreter 
./usr/share/ray/scripts/plot-coverage-distribution.R #!/usr/bin/Rscript
W: ray-extra: unusual-interpreter 
./usr/share/ray/scripts/plot-library-distribution.R #!/usr/bin/Rscript

r-base-core installs /usr/bin/Rscript so it is legit.


Yes.  I would consider this a lintian bug.


Maybe /usr/bin/Rscript should be added in

 /usr/share/lintian/checks/scripts (package lintian)


Using /usr/bin/env Rscript also throws unusual-interpreter.


I do not see any reason to try tricks to circumvent lintian problems.
If I were you I would use `reportbug lintian` to report this problem.



I reported the bug and provided a patch for lintian:

"lintian says that /usr/bin/Rscript (r-base-core) is unusual"
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=692232


If you are bored by these lintian warnings you could meanwhile add
the lines above in a file

 debian/ray.lintian-overrides

and lintian will ignore the warnings.



I added this.


In the Debian Med team we are using the team mailing list as maintainer
so everybody gets informed about uploads, bugs etc.  It is easy to add
other uploaders (people feeling responsible and adding code to the
packaging).



Done.


Great.


and install these using a file ray.manpages containing just this string
(see dh_installman).  A patch would be fine as well but editing a plain
file is way more comfortable than handling a patch and it is easier to
point upstream to a plain file when asking upstream to include this file.



I added debian/Ray.1 that is simplier than a patch.


Fine.


Keep on your good work and feel free to keep on asking if something else
might remain unclear



My question: what's next ?


I think the package deserves a watch file to enable uscan detecting new
versions.  Unfortunately it is not that obvious to me how to download a
versioned source tarball - otherwise I would have just injected such a
watch file.



I added the watch file.

The tarballs are available on

http://sourceforge.net/projects/denovoassembler/files/


I think debian/2.1.0-4 should do it, unless you have other suggestions, in 
which case
I will gladly implement them.


I just have one remark for the versioning:  While it is a good idea to
have differently versioned packages if there are changes it is kind of
usual to uploading a *new* package starting with version 2.1.0-1.  The
rationale is that changes done before the first upload to the official
Debian mirror are not really relevant.  While this is rather some
"usual" thing to do there is no real need I would like you to consider
this.


I fixed this. Now there is only 2.1.0-1 and I removed the
git tags debian/2.1.0-3 and debian/2.1.0-4.

There are no debian/* tag at the moment.

I guess you (or I) can add the tag debian/2.1.0-1 on the HEAD of branch master
if you feel that everything is fine.


However, if there was some non-Debian upload (I have noticed that
2.1.0-2 was targeting at precise) you should definitely leave the
versioning as it is now.



The precise target was added by Tim Booth. I guess the target for 2.1.0-1
can be edited for Ubuntu releases using the same files. I prefer to start
at 2.1.0-1 in Debian. It makes more sense.


What needs to be done before the upload is to file an so called ITP
(=Intend To Package) bug using

  reportbug wnpp



"ITP: ray -- Parallel genome assemblies for parallel DNA sequencing"
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=692238


and following the menu.  It makes sense to add a final remark that the
package will be maintained in the Debian Med team and also give the
Vcs-Git location.



Fixed the Vcs-Git in debian/control.


Kind regards and thanks for your work on this



Thank you (and Tim as well) for the mentoring. I really appreciate the time you
have put on

Re: Packaging Ray for Debian Med

2012-11-03 Thread Charles Plessy
Le Sat, Nov 03, 2012 at 06:39:30PM -0400, Sébastien Boisvert a écrit :
> 
> The pristine-tar branch contains the content of Ray-v2.1.0.tar.bz2 (tag
> pristine-tar/2.1.0)
> 

Bonjour Sébastien,

this is not the purpose of the pristine-tar branch, as otherwise
pristine-tar/2.1.0 would be entirely rendundant with upstream/2.1.0.

Pristine-tar solves a techinical problem that is not obvious at first. If on
two computers one regenerates a tarball using upstream/2.1.0, results man not
be byte-identical.  Difficulties go beyond simple timestamp gymnastics.

Pristine-tar was therefore invented to store a binary delta, with which a
regenerated tarball is made byte-identical with a reference tarball, in a
space-efficient manner (storing all the tarballs would take unreasonable
space).

This is important because when a Debian package is updated with no changes
to the upstream sources (that is, when only the debian revision number changes),
the previously uploaded upstream source is used.  The "orig.tar.gz" of the 
updated
source package will not be uploaded, but as you can see in the Debian source
control file (".dsc"), its check sums are recorded, for comparison with the
one in the Debian archive.

You can see examples of pristine-tar branches in the Git repositories for bwa
or samtools for example. 

