Well, I have some good news and bad news.

The good news is that I finally got accepted into grad school[*]. The bad
news is that I'm going to be busy dealing with reviewing the CS programs
and also dealing with job-related projects, so, starting in a few days, I
won't be able to do much Squeak work. Nor will I be monitoring the list as
closely.

Thus, I've loaded and verified all of my changesets since I started
working on Squeak, 5 months ago. And now I'm going to try to send a
halfway organized dump to the list.

I've done a fair number of things for the image and I have quite a few
changesets. I've updated a few of my past changesets with more comments or
small fixes. If anyone (including any harvesters) has any problems or
questions on these changesets, don't hesitate to CC me the questions.

I also expect to be porting the new method cache to BC2 when we figure out
why I can't get a running BC2 image. I'm also available for any quick
questions or help on the fulltext engines.

--

These changesets have been filed into an image thats been updated from
#4164 to #4646, with the VM rebuilt [**] based off of the UNIX/Linux
Squeak-3.1a-4164 source tree. No other changesets have been loaded.

For the most part, all of these sets of changesets are independent. Except
that MiscEnhancements must be filed in before DemoAdaptors. Other than
that, I filed them all in in alphabetical order without any problems.

I tested the fulltext indexing by filing in all of these changesets into a
clean #4646 image, and running it on both my VM and the stock VM.

--

Yes, the fulltext engines rock. On my VM, with my GC parameters, on a
P2-450, I can fulltext every method (13mb of source) in *6 minutes*, with
60% of *that* spent in CompiledMethod>>getSourceFromFile. Performance on a
stock VM will be about half/third of that.

--


Scott

[*] Rice and UIUC so far.

[**] Due to minor changes in the C-side source, a couple of hacks are
needed to build.