How to proceed for the kernel cleaning harvesting

Hi doug and others

We are starting to make progress on the kernel cleaning. However I 
would like to avoid to have 200 changesets to be included before 
putting some of them in the stream.

What is important is that we are decomposing our changes (with pain 
instead of doing three changes at once we really do them one by one 
slowly) into
small changesets so that people can understand them. We are also 
building a testSuite
that we will keep in the future as a kind of specification of what we 
did. We are four working on that and we have our own internal 
reviewing. We are also introducing a deprecation schema so that tools 
build before the cleaning will be notified when they used an old method 
and that people can query the deprecated method. This way we could 
build a simple tool to identify deprecated methods and build a database 
with that information for later migration.

Now the problem is that cleaning the kernel is somehow special and the 
order of the changesets may be really important. So change 0001 should 
be put in before 0004 for example. We are focusing on specific parts 
and we have some major milestones as explained on the page. 
http://minnow.cc.gatech.edu/squeak/3083

We can document what we are doing but if for every method move we have 
to write 10 lines this will never scale. Normally there is a preamble 
that explain the change but often we cannot spend time motivate them 
because this is obvious. So I would like to know if some harvesters are 
really interested in reviewing what we are doing, we are looking for a 
meta-aware one. I asked Noury (the best meta programmer I know in 
france father of metaclassTalk which already cure some part of the VW 
kernel for his PhD) to be an external reviewer. But he is not an 
harvester but could become one (Noury?)

So as soon as our own reviewing process will finish for the first 40 
changesets we will send them into the stream. But we would like to know 
how to minimize energy.


Stef




Prof. Dr. Stéphane DUCASSE
http://www.iam.unibe.ch/~ducasse/
  "if you knew today was your last day on earth, what would you do 
different? ...  especially if,
  by doing something different, today might not be your last day on 
earth" Calvin&Hobbes

"The best way to predict the future is to invent it..." Alan Kay.

Open Source Smalltalks: http://www.squeak.org, 
http://www.gnu.org/software/smalltalk/smalltalk.html
Free books for Universities at 
http://www.esug.org/sponsoring/promotionProgram.html
Free Online Book at 
http://www.iam.unibe.ch/~ducasse/WebPages/FreeBooks.html