[GRASS5] GRASS 6.0.1 release question

[Markus Neteler]

On Tue, Jul 05, 2005 at 04:30:48PM +1200, Hamish wrote:
> > > > > e.g. "GRASS does not compile with gcc 4.0" is a critical fix.
> > > >
> > > > Is it really critical for 6.0.1? Which distro comes with gcc4?
> > > > Since it's half done only for 6.1, I suggest to postpone it
> > > > for 6.0.2 so that we have time for testing.
> 
> AKAIK:
> Fedora Core 4, MacOSX 10.4; Debian's development branches, but they are
> currently undergoing a lot of change so I think we can get away without
> supporting them for a little while.

Ok, but today a major part of the users will still use Fedora Core 3
and MacOSX 10.3. Releasing 6.0.1 now helps for all these users
(also Mandriva, Windows etc pp).

etc. I guess I don't have to repeat :-)

> It's easy enough to patch 6.0.1 with
> the existing gcc4 fixes and let any problems get sorted out in the
> DebianGIS repository before committing to Debian/unstable; and if that
> goes smoothly look at putting them in the GRASS 6.0 release branch.
> (use testing for testing and release branches for stable code after all)
> 
> Debian has a gcc-3.4 package that can be installed along side a default
> install of gcc-4.0 too.....
> 
> 
> > Well, I still think that it is a good idea to get the current
> > code published. Changes for gcc4 are quite intrusive (correct me
> > if I am wrong), are not yet finished, not tested completely etc.
> > There is a possibility to publish 6.0.2 for that.
> 
> Although I don't have any experience with them, the gcc4 fixes are
> indeed new and lightly tested.

I think Brad is doing a great job. But even he says that things are
not complete yet.

> So there remain several bugs in various places that we don't like in
> 6.0.1, but I agree there is no strong reason to wait. As you say, there
> can always be a 6.0.2. In fact I think we should certainly keep the
> 6.0.x line alive for a while and there should one-day be a 6.0.3 and
> even .4 -- give you something stable to write a 3rd ed. book on e.g.

We are thinking about it :-)
 
> If d.where is giving wrong results, I do hope we can fix that before 
> we get to 6.0.1-final ... Segfaulting etc is an inconvenience, but
> silently reporting the wrong answer is very bad.

I agree.
 
> On the other side, should we plan on gcc4 support for 5.4.1? --As 5.4.1
> might be the last of the GRASS 5's? Depends on how serious the imagery
> lib changes are I guess.
> 

No idea. Paul?

Markus