>  * the second one is that librmagick-ruby is very easily broken by
> newer versions of imagemagick, which is actually very hard to track.
> This will need some thinking; possibly a solution would be a strict
> versioned dependency along if only binNMUs for each imagemagick
> release ?

What happened to this idea? I think this is the (a?) reason ruby-rmagick
is currently failing its autopkgtests, using an imagemagick at runtime
that it is not built against?

We have seen this on Ubuntu as well, and I wonder if it would be
appropriate to just extract the libmagickcore version used at build-time
and override the shlib-depends for that package with specifically the
built-using version? Thanks to Adam Conrad for this suggestion.

I can work on a patch, if it would be acceptable.

-Nish

-- 
Nishanth Aravamudan
Ubuntu Server
Canonical Ltd

Reply via email to