On Wed, Jan 11, 2012 at 05:12:44PM +0100, Olivier Sallou wrote: > I also remark in beast.jar MANIFEST that there is a carriage return: > > Class-Path: /usr/share/beast-mcmc/beagle.jar /usr/share/java/jebl.jar > /u <---- there should not be a carriage return here > sr/share/java/jam.jar /usr/share/java/treeview.jar > > this may cause issue in classpath read.
I do not have the slightest idea how this CLASSPATH information is created (neither do I finally understand why those Java guys are entertaining different spellings for the same thing (CLASSPATH versus Class-Path). The only file I'm providing is debian/beast-mcmc.manifest which is probably / hopefully used by jh_manifest (triggered by the use of javahelper). I did some experiment by injecting a random string and it was moved unchanged to the resulting MANIFEST.MF file. However, as you can surely verify debian/beast-mcmc.manifest does not contain any line breaks - just one line for Class-Path. I do not have any idea how these line breaks will be inserted nor whether they do really harm. Could anybody from debian-java please comment on this. We are talking about the preliminary packages at http://people.debian.org/~tille/packages/beast-mcmc-help-wanted/ where I needed to inject explicite CLASSPATH in all wrapper scripts because the information in MANIFEST.MF seems to be perfectly ignored. I also do have another question to the Java experts: We have quite large data sets and faced situations where the setting -Xmx2048m is not enough. Is there any reasonable limit for this parameter? Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/[email protected]
