Daniel E. Macks <dma...@netspace.org> said: > Executive summary of heuristic: if the package uses Module::Builder > (for example, CompileScript invoking Build.PL) rather than MakeMaker > (the default script for type:perl, invoking Makefile.PL). > > Thanks to recent buildworlding, we've found a user-visible packaging > problem: if a perlmodule package is the first of its perl variant (any > certain XXX of pmXXX, or agnostic/nonvarianted pm) asserted > UpdatePOD:true in its .info but does not have a perllocal.pod > fragment, PostInst crashes. The UpdatePOD tag is specifically to > handle the perllocal.pod fragments with lots of internal (but > apparently fragile) magic, so the solution is just to avoid using that > tag if there is no such fragment.
Thinknig more, an alternative solution is to have the perl core packages themselves supply starter/dummy files so that even if the perlmodule package has no perllocal.pod fragment itself, the postinst won't crash. That's cleaner. Sort-of...protecting maintainers from making mistakes by hiding the mistakes is friendly I guess but icky. OTOH, if Module::Build ever does change its behavior, we would have to re-add IUpdatePOD:true (since it's a more severe packaging error to *not* declare it when there *is* a perllocal.pod). Advice please... dan -- Daniel Macks dma...@netspace.org ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Fink-devel mailing list Fink-devel@lists.sourceforge.net http://news.gmane.org/gmane.os.apple.fink.devel Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-devel
