Philip Lamb wrote:
My updated solution to this problem is to create 2 different shlib splitoffs, with different version numbers.. this appears to be what the openvrml project actually does for their .rpm packages.
Following what you see in the .spec file is a good idea.
So now I'm waiting for the openvrml package to be moved into -unstable by someone with the authority to do that. I hope someone can find time, as we have other projects here which are depending on openvrml being available in fink before we can release. Hopefully the results of this will be more packages for fink...
On a related note, is it just me or are there issues at the moment with getting submitted packages into the unstable tree in a timely fashion? I understand the contraints on the time of the core fink team, I'm just wondering if the bar over which packages have to leap to be accepted into -unstable is a bit too high. Provided a package has been validated and builds ok for the maintainer, shouldn't the job of installing and testing be done by the fink community at large through the -unstable tree, rather than depending on the scarce resources of the core fink team?
For many reasons. Firstly, if we allow a user without any previous experience to put packages into the tree without any testing at all, it leads to all kinds of security issues. Mr. Doe could add a package which does "rm -rf /" in it's postinstall script, the users who found the problem would report it, but how many users would be badly affected? Even those of us who have been doing this for a while have occasionally managed to remove a few users /etc symlink or worse.
A maintainer must go through a learning curve, like everything else in life it takes time, once a current cvs committer is sure that a maintainer is consistently submitting pacakges which require little review before being added to the tree that committer should notify fink-core (or any one of the core members) to add the maintainer as a cvs committer. There may be a better way of doing this, such as weeding out the packages which do not even pass the validator tests before human interaction is required, but we have not implemented such a thing as yet.
Peter -- Peter O'Gorman - http://www.pogma.com
------------------------------------------------------- This SF.Net email is sponsored by: InterSystems CACHE FREE OODBMS DOWNLOAD - A multidimensional database that combines robust object and relational technologies, making it a perfect match for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8 _______________________________________________ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
