"Peter O'Gorman" <[EMAIL PROTECTED]> wrote: > Martin Costabel wrote: > > More precisely: In the file atk1.info in 10.3/unstable, which is version > > 1.4.1-3 and which I have installed, there is no such field. In > > 10.3/stable in version 1.4.1-2, there is such a field. The fact that > > this is taken into account looks like a bug. > > I have to agree, we do not need to take older versions into account. > PkgVersion.pm says ' @dependslist = $package->get_all_providers();' and > loops through everything looking for something to complain about. > > Dave, did you intend this? >
Well, I didn't even consider it. It's hard to imagine a situation, other than correcting a typo or a radical change in packaging, in which BuildDependsOnly should change from version to version. In particular, I consider it an error that BuildDependsOnly: true was ever removed from the atk1 package. It contains headers and unversioned dylibs, so it should be BuildDependsOnly:true. I'm planning to add a routine to Validation.pm which looks to see if the package provides headers, and if so, checks to see if BuildDependsOnly:true has been set. If not, it issues a warning. Our shlibs system will only work if all packages adhere to it. -- Dave P.S. I suspect that atk1 was changed due to the fact that many of the new gnome packages depend on it. They shouldn't depend on it. However, there is a real problem with that set of packages: a problem that could be addressed by my proposed "InheritedBuildDepends" extension to fink. ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
