Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 20 Feb 1996 23:09:38 -0800
From:      asami@cs.berkeley.edu (Satoshi Asami)
To:        mark@grondar.za
Cc:        ports@freebsd.org
Subject:   Re: Improvement? for the files/ directory...
Message-ID:  <199602210709.XAA12839@sunrise.cs.berkeley.edu>
In-Reply-To: <199602210700.JAA18191@grumble.grondar.za> (message from Mark Murray on Wed, 21 Feb 1996 09:00:54 %2B0200)

next in thread | previous in thread | raw e-mail | index | archive | help
 * Duh. 'Cos I didnt actually look at what md5 looked like. Although
 * If I wanted to write a script that nuked all old distfiles, I
 * couldn't use files/md5 right now, as that would kill those ports
 * which do not have md5 files (TeX, LaTeX spring to mind).  Any way
 * we could do something similar to my idea that would be compulsory
 * for all ports that could be used to give a definitive list of
 * relevant distfile names? Actually - I could probably come up with
 * this given ten minutes or so. Interested?

You mean for the ports that can't have md5 for some reason?

 * I know the job is a bit of a bitch as some authors do not change the
 * tarball name when they change the contents of the file. (TeX, LaTeX
 * and expect are culprits). Still, the names are still useable?

That's not the problem.  Missing md5's are usually because the
tarballs are built on the fly (and the dates and such can screw up the 
checksum).

I don't like the idea of adding yet another file, as most of the
information is already there in files/md5.  All we need to do is to
add a special treatment in files/md5 for "ignorable" checksums or
something.

Satoshi



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199602210709.XAA12839>