Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 23 Sep 1998 02:46:37 +0200
From:      Stefan Esser <se@mi.uni-koeln.de>
To:        Satoshi Asami <asami@FreeBSD.ORG>, ady@warpnet.ro
Cc:        ports@FreeBSD.ORG, Stefan Esser <se@mi.uni-koeln.de>
Subject:   Re: fetch + size
Message-ID:  <19980923024637.A7692@mi.uni-koeln.de>
In-Reply-To: <199809230014.RAA09344@silvia.hip.berkeley.edu>; from Satoshi Asami on Tue, Sep 22, 1998 at 05:14:58PM -0700
References:  <Pine.BSF.3.96.980922224533.12618K-100000@ady.warpnet.ro> <199809230014.RAA09344@silvia.hip.berkeley.edu>

next in thread | previous in thread | raw e-mail | index | archive | help
On 1998-09-22 17:14 -0700, Satoshi Asami <asami@FreeBSD.ORG> wrote:
> Sorry, I forgot to mention that it will be merged to -stable soon.

> Sorry, forgot to mention that too.  The point is that if there is a
> master site with a different file but with the same name, fetch can
> (usually) tell right away that the size is wrong.  On the other hand,
> you will have to fetch the whole file to see that the checksum is
> wrong (at which point the fetch loop will terminate with an error).
> 
> This will help in cases where there are multiple master sites and only 
> a few have the correct file.  Then the new "fetch" will return an
> error for size mismatch and bsd.port.mk will try to continue on a new
> site.
> 
> Stefan, did I get that right?

Perfectly right, of course ;-)

You may remember that I suggested that change a long time ago,
but did not have time to prepare an example implementation, at
that time. The reason that make me wish that feature (which you
explain above), still exist, and I found what I consider a 
minimally intrusive method to add the size check feature ...

If you don't feel confident about the change to bsd.port.mk,
which passes the "-S" option to fetch, than you may still want 
to add the one line change to "makesum", which adds valuable 
information about the amount of data to retrieve.

I agree, that "files/md5" doesn't seem like the best place for
file size information. But I'd rather change the name of that 
file, then introduce another one. And then, you may as well keep
the name unchanged. Nobody will ever notice, that there is more
in files/md5 than MD5 checksums, unless you tell him ;-)

When I prepared the changes, I was careful to keep the "fetch"
binary part (i.e. "-S" option), the "makesum" target (to add the
file size information) and the "fetch" target (to pass the size
via the -S option to the fetch binary) separate. You can have 
most any combination of them in a system, with the exception of
"fetch" target, md5 file created by a patched "makesum" target,
but fetch binary without support for "-S". That's the only case
that will fail, and that's why I wanted to make sure 3.0 will
have fetch -S.


Regards, STefan

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980923024637.A7692>