On Mon, 4 Aug 2014 09:20:34 -0500, Hanspeter Niederstrasser
wrote:
On Sun, August 3, 2014 10:09 pm, Daniel Macks wrote:
> > Should we make a validator test for if:
> > Â Package:*-pm(\d+)
> > then require:
> > Â Depends:perl\1-core
> > ?
> >
> > Should it be a .info test or a .dep test?
> >
> > I tried a few grep-based heuristics on dists/10.7/finkinfo and found 40-80
> > such missing.
>
> Probably doesn't hurt to have a val test.
>
> It should be a .info test, I think. If a deb is built w/ perlXXX-core
> present, but used where absent, the runtime error is pretty obvious. But a
> build w/out perl5.x.y around will fail.
Added to git master (and also for python and ruby). I allowed a
dependency on another package with matched variant as an alternative to
a dependency on the language itself. That is,
bar-pm5124:Depends:foo-pm5124 is sufficient as it inherits an indirect
dep on perl5124-core via foo-pm5124 (which either has an explicit
Depends:perl5124-core or else recurses down to other *-pm5124).
It uncovered a handful of cases in 10.7/finkinfo, all fixed now. Please
let me know if your build/run testing finds any others (== this new
test is leaky) or other problems with it.
dan
--
Daniel Macks
dma...@netspace.org
--
Infragistics Professional
Build stunning WinForms apps today!
Reboot your WinForms applications with our WinForms controls.
Build a bridge from your legacy apps to the future.
http://pubads.g.doubleclick.net/gampad/clk?id=153845071&iu=/4140/ostg.clktrk
___
Fink-devel mailing list
Fink-devel@lists.sourceforge.net
List archive:
http://news.gmane.org/gmane.os.apple.fink.devel
Subscription management:
https://lists.sourceforge.net/lists/listinfo/fink-devel