Re: [gNewSense-users] Re: speaking of scripts...

[Brian Brazil]

On 10/18/07, Markus Laire <address@hidden> wrote:
> On 10/18/07, Chris Andrew <address@hidden> wrote:
> >
> > This is a great idea, that could potentially save many hours of work.
> >
> > What does anyone think?
> >
> > > Realistically, 99% of the time the copyright won't change from version
> > > to version but we still need to check every package each time we
> > > distribute a new gNS. Having a program deal with this would save time,
> > > no? We should take advantage of the fact that we now have (or almost
> > > have) a database from which to do comparison from. Otherwise, we have
> > > to do all of the work again from scratch.
> > >
> > > Can the script by Tryggvib and Lee be cannibalized to do something like 
> > > this?
> > >
> > > Am I making sense? Or have I missed something in this process?
>
> If copyright-files don't usually change *at all* (not even a single
> character) between releases, then trivial solution would be to
> calculate e.g. SHA1-checksum for all copyright-files and check whether
> the checksum has changed.

Considering the small amount of data involved in a given file, doing a
full comparison will probably be faster and more accurate in the
two-revision case.

I had already been thinking along these lines for Hardy, as it clearly
is a big time saving and will greatly reduce the amount of time we
spend on packages. I'm hoping to use the time we have freed up to
check every file in the Hardy kernel before it is released.

Brian