[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#155879: Willing to NMU



> Hi there,
> 
> I am willing to NMU the package libcarp-assert-perl as I need it to
> get some other packages into Debian. I have downloaded the package
> from your website and reviewed the packaging. It does currently need a
> few changes, these are:
> 
> 1) You should remove the debian/README.Debian as it does not contain
> any useful information which is additional to what is already stated
> elsewhere in your package.
> 
> 2) You need to add the correct url for original download. It is
> actually something like:
> http://search.cpan.org/CPAN/authors/id/M/MS/MSCHWERN/.
> 
> 3) You need to put a note as to location of the Perl license into
> debian/copyright (/usr/share/doc/perl/copyright).
> 
> 4) debian/rules should be corrected so the build is done in
> binary-indep not binary-arch, (basically just swap the binary-indep
> and binary-arch targets around).

The first four were pointed out to me by another kind Debian developer and have been remedied on my local build for some time.

> 5) debian/rules needs to be fixed so it creates *.3perl man pages not
> *.3pm, this is done by using the MAN3EXT=3perl option when doing the
> make and make install phases. e.g.:
> 
>    $(MAKE) OPTIMIZE="-O2 -g -Wall" MAN3EXT=3perl
> 
>    $(MAKE) install PREFIX=$(CURDIR)/debian/libcarp-assert-perl/usr MAN3EXT=3perl

Interesting. I've made that change.

> 6) You should be aware that although removing all empty directories is
> a good idea there is a subtle bug in your find command in that xargs
> splits on white space so a directory path containing whitespace (which
> could happen with Linux) will be broken up into two incorrect
> pieces. You should look at using the -0 (--null) option with xargs to
> make it safer.

Here is one where I hit a snag.  I made the xargs and find changes per their documentation, but what kills me is this. I tried it with a directory upstream called "Test Directory"
/home/jaybonci/Test Directory/deb/libcarp-asser-perl....

Everything worked a-ok until I got to the file installation part. I enclosed the PREFIX="CURDIR/stuph" in quotes, and it did end up creating a deb with no files (files ended up in /home/jaybonci/Test/usr etc). Lintian proceeded to choke with an internal error, with other errors from the file utilities.

Basically there are more than a few bugs going down here that really need to be fleshed out and made into good bug reports. I shall do so tonight, but in the meantime, I'd like to forge onwards with this package. It seems these bugs are not specific to my build process.

> 7) You are including Makefile.old in your package which accounts for a
> huge portion of the created diff (806 lines out of a 956-line diff). I
> think you need to do a 'make distclean' in your clean target in
> debian/rules rather than just 'make clean'. And then possibly do a
> 'debuild clean' before you next build the package. The diffstat
> package is a great tool for examining the diff output of debuild.

I took a look at the makefile, and I've slimmed it's diff down greatly.

> Basically most of these points (2,3,4 & 5) are necessary to make the
> package comply with debian policy. You should read through the
> perl-policy document
> 
> http://www.debian.org/doc/packaging-manuals/perl-policy/

I've read that several times, but I think a lot of this will come with experience, and can be filed under "Best practicing packages" in my mind.  Thanks for all of the advice. It is sincerely appreciated.

> If you fix all these issues, let me know and I will happily upload the
> fixed package. I am looking to get my own packages in fairly soon so I
> will go ahead and fix the bugs and upload if I don't hear anything
> from you in the next couple of weeks.

I'll attach packages to a separate email (they are quite small). It's going to take time to rebuild and redeploy all of my packages to the website, but if you can sponsor/NMU the libcarp-assert-perl I would be very grateful. That is the last step in my Debian application process. and I'm very eager to get started.

Thank you
	--Jay Bonci

-- 
Jay Bonci    | <jay@bonci.com>
GPG: E0B8B2DE| 562B 35DC BE8D 7802 DB31  6423 64D8 790F E0B8 B2DE

Attachment: pgpfBXVBjPxFa.pgp
Description: PGP signature


Reply to: