On Sun, 1 Jul 2007 15:05:54 -0500
"Raphael Geissert" <[EMAIL PROTECTED]> wrote:

> > > I have an alias in my .bashrc which helps me with this stuff:
> > >
> > > alias checkdeb='linda -i *.deb; linda -i *.dsc; lintian -i -I
> > > *.deb;
> > > lintian -i -I *.dsc'

What's wrong with just passing the .changes? That checks the .dsc and
each .deb

I stopped using linda some time ago - it is just incredibly buggy. I
admit I didn't file a bug report for every false positive or other
linda problem. I just uninstalled linda.

It shouldn't be too hard for someone to identify all packages in the
archive that are lintian-clean (without overrides) and run them through
linda to get a report of where lintian and linda disagree. Not all of
those will be bugs (because linda works differently to lintian and can
catch problems that lintian cannot and vice versa) but it's not
something I would want to do.

> > Linda has way too much false-positives (the above is just
> > one example) which is really bad.
> > bugs.debian.org/linda vs bugs.debian.org/lintian does look
> > like this as well.
> 
> There are only 5 bugs in linda:
> http://bugs.debian.org/cgi-bin/pkgreport.cgi?which=src&data=linda&archive=no&pend-exc=pending-fixed&pend-exc=fixed&pend-exc=done&sev-inc=important&sev-inc=normal

Minor and wishlist bugs still count - particularly for things like
linda and lintian, which leads to 16 outstanding. (The lintian
maintainer correctly insists that lintian - and therefore linda - are
only indicators of problems and cannot catch all such problems, they
are not intended to be the authoritative implementation of Policy.
Therefore, bugs in linda or lintian that relate to false positives,
missing checks and wrong results are requested to be filed as wishlist
only.) On that score, lintian has 99 bugs - yet lintian is still the
better of the two. 

As indicators of problems, lintian is the one to use but you and I
know that just because a package is lintian clean does NOT mean that the
package is in a good enough condition to be uploaded. Equally, there
are a small number of instances where a lintian override is the correct
response to a lintian error but sponsorees should not use
overrides without talking to their sponsor first!

-- 


Neil Williams
=============
http://www.data-freedom.org/
http://www.nosoftwarepatents.com/
http://www.linux.codehelp.co.uk/

Attachment: pgpmcAMKvDLjW.pgp
Description: PGP signature

Reply via email to