cvs commit: CVSROOT modules ports/shells Makefilepkg-install patch-bashline.c patch-builtins_shopt.def patch-config-bot.h ...

[Joe Marcus Clarke]

On Wed, 2004-08-11 at 08:04, Maxim Sobolev wrote:
> P.S. My recommendation is to kill both bash1 and bash2, repo-copy bash3 
> into just bash and remove bash3 as well.

Actually, we can't do that.  In order to preserve all histories, we
should delete bash2 and bash3, and update bash to 3.0....that is, if
that's what all the maintainers agree to.

Joe

> 
> Maxim Sobolev wrote:
> > Oliver Eikemeier wrote:
> > 
> >> Maxim Sobolev wrote:
> >>
> >>> Hmmm, why do we have those "bash", "bash2" and "bash3"? There may 
> >>> have been some historical reasons for bash/bash2 separation, but I 
> >>> wonder if they are still valid for the bash2 vs bash3 case.
> >>
> >>
> >>
> >> I guess bash 3.0 (like most .0 releases) has still some bugs to be 
> >> ironed out, see for example:
> >>   <http://lists.freebsd.org/pipermail/cvs-ports/2004-August/043003.html>
> >>   <http://lists.freebsd.org/pipermail/cvs-ports/2004-August/043006.html>
> > 
> > 
> > Well, all software have bugs, if bash maintainers think that it is ready 
> > for release, shouldn't we just agree with their decision? Critical bugs 
> > can be backported into the ports tree if necessary until next stable 
> > release is out. That is how our ports tree works. If somebody wants 
> > previous version he can get it from pre-compiled packages or from cvs repo.
> > 
> >> Therefore it seems wise to keep bash2 to run scripts until bash3 is 
> >> mature.
> >> OTOH people might want to use the new bash3 features:
> >>   <http://cnswww.cns.cwru.edu/~chet/bash/NEWS>
> >>
> >> So having bash2 and bash3 is justified. Do you think the directories 
> >> should have different names?
> > 
> > 
> > I still don't see the reason for having bash2/bash3. We have more than 
> > 10000 ports in the tree, most of them are routinely being updated to the 
> > new major release without creating those ugly new fooN ports.
> > 
> > Creation of fooN is only justified if it is backward incompatible with 
> > foo{N-1}, while there are still ports in the tree that rely on previous 
> > version. Hypotetical bugs in .0 release does not justify it.
> > 
> > -Maxim
> > 
> > 
> > 
-- 
PGP Key : http://www.marcuscom.com/pgp.asc


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: This is a digitally signed message part
Url : http://lists.freebsd.org/pipermail/cvs-ports/attachments/20040811/039e4baa/attachment.bin