[GRASS5] Re-release of beta11

GRASS GRASS developer
1

Hi all,

due to the still existing GISBASE/dev bug I have to re-release
beta11 (already removed from web). As the create_*.sh scripts have not
been copied into the binary package, the grass5install.sh wipes out
the fifos without being able to reproduce them (was working before).
This simply makes GRASS beta11 unusable. :-((

Lucily the beta11 was not yet officially announced.
So I am preparing beta11b now.

Next time please don't break GRASS few minutes before releasing.
We already had two test phases, after that the bug was introduced.
Such heavy changes (related to IPC/fifos) should take place
*after* releasing a new version. Otherwise a new test phase would
be required.

As we already had severe bugs in beta10, I simply don't like that
the beta11 from sunday was also not working. That's no good reputation.

We need more either more discipline when submitting changes or
lock the CVS completely during the test phase to avoid such problems.

It takes me hours to build such binary release, update everything etc.
Now I have to do it again for beta11... (I have other work to do as well!)
And the users get confused: "why another beta11?".

So far the status,

Markus

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

2

On Tue, Feb 06, 2001 at 05:02:23PM +0000, Markus Neteler wrote:

due to the still existing GISBASE/dev bug I have to re-release
beta11 (already removed from web). As the create_*.sh scripts have not
been copied into the binary package, the grass5install.sh wipes out
the fifos without being able to reproduce them (was working before).
This simply makes GRASS beta11 unusable. :-((

Lucily the beta11 was not yet officially announced.
So I am preparing beta11b now.

No. Please release it as beta12!
Because I have got a beta11 release tarball on my machine.

There is no shame in bumping the version number up.
We might have gone with another pre release, if we have have felt
the need to test a bit further.

  Bernhard

--
Professional Service around Free Software (intevation.net)
The FreeGIS Project (freegis.org)
Association for a Free Informational Infrastructure (ffii.org)
FSF Europe (fsfeurope.org)

3

On Tue, Feb 06, 2001 at 05:50:11PM +0100, Bernhard Reiter wrote:

On Tue, Feb 06, 2001 at 05:02:23PM +0000, Markus Neteler wrote:
> due to the still existing GISBASE/dev bug I have to re-release
> beta11 (already removed from web). As the create_*.sh scripts have not
> been copied into the binary package, the grass5install.sh wipes out
> the fifos without being able to reproduce them (was working before).
> This simply makes GRASS beta11 unusable. :-((

> Lucily the beta11 was not yet officially announced.
> So I am preparing beta11b now.

No. Please release it as beta12!
Because I have got a beta11 release tarball on my machine.

O.k.

There is no shame in bumping the version number up.

It *is* in my opinion. As it becomes happening regularly.
As we always say: latest version before stable release,
we are confusing the users. I definitly dislike this.

We might have gone with another pre release, if we have have felt
the need to test a bit further.

Well, there is much pressure to get out the stable. Even
beta11 was taking too long (you reminded me here :-). So
if again having a twofold test-phase, it takes March to get
out beta12. Currently there is much waste of time in this.
Sorry, probably I am too overworked... but this is all a bit
odd as we regularly break promises.

Markus

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'

4

Hi Markus

Markus Neteler wrote:

Hi all,

due to the still existing GISBASE/dev bug I have to re-release

Actually, it was fixed on Monday in version 1.77 of Makefile.in, but of
course that was too late for the beta11 tag.

We need more either more discipline when submitting changes or
lock the CVS completely during the test phase to avoid such problems.

I agree that we need to keep new code out of the sources until a release
is announced once we enter the pre-test phase. Demanding more discipline
from developers (that is, we do not commit changes unless they fix a bug
discovered by a pre-tester) is one thing we can do. However, it probably
won't catch all problems since somebody is bound to forget. Locking the
CVS tree completely is not a good option since fixes for bugs discovered
by the pre-testers can't be fixed easily.

The only other option I know of is that once we decide to send the code
to the pre-testers, we create a branch of the CVS tree and developers
can only commit fixes for bugs found by pre-testers to the branch. Then
once the release is established and announced, we merge the branch back
in. I'm not sure if this is feasible, but it seems to be the ideal
solution to me. What do people think?

Also, do we have a step by step procedure for our pre-testers? If not,
then how about something like

For the source code
o Download and unpack new distribution in a new directory
o Run configure
o Run make
o Run make install
o Check if Grass runs (does "make check" work?) - maybe better to have
  this before "Run make install"
o Run make uninstall
o Run make distclean

For the binary distributions
o Download and unpack new distribution in a new directory
o Run grass5install.sh
o Check if Grass runs (maybe use testgrass.sh or some script created
  from the testsuite directory)
o Run grass5uninstall (new feature in CVS tree, not in beta 11)

Anyway, just my 2 cents worth

--
Sincerely,

Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, Thailand

People who think they know everything are very irritating to those
of us who do. ---Anonymous

Jazz and Trek Rule!!!

----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'