On Fri, Nov 09, 2001 at 11:51:40AM -0800, Kevin A. Burton wrote: > > s/doesn't/does/ ? > > yup... sorry :( should have been does > <snip>
:) > > Yes that is what you have to do. Libbar-1.1.1 needs to be a package with > > that > > name so an upgrade does not introduce conflicts. I think the new proposed > > java > > policy adresses this quite good. > > Is there a link to the new proposed policy.. I think I saw it posted but > can't > remember where I put it :( > <snip> > > <snip> > > > > Are you saying that .WAR files are also incorrect? > > > > Yes they are most probably incorrect. > > ... perhaps on debian systems. It does make it easy to deploy java > applications especially if you want to be as 100% java as possible. > > I don't think we will get past this issue though. The WAR approach does allow > one VM to load code from multiple places due to the classloader approach. > > If we want to push this 100% we will have to use a more modern classloader > (IMO > the standard classloader suck). > > > > Just because it isn't the UNIX approach doesn't make it incorrect. > > > > > > I have an Open Mind (TM) on the subject so I am listening... Explain why > > > the Servlet WAR spec is incorrect and how it could be done better. > > > (Specifically WRT the WEB-INF/lib approach). > > > > Wars are not needed if you have deb packages. > > Not necessarilly. In the above scenario they are needed. No it is not needed. See below. > > Why have a packaging system (and a very bad one not allowing symlinks) > > inside > > an other packaging system? > > There are advantages. Again nothing is black and white. :) > > > You can use simple directories instead of the war files. :) > > I don't understand??? > <snip> You say that the war-files allow the classloading. That is not entirely true. Take tomcat for example. * You place a war-file in the specified directory. * Restart tomcat. * Tomcat now unzips this file to the webapp directory. * Tomcat is started and uses the directories that has the WEB-INF dirs and more. Conclusion: The war-file is not used directly. It is mearly unzipped and because we have a deb arproach in debian the war-files is not needed. War files have a major disadvantage and that is that they are simple zip- archives which means that they will not support symbolic links. I think we should cover the war-files in the policy too. They should be avoided, right? Regards, // Ola > - -- > Kevin A. Burton ( [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] ) > Location - San Francisco, CA, Cell - 415.595.9965 > Jabber - [EMAIL PROTECTED], Web - http://relativity.yi.org/ > > Resistance is *not* futile! > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.0.6 (GNU/Linux) > Comment: Get my public key at: http://relativity.yi.org/pgpkey.txt > > iD8DBQE77C+QAwM6xb2dfE0RAu2GAKDRLpPrG+kfJXbykxflueaQLyecgQCgiRLj > ZMGGVdzAeEaN1VIMt1Dacn4= > =/hOr > -----END PGP SIGNATURE----- > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- --------------------- Ola Lundqvist --------------------------- / [EMAIL PROTECTED] Björnkärrsgatan 5 A.11 \ | [EMAIL PROTECTED] 584 36 LINKÖPING | | +46 (0)13-17 69 83 +46 (0)70-332 1551 | | http://www.opal.dhs.org UIN/icq: 4912500 | \ gpg/f.p.: 7090 A92B 18FE 7994 0C36 4FE4 18A1 B1CF 0FE5 3DD9 / ---------------------------------------------------------------
