[Debian-med-packaging] Bug#882169: seqan2: FTBFS on indep-only build

[Andreas Beckmann]

Source: seqan2
Version: 2.3.2.000platform-issues6-9cf5a69+dfsg-3
Severity: serious
Justification: fails to build from source (but built successfully in the past)

Hi,

seqan2 FTBFS during the indep-only build of the arch:all packages:

https://buildd.debian.org/status/fetch.php?pkg=seqan2&arch=all&ver=2.3.2.000platform-issues6-9cf5a69%2Bdfsg-3&stamp=1511019385&raw=0

   debian/rules override_dh_installman
make[1]: Entering directory '/<<BUILDDIR>>/seqan2-2.3.2.000platform-issues6-9cf5a69+dfsg'
$DEB_BUILD_OPTIONS is [parallel=4]
dh_link
find -L /<<BUILDDIR>>/seqan2-2.3.2.000platform-issues6-9cf5a69+dfsg/debian/seqan-apps/usr/bin -name . -o -type d -prune -o -type l -exec rm {} +
find: '/<<BUILDDIR>>/seqan2-2.3.2.000platform-issues6-9cf5a69+dfsg/debian/seqan-apps/usr/bin': No such file or directory
debian/rules:107: recipe for target 'override_dh_installman' failed
make[1]: *** [override_dh_installman] Error 1


You probably want a override_dh_xxx-arch target there, but looking at the
rules file I don't get what you are doing there. Why is dh_link called
in override_dh_installman? And are the two find calls in in
override_dh_link-arch and override_dh_installman not doing the same?
And could be simplified to 

	# delete broken symlinks - some apps are not built on all architectures
	find -L $(CURDIR)/debian/$(pkgapps)/usr/bin -type l -delete

Then there is this comment:
	# we generate only those manpages where binaries are linked to /usr/bin
Which somehow implies: the list of manpages included some arch:all doc
package depends on the architecture where the package was built ???
Shouldn't that rather be a superset?


Andreas

PS: using the full hash would only marginally increase the length of the version number ;-)