Subject: Re: (a little off topic) sup
To: None <cwood@ichips.intel.com>
From: Ken Nakata <kenn@synap.ne.jp>
List: port-mac68k
Date: 05/04/1998 10:03:15
On Sun, 3 May 1998 16:36:43 -0700 (PDT),
Colin Wood <cwood@ichips.intel.com> wrote:
> SamMaEl wrote:
> > 	Also, a sup -v showed me this... I thought it kind of odd.
> > 
> > SUP Requesting changes since Dec 31 16:00:00 1969
> > 
> > 	Is there a flag or option I missed? Kind of defeats the purpose if
> > it has to download EVERY file (and I'm sure that every file has changed
> > since Dec 31, 1969 ;-))
> 
> This is just what it does the first time.  After it has completed the
> first run, it will only pull down the most recent files.

Provided that you know what you're doing, there's a way you can defeat
this.

Sup writes the time of last supscan on the server into files
${prefix}/sup/${collection}/when.${release} (*) after completion of a
successful run.  The time of last supscan is expressed as the number
of seconds from the Epoch (see date(1) -r).

*: prefix is where you have your source tree (e.g. /usr/src),
collection is usually current, and release is for instance,
ksrc-common or ksrc-mac68k.

So, you can fake these files by putting the time of last supscan
before your tar balls were created (current-source list subscribers
get the daily CVS update outputs).  If you don't have that
information, it's probably safe to use the time 24 hours earlier than
the tar balls creation time.

Having said that, however, it's probably safer just to run sup, and
let it update every file in the tree.  By default, sup does *not*
receive source files that are already up to date.  You can see the
difference when you run sup with the option -v.  The second sup and
after will be much shorter than the first run.

Later,

Ken