Finally I'm making the progress public *officially*  :P

I've written some posts in my blog about the bootstrap progress:

http://playingwithobjects.wordpress.com/2012/06/08/the-bootstrap-chronicles-chapter-1-preparing-the-soil/
http://playingwithobjects.wordpress.com/2012/06/11/the-bootstrap-chronicles-chapter-2-do-not-mess-with-the-vm/
http://playingwithobjects.wordpress.com/2012/06/17/the-bootstrap-chronicles-chapter-3-its-alive/

The last post is probably the most interesting one, telling how you can try
the code *(at your own risk)*.

If you do not like reading long posts, here is a summary:

- code is available to bootstrap a pharo image.
- the bootstrap takes a sample on the current running pharo image.  It does
not start from source code yet.
- the bootstrapped image doe some nice reporting like space tally and a
list of broken references and not initialized stuff. This list is a point
to start improving the pharo core packaging, to make the core not dependent
on external packages.
- the size of the bootstrapped image fluctuates between 1.1 and 2 megabytes
depending on the initialization and sample you take to bootstrap. This is a
point to attack using the spacetally and some other reports probably.

You can also see a jenkins job at Ecole des Mines where that report is
printed as a Junit report. As long as that list has items to fix, the job
will remain yellow :).

http://car.mines-douai.fr/ci/job/Seed%20Broken%20References%20Report/
http://car.mines-douai.fr/ci/job/Seed%20Broken%20References%20Report/lastCompletedBuild/testReport/

My next steps are making the bootstrap start from a source code
specification and work a bit on modularization.

Enjoy,
Guille

Reply via email to