[pybsddb] Distributed