[GRASS5] GRASS stable/exp branches

Justin Hickey jhickey at hpcc.nectec.or.th
Mon Mar 5 04:11:14 EST 2001


Hi Radim

Radim Blazek wrote:
> When do you think development in 5.1 will be started (event + date)?
> Yes, 5.1 is almost empty and it will always be until we start there.

I'm not sure of the date, but I feel that moving the code to the 5.1
tree should be the number 1 priority after 5.0 is released. Then
development can continue in the new structure. Furthermore, I think the
safest way to move the code is to prohibit any commits at all while the
code is being moved. Perhaps we can define the new directory structure
as well as which modules will go where. Then we prohibit commits, move
the code, and then set the 5.0 tree to dead. From then on, all commits
will be to the 5.1 tree.

> Will be new features commited into 5.0 (devel branch) released in 
> some 5.0.x?

I would say no, but this is based on my understanding of version
numbers. I could be totally wrong so somebody please correct me if I am.
My experience indicates that the 3 numbers in a version number x.y.z
have specific meanings. The z value is incremented for bug fixes. The y
value is incremented for the addition of minor new features and of
course bug fixes. And the x value is incremented for the addition of
major new features. Thus, version 5.0.1 should contain only bug fixes
and not any new features.

This brings up the point of what version we should associate with the
new directory structure. Since the new directory structure and new
Makefile system is a radical change to grass, then I would suggest that
this code be labeled 6.0 not 5.1, and that all the CVS file versions get
bumped up to 2.0 (or 2.1.1.1 - I don't know how CVS versions work, the
initial version was 1.1.1.1). Since most of the files will reside in new
directories a major increment in their version number is necessary IMO.
If we decide to do this and it turns out that a few new features are
added to the source before we finialize the directory structure, then we
can release a 5.1 version with these new features just before we move
the code.

This is just an idea, but I really don't think such drastic changes
deserve a minor increment in version number.

> 
> Why:
> 5.0b11 - -> 5.0.0 -> 5.0.1 -> ...    (stable)
>            \          ^          ^
>              -> devel branch  - -> stop
>                                        \
>                           -> 5.1 -> x -> 5.1 ->
> 
> instead of:
> 5.0b11 - -> 5.0.0 -> 5.0.1 -> ...    (stable)
>            \
>              -> 5.1 ->
> 

The problem with the second diagram is that you do not indicate where
the switch to the new tree would occur. If the switch is before the
branch to 5.1 then we will have to wait quite a while before we release
5.0, since we need to finalize the directory structure before we can
switch trees. But we want to release 5.0 now, thus this option is out.
If the switch is after the branch to 5.1 then we have the first diagram.
As I indicated before, we need to create a branch if we want to release
5.0 to minimize the type of error that occured with beta 11, thus the
first diagram seems to be the best alternative.

-- 
Sincerely,

Jazzman (a.k.a. Justin Hickey)  e-mail: jhickey at 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 at geog.uni-hannover.de with
subject 'unsubscribe grass5'



More information about the grass-dev mailing list