On Wed, 02 Jul 2014 14:26:09 +0200, Jonas Smedegaard wrote:

> Hmm, I am getting doubts again...

Let's see ...
 
> Niko's declaration, with Gregor's correction, is this:
> 
>   Build-Depends: perl (>= 5.17.1~) | libmodule-build-perl (>= 0.400000), 
> libmodule-build-perl
> 
> That declaration relies on the assumption that any system with a new 
> enough perl will also have a new enough libmodule-build-perl - either 
> provided by perl(-modules) or standalone.  Is that safe to assume?

Good question.

I think it is safe now [0], as we have:
perl-module 5.20.0-1
Breaks: libmodule-build-perl (<< 0.420500)
and no Provides, so the single 'libmodule-build-perl' can only be the
real package, and must be >= 0.420500.

For a future perl-modules 5.22, if the Breaks is kept, everything
should be fine. If the Breaks is dropped there is indeed a loophole
(for versions >= 0.420500), if I'm seeing this correctly.
 
> I believe this more strict declaration covers that cornercase:
> 
>   Build-Depends: perl (>= 5.17.1~) | libmodule-build-perl (>= 0.400000), 
> libmodule-build-perl (>= 0.400000) | perl (<< 5.21~)

(I guess I'd still go for a plain 'libmodule-build-perl (>= 0.400000)'.
If only to save us all from further headaches :))


Cheers,
gregor


[0] besides that fact that the (depracated) M::B is still included

-- 
 .''`.  Homepage: http://info.comodo.priv.at/ - OpenPGP key 0xBB3A68018649AA06
 : :' : Debian GNU/Linux user, admin, and developer  -  http://www.debian.org/
 `. `'  Member of VIBE!AT & SPI, fellow of the Free Software Foundation Europe
   `-   NP: George Harrison: I Remember Jeep

Attachment: signature.asc
Description: Digital Signature

Reply via email to