New john packages for Debian sid?

[Javier Fernández-Sanguino Peña]

--adJ1OR3c6QgCpb/j
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Jul 05, 2004 at 07:52:10AM -0300, Guilherme de S. Pastore wrote:
>=20
> Haven't you received my e-mail regarding an upload to unstable? Well,
> check the list archives if you didn't :)

Oh, I didn't see that one sorry. Actually, there are no list archives in=20
the project I can see. Where can I retrieve them?

> It should be already there, but my sponsor didn't upload it because of
> some lintian warnings, especially the one about the configuration file
> check being in the postinst. I moved the check to config and added a
> real alternative to the suggestion on wordlist, and he's going to upload
> it in some hours.

Ok. Are you going to publish it somewhere so we can review them before the=
=20
upload?

> I'll commit the changes to the CVS as soon as I get time (/me add's it
> to his new shiny to do list, so he doesn't forget and can define a
> deadline :-))
>=20
> About the bugs that should be closed, we can discuss more later on, I
> don't really have time to read all of your opinions right now, I'm in a
> hurry, sorry :/ But one called my attention, the one about using generic
> to build all archs. I don't really want to have problems again of john
> not entering testing because of failing builds. I totally agree with

That's what happening now because it won't compile on sparc.

> using the generic target, and even worked on that during debconf, but
> before uploading, we should put it in a separate repository and, IMHO,
> ask for people who could test if it builds on different archs on -devel.
> Then I think we could safely (and perhaps gradually) add archs that
> work OK :)

There are two independent things here:

1.- Have john Arch: all. That change ws done in experimental, and the=20
people can be asked to build that one and see if it works.

2.- Add code in debian/rules to have john build using 'generic' save for=20
some architectures.

1) is not necessary, but, even if it was added, if new architectures failed=
=20
building it would not hinder john from transfering to testing. Actually,=20
only build failutres on architectures that _previously_ built properly stop=
=20
a package from progressing to testing.

2) is needed in order to have a succesful sparc build of john. Until this=
=20
is fixed, no new john versions will progress from sid to sarge at all.

Given that 1) will not hinder the package and that it can make us take=20
advantage of the build daemons to detect build failures (if any) I would=20
suggest doing it. In any case 2) is mandatory.

> And also, as I've said in a previous e-mail I'm not sure you received,
> I think it's not utterly necessary to run after these changes for the
> unstable version, since we might have progress enough to put the devel
> snapshot from experimental in unstable, hopefully =3DD
>=20
> What do you think? =3D)

I don't believe that the version from experimental is suited to be in=20
unstable. There are several things that need to happen before that version=
=20
is usable:

1.- There needs to be code to migrate the old configuration file locations=
=20
=66rom /etc to /etc/john

2.- We have to discuss wether we want the --private flag or, as an=20
alternative solution, we want the cronjob to run as a new 'john' user and=
=20
have it's homedir be /var/lib/john.

3.- The new version lacks documentation

In the meantime the unstable version suffers a number of issues we should=
=20
be quick fixing:

a) as pointed out by Solar, the wordlist approach is incomplete unless we=
=20
merge the wordlist selected by the user and the default password list=20
provided by John

b) the cronjob does not work at all as expected and is awfully broken. I=20
believe a better approach should be taken using start-stop-daemon as I've=
=20
implemented in the experimental branch. Check version 1.4.24 for example,=
=20
if you review the patch available at
http://cvs.alioth.debian.org/cgi-bin/cvsweb.cgi/john/debian/extra/cronjob.d=
iff?r1=3D1.4&r2=3D1.4.2.4&cvsroot=3Dpkg-john&only_with_tag=3Ddevel
you will see a number of changes some of which could be integrated into the=
=20
main branch. The way the cronjob 'kills' anything that resembles john is=20
quite badly designed.

As a side note: the start-stop-daemon calls could be used to make the=20
cronjob run as a non-root user which is better than having john run as root=
=20
IMHO.

Actually, it would be nice if some of these things could be added as=20
'Tasks' or 'Feature Requests' in the Alioth home page so we could easily=20
keep track of what needs to be done in a more organized manner...

Regards

Javier

--adJ1OR3c6QgCpb/j
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA6kwoi4sehJTrj0oRAj15AJwLuiHyxbTOm8KApQ5VAFAK2hra0QCfcIGD
mVztofD3jVUiNUSpcYzbOdA=
=H8ih
-----END PGP SIGNATURE-----

--adJ1OR3c6QgCpb/j--