Re: I want to colaborate with debian-med

[Andreas Tille <tillea@rki.de>]

On Sun, 22 May 2005, Adrian Mastronardi wrote:

> On 5/22/05, Andreas Tille <tillea@rki.de> wrote:
...
Correction for my first mail:
  In item 3. I mentioned you should flag the bug RFP.  For sure I meant
  ITP (Intent To Package).

> > $ ldd -v debian/mummer/usr/bin/annotate
> >          not a dynamic executable
The situation is really strange.

$ cd mummer-3.17/src/tigr
$ make gaps
g++  tigrinc.cc -c -o tigrinc.o
g++  gaps.cc tigrinc.o -o /home/andreas/debian-maintain/sponsor/adrian/mummer/mummer-3.17/src/tigr/gaps; chmod 755 /home/andreas/debian-maintain/sponsor/adrian/mummer/mummer-3.17/src/tigr/gaps
$ make annotate
g++  annotate.cc tigrinc.o -o /home/andreas/debian-maintain/sponsor/adrian/mummer/mummer-3.17/src/tigr/annotate; chmod 755 /home/andreas/debian-maintain/sponsor/adrian/mummer/mummer-3.17/src/tigr/annotate
$ ldd annotate gaps
annotate:
        not a dynamic executable
gaps:
                libstdc++.so.5 => /usr/lib/libstdc++.so.5 (0xb7f19000)
        libm.so.6 => /lib/tls/libm.so.6 (0xb7ef6000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x41312000)
        libc.so.6 => /lib/tls/libc.so.6 (0xb7dc1000)
        /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0xb7fea000)

I'm no C++ expert but I really fail to see the difference which causes
ldd to fail. Perhaps we should discuss this on debian-devel list.  I'm
not sure whether such things might be caused by different libraries and
you need to add a certain build dependency.

> Running lintian I found below 'warnings' as well
>
> W: mummer: interpreter-not-absolute ./usr/bin/exact-tandems #!-f
> N:
> N:   This script uses a relative path to locate its interpreter. This path
> N:   will be taken relative to the caller's current directory, not the
> N:   script's, so it is not likely to be what was intended.
> N:
> W: mummer: unusual-interpreter ./usr/bin/exact-tandems #!-f
> N:
> N:   This package contains a script for an interpreter that the Lintian
> N:   maintainers have not thought of. It could be a typo for a common
> N:   interpreter. If not, please file a wishlist bug on lintian so that the
> N:   Lintian maintainers can add this interpreter to their list.
> N:
> W: mummer: interpreter-not-absolute ./usr/bin/run-mummer1 #!-f
> W: mummer: unusual-interpreter ./usr/bin/run-mummer1 #!-f
> W: mummer: interpreter-not-absolute ./usr/bin/run-mummer3 #!-f
> W: mummer: unusual-interpreter ./usr/bin/run-mummer3 #!-f
This is strange because what is moved tp my debian/mummer/usr/bin
directory just contains
   #!/bin/csh -f
which is definitely an absolute path.  But it has to be changed anyway
because it would add a really unneded csh dependency.  The quite
simple scripts should be changed to POSIX shell syntax.

> I'll check debian/install and will let you know when you can download
> the new deb file.
Further hint: dh_install is mostly used without (=instead) of
      make install
If you think it makes sense to use
      $(MAKE) install DESTDIR=$(CURDIR)/debian/mummer
(better would be
      $(MAKE) install DESTDIR=$(CURDIR)/debian/tmp
) then dh_move will be your friend.

> Many thanks for your help,
I regard this as an important part of my Debian-Med work - so no
thanks is necessary.

Kind regards

          Andreas.

--
http://fam-tille.de