Bug#753657: mk-build-deps: -i broken

[James McCoy]

On Fri, Jul 04, 2014 at 02:30:36AM +0200, Jan Braun wrote:
> the --install option of mk-build-deps is not working for me:

It is and it isn't. :)

> | $ mk-build-deps -s sudo -i polipo
> | dpkg-gencontrol: warning: File::FcntlLock not available; using flock which is not NFS-safe
> | dpkg-deb: building package `polipo-build-deps' in `../polipo-build-deps_1.1.1-1_all.deb'.
> | 
> | The package has been created.
> | Attention, the package has been created in the current directory,
> | not in ".." as indicated by the message above!
> | dpkg-gencontrol: warning: File::FcntlLock not available; using flock which is not NFS-safe
> | dpkg-deb: building package `polipo-build-deps' in `../polipo-build-deps_1.1.1-1_all.deb'.
> | 
> | The package has been created.
> | Attention, the package has been created in the current directory,
> | not in ".." as indicated by the message above!

For some reason, “apt-cache showsrc polipo” listed the same stanza
twice, so we built two polipo-build-deps_1.1.1-1_all.deb packages.  Can
you show what “apt-cache showsrc polipo” and “apt-cache policy polipo”
output?

> | Use of uninitialized value $deb_files[1] in system at /usr/bin/mk-build-deps line 354.
> | (Reading database ... 209746 files and directories currently installed.)
> | Preparing to unpack polipo-build-deps_1.1.1-1_all.deb ...
> | Unpacking polipo-build-deps (1.1.1-1) over (1.1.1-1) ...

This installed the deb that was created.

> | dpkg: error processing archive  (--unpack):
> |  cannot access archive: No such file or directory

But then failed on the empty argument.

> The immediate reason is an extra empty argument to dpkg, as seen by:
> | $ mk-build-deps -s doom -i polipo
> | [...]
> | Use of uninitialized value $deb_files[1] in system at /usr/bin/mk-build-deps line 354.
> | 4 args: <dpkg> <--unpack> <polipo-build-deps_1.1.1-1_all.deb> <>

The way we generate the arguments is by globbing for debs matching the
package that was just created (polipo-build-deps_*.deb).  In this case,
since we gave it the same argument, it appears to have triggered the
"return undef when (matching) list is exhausted".

That can easily be handled by checking for undef, but the bigger
question is why we thought we had to create two of the same package.

Cheers,
-- 
James
GPG Key: 4096R/331BA3DB 2011-12-05 James McCoy <jamessan at debian.org>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 949 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/devscripts-devel/attachments/20140704/7f468b9f/attachment.sig>