[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: fiji packages



On Fri, 1 May 2009, Mark Longair wrote:

Actually, I hadn't tried building with openjdk for quite a
while, and now there are only a couple of errors left, which is
great.  They should be fixed soon.

This would be great.

...
ImageJ.  The "fiji" branch is based on the "imageja" branch and
has some additional patches.

 [1] http://repo.or.cz/w/imageja.git

Thanks for the clarification.  I have not yet made up my mind
what might be the best solution to integrate ImageJA.

Is there somewhere I can download the buggy .deb?  I've tried
building from the debian-med svn repository but the build rules
don't seem to unpack the upstream source correctly.  I'm doing:

  $ svn co svn://svn.debian.org/debian-med/
  [... lots of output ...]

  $ cd debian-med/trunk/packages/imagej/trunk/

  $ echo "origDir=.." >> .svn/deb-layout && uscan --force-download
  imagej: Version (1.42n) available on remote site:
    http://rsb.info.nih.gov/ij/download/src/ij142n-src.zip
    (local version is 1.42n)
  imagej: Successfully downloaded updated package ij142n-src.zip

  $ debian/rules get-orig-source
  . debian/get-orig-source

This should create the orig.tar.gz tarball I was using.  I then
copied the debian directory from SVN to the unpacked tarball and
used (p)debuild.

  [... lots of output, included in full below, but ending: ]

  cd /mnt/holyrood-backup/med-svn/debian-med/trunk/packages/imagej/build-area/imagej-1.42n && /usr/lib/jvm/java-6-openjdk/bin/java -classpath /usr/share/ant/lib/ant.jar:/usr/share/ant/lib/ant-launcher.jar:/usr/lib/jvm/java-6-openjdk/lib/tools.jar  -Dant.home=/usr/share/ant org.apache.tools.ant.Main -Dcompile.debug=true -Dcompile.optimize=true     build
  Buildfile: build.xml does not exist!
  Build failed
  make: *** [debian/stamp-ant-build] Error 1
  dpkg-buildpackage: failure: debian/rules build gave error exit status 2
  Command dpkg-buildpackage -us -uc -rfakeroot failed in <unknown>, how to continue now? [Qri?]:

Is there step I'm missing there?

That's strange - but whatever happened - I just uploaded my sources to

   http://people.debian.org/~tille/packages/imagej/

Incidentally, the "svn co" fails half way through with:

 Fetching external item into 'trunk/packages/med-doc/trunk/05_biomed_bio'
 svn: URL 'svn://svn.debian.org/svn/cdd/projects/med/trunk/papers/05_biomed_bio' doesn't exist

Thanks for the hint - this is fixed now.

I'll discuss this with the other developers.  There are several
plugins in Fiji that rely on features that are only in ImageJA
and I'm not sure what the history of those is with regard to
upstream ImageJ.

Well, we should finally find out what might serve our users best.
If you consider ImageJA stable enough to be used by *every* Debian
user it might probably make sense to replace ImageJ in Debian by
ImageJA - but my feeling says that it might be better to ask
ImageJ authors to accept those patches that enable the Plugins in
question.

That's broadly what I think, but there's quite a lot of work
involved.  (Ideally I would split the plugins into about 40
different packages, having done a quick count of those with
distinct Free Software licenses and authors.)  I do have plans,
however, for automatically generating more modular Debian
packages at this kind of granularity, and we're likely to work
on this over the summer.

This would be really great.

Certainly in theory both 1 and 2 would be the best way forward,
but pragmatically it might be better to just fix the OpenJDK
bugs, replace the obviously duplicated components with Debian
dependencies and get the monolithic package into contrib as soon
as possible.

I doubt that ftpmaster will accept duplication of large chunks
of code.  So even if you manage to solve the OpenJDK issues
this road does not sound very promissing to my knowledge.

However, I'm going to be working towards the same
essential goal whenever I have a chance.  Unfortunately I'm just
very short of time right at the moment.

That's the usual thing.  So even if it will take you more time
to do the modularisation this road will probably last longer but
will finally be successful (IMHO).

Kind regards

     Andreas.

--
http://fam-tille.de


Reply to: