[V3dot10] needed tool
johnson at cs.uiuc.edu
Sun Jan 21 20:00:32 UTC 2007
On 1/21/07, Keith Hodges <keith_hodges at yahoo.co.uk> wrote:
> Ralph Johnson wrote:
> > Markus said that one of the big problems during harvesting is that
> > when you read change sets into your image, some of the methods get put
> > into unexpected MC packages, and the harvester doesn't save those
> > packages and so loses the code.
> This is a limitation of the linear incremental development process that
> has previously been following. You harvest, you save, you release an
> update, you harvest you save you release an update, you discover that in
> a previous save you missed saving the appropriate package, so you have
> to go back and retrieve it and sort out the mess.
> With the build script approach, the script loads in a load of fixes, and
> then as a later script step, saves ALL the packages, then you don't
> loose anything.
I can see how it loads in a set of fixes, but how would it know which
packages to save? Since it is repeatable, you can always go back and
change the script if you discover that you made a mistake, but in fact
3.9 has a script and you can change it, so there isn't much
One of the differences is that the 3.9 script points to MC packages
instead of change sets. Mantis mostly contains change sets. So, the
change sets in Mantis had to be converted to MC. It seems to me that
if you mostly use MC then you'll always have to convert change sets to
MC and will have the same problem.
More information about the V3dot10