Pristine-tar branches are not strictly mandatory, but are especially useful
when working collectively on a package.  In the absence of pristine-tar
branches, the best way retreive the proper "orig.tar.gz" is to downolad the
previous source package from the Debian archive. 

Bon dimanche !

-- 
Charles Plessy
Debian Med packaging team,
http://www.debian.org/devel/debian-med
Tsurumi, Kanagawa, Japan


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121104022358.ga25...@falafel.plessy.net



Re: Bowtie2 ready for upload?

2012-11-03 Thread Charles Plessy
Le Fri, Nov 02, 2012 at 11:25:54PM -0400, Carlos Borroto a écrit :
> 
> I uploaded latest upstream version for Bowtie2, which dropped the beta
> tag. The package seems ready, but I haven't worked in this package
> before. Alex, could you please take a look and see if things look
> good.

Hi Carlos and Alex,

Bowtie 2.0.2 looks good to me.  I pushed two minor commits to remove obsoleted
Lintian overrides and use pristine-tar by default.

I recommend to upload directly to Unstable.  In case a RC bug were found in
2.0.0-beta6-3, it would not make much sense spending time correcting this beta
version anyway.

Lastly, have you considered adding regression tests using the example data ?
You can see http://dep.debian.net/deps/dep8/ for one framework available in
Debian and Ubuntu.

Have a nice Sunday,

-- 
Charles Plessy
Debian Med packaging team,
http://www.debian.org/devel/debian-med
Tsurumi, Kanagawa, Japan


-- 
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121104043829.gb25...@falafel.plessy.net



Re: Packaging Ray for Debian Med

2012-11-03 Thread Sébastien Boisvert

OK, so the things in my pristine-tar branch are not good at all !

I checked the pristine-tar branch of debian-med/bwa.git, and it contains
delta between orig tarballs.
But is the first orig tarball stored somewhere ?

The ray tarball from sourceforge (Ray-v2.1.0.tar.bz2) is constructed
from two git repositories with a script called scripts/ShipProduct.sh
in the ray upstream repository.


The script does this


release=$1
base=~/git-clones

RayGit=$base/ray
RayPlatformGit=$base/RayPlatform
productName=Ray-$release
productDirectory=~/ProductReleases/$productName

mkdir -p $productDirectory
cd $productDirectory

git clone $RayGit $productName &> /dev/null
cd $productName
git log|grep ^commit|head -n1|awk '{print $2}' > tag.txt
echo https://github.com/sebhtml/ray > github.txt
echo git://github.com/sebhtml/ray.git > git.txt
rm -rf .git
rm -rf RayPlatform

git clone $RayPlatformGit &> /dev/null
cd RayPlatform
git log|grep ^commit|head -n1|awk '{print $2}' > tag.txt
echo https://github.com/sebhtml/RayPlatform > github.txt
echo git://github.com/sebhtml/RayPlatform.git > git.txt
rm -rf .git
cd ..


tar cjf $productName.tar.bz2 $productName &> /dev/null


RayPlatform is a separate project, but I am not sure that it
deserves its own package at this point as it is only used by
Ray as far as I know.


On 11/03/2012 10:23 PM, Charles Plessy wrote:

Le Sat, Nov 03, 2012 at 06:39:30PM -0400, Sébastien Boisvert a écrit :


The pristine-tar branch contains the content of Ray-v2.1.0.tar.bz2 (tag
pristine-tar/2.1.0)



Bonjour Sébastien,

this is not the purpose of the pristine-tar branch, as otherwise
pristine-tar/2.1.0 would be entirely rendundant with upstream/2.1.0.

Pristine-tar solves a techinical problem that is not obvious at first. If on
two computers one regenerates a tarball using upstream/2.1.0, results man not
be byte-identical.  Difficulties go beyond simple timestamp gymnastics.

Pristine-tar was therefore invented to store a binary delta, with which a
regenerated tarball is made byte-identical with a reference tarball, in a
space-efficient manner (storing all the tarballs would take unreasonable
space).

This is important because when a Debian package is updated with no changes
to the upstream sources (that is, when only the debian revision number changes),
the previously uploaded upstream source is used.  The "orig.tar.gz" of the 
updated
source package will not be uploaded, but as you can see in the Debian source
control file (".dsc"), its check sums are recorded, for comparison with the
one in the Debian archive.

You can see examples of pristine-tar branches in the Git repositories for bwa
or samtools for example.

Pristine-tar branches are not strictly mandatory, but are especially useful
when working collectively on a package.  In the absence of pristine-tar
branches, the best way retreive the proper "orig.tar.gz" is to downolad the
previous source package from the Debian archive.

Bon dimanche !




--
Sent from my IBM Blue Gene/Q


--
To UNSUBSCRIBE, email to debian-med-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/5095f765.7020...@ulaval.ca