As part of preparing for a FOSDEM 2010 talk on TPPM outlined here
	http://rpm5.org/community/rpm-devel/3995.html
I've spent a week or two ensuring that I have sufficient process in place
to ensure that I do, indeed, have "working" code @rpm5.org.

This snapshot is most of the "database ACID" first phase of TPPM.

What I most needed to see was the coverage of development over the
last year. This link has LCOV/GCOV reports for current "make check"
coverage
	http://jbj.org/files/rpm/rpm-5.3/lcov/

A "make dist" tar ball (untested, I build from CVS) snapshot of the sources used
for the coverage testing is here:
	http://jbj.org/files/rpm/rpm-5.3/rpm-5.3.DEVEL.tar.gz

(aside)
If you are interested in trying rpm-5.3, I *STRONGLY recommend building
from CVS. I am not going to have time (nor is it useful imho) to distribute
patches against the tarball over the next month or so.

I will use the coverage (and tarball) more-or-less as a reference point
for callgrind perfomance measurements, the results will be reported here
on <rpm-devel@rpm5.org>.

Please expect both the LCOV and tarball to change rapidly over the next month
as I get back to active development rather than boring release engineering. SO
far (see the LCOV results) many code paths in rpmlib are well covered with
"make check". I expect to be able to get ~70% coverage in place for rpm-5.3 if/when I
flip to writing tests rather than code.

With luck, rpm-5.3 will be released in late February or early March. We'll
see, I fully expect the "Berkeley DB haters" to start complaining Real Soon Now, sigh.

Enjoy!

73 de Jeff