Mail Archives: djgpp-workers/2001/12/26/10:07:28
> I think _everything_ should be updated. But that doesn't mean that
> you, Charles, need to be the one who does it all as part of the
> refresh.
Everything is a "big" order - since that also means that every package
on every computer and every CD everwhere needs to be swapped out. I
don't disagree - but I'd like to know where to focus - which
distributions are absolutely REQUIRED to get a functional package.
So when someone on a low bandwidth connection says they are on WinXP
with this CD full of 600Mb of DJGPP zips - we can say do these first.
Since that list is already a minimum of of around 7Mb+ ...
> As far as the packages I maintain, I was waiting for the refresh to
> stabilize before rebuilding them.
There are also some of the very old releases on Simtel (less features,
work great!) which should also be rebuilt.
It's probably a good idea to discuss an update strategy with everyone.
For example, UPX binary .exes ? DSMs in all packages?
> I agree that Fileutils should be refreshed ASAP, but I'm guessing
> that people who use them routinely did that already ;-).
I was surprised the test build environment had built/tested so many
things without any failures.
> > I so far have not seen problems with sed, bsn, txi, grep, txt, shl,
> > dif, gwk or pat without refreshing - but some of them seem to have the
> > nesting signature.
>
> I'd guess you didn't try all the obscure options of those packages,
> either. So problems might still lurk in there.
I don't doubt problems lurk in some obscure options. For example, info
allows you to spawn a command when running. But to trigger the bug you
would need to launch info from a 3rd image (ide/editor), then launch
a djgpp image from spawn inside info. Possible, but obscure.
There are some programs which obviously spawn - but I don't even know
how to make them do that :-)
- Raw text -