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

Re: Help with java package (beast-mcmc) needed

Le 1/10/12 2:11 PM, Andreas Tille a écrit :
> On Tue, Jan 10, 2012 at 12:01:57PM +0100, Olivier Sallou wrote:
>>>> Usually, this kind of problem is due to the CLASSPATH not containing
>>>> jam.jar
>>> $ grep jam debian/rules
>>> export CLASSPATH := $(DEBJAR)/itext.jar:lib/beagle.jar:lib/mpj.jar:lib/org.boehn.kmlframework_20090320.jar:$(DEBJAR)/junit4.jar:$(DEBJAR)/figtree.jar:lib/colt.jar:lib/options.jar:lib/mtj.jar:$(DEBJAR)/jam.jar:$(DEBJAR)/jdom1.jar:$(DEBJAR)/jebl.jar:$(DEBJAR)/commons-math.jar
>>> As far as I understood you do not need to explicitely set CLASSPATH at
>>> runtime (and it would not explain why the other executables are
>>> perfectly finding the needed jars.
>> For classpath, at runtime, all depends on how jar is generated. If it
>> contains a MANIFEST file with the classpath defined, it will be able to
>> find the JARS (supposing that libraries path are the same). If it dies
>> not contains the classpath in the MANIFEST file, classpath must be set
>> explicitly to each jar file in the command line (usually via a wrapper
>> shell)
> Apropos MANIFEST:  I formerly fiddled around with packaging using jh_manifest
> and there is also some alternative packaging method at
>    http://people.debian.org/~tille/packages/beast-mcmc-help-wanted/manifest/
> featuring a debian/beast-mcmc.manifest file.  Believe it or not it shows
> the very same behaviour.  The most strange fact for me is that the two
> executables loganalyser and logcombiner are very similar but only one of
> them runs and the other fails.
> Do you think I should simply add a CLASSPATH variable to those scripts that
> are failing without understanding why the others are working?

classpath in beast-mcm.manifest looks fine but did you check the content
of generated JARS with their manifest content?
(you can exec jar -xf myfile.jar to "unzip" it and look in sub dir
If generated manifest is file in one, and not in the other, it might a
simple typo issue with generation.

> Kind regards
>          Andreas.

Olivier Sallou
IRISA / University of Rennes 1
Campus de Beaulieu, 35000 RENNES - FRANCE

gpg key id: 4096R/326D8438  (pgp.mit.edu)
Key fingerprint = 5FB4 6F83 D3B9 5204 6335  D26D 78DC 68DB 326D 8438

Reply to: