Bug#1076560: pfstools: FTBFS in experimental: ImageMagick silently replaced by GraphicsMagick

[Yavor Doganov]

Source: pfstools
Version: 2.2.0-7
Severity: minor
Tags: sid trixie experimental ftbfs

While build-testing this package for the forthcoming GSL transition, I
encountered a build failure which might be worth fixing.  The precise
command used was:

$ sbuild --extra-repository='deb http://deb.debian.org/debian
  experimental main' --build-dep-resolver=aspcud
  --add-depends='libgsl-dev (>= 2.8)' --arch-any --no-arch-all --no-clean

From the build log:

Get:178 http://deb.debian.org/debian unstable/main amd64 libgraphicsmagick-q16-3t64 amd64 1.4+really1.3.43-1+b2 [1200 kB]
...
Get:219 http://deb.debian.org/debian unstable/main amd64 libgraphicsmagick1-dev amd64 1.4+really1.3.43-1+b2 [1451 kB]
Get:220 http://deb.debian.org/debian unstable/main amd64 libgraphicsmagick++-q16-12t64 amd64 1.4+really1.3.43-1+b2 [119 kB]
Get:221 http://deb.debian.org/debian unstable/main amd64 libgraphicsmagick++1-dev amd64 1.4+really1.3.43-1+b2 [308 kB]
Get:222 http://deb.debian.org/debian unstable/main amd64 libgraphics-magick-perl amd64 1.4+really1.3.43-1+b2 [64.1 kB]
Get:223 http://deb.debian.org/debian unstable/main amd64 graphicsmagick-libmagick-dev-compat all 1.4+really1.3.43-1 [18.8 kB]
...
-- Could NOT find ImageMagick (missing: ImageMagick_Magick++_LIBRARY ImageMagick_MagickCore_LIBRARY)
-- ImageMagick not found. The following commands will not be
compiled: pfsinimgmagick pfsoutimgmagick pfsouthdrhtml.
...
dh_install: warning: Cannot find (any matches for) "usr/share/pfstools/hdrhtml_*" (tried in ., debilan/tmp)

dh_install: warning: pfstools missing files: usr/share/pfstools/hdrhtml_*
dh_install: error: missing files, aborting
make[1]: *** [debian/rules:55: override_dh_install] Error 255


I had to replace libmagick++-dev with libmagick++-6.q16-dev in
debian/control to obtain a successful build.  Since the former is a
dummy package, I would suggest to do that change.

This is unlikely to cause any trouble with sid uploads (most probably
the different behavior is due to the aspcud resolver) but it would be
nice to fix so that pfstools can be build-tested against libraries in
experimental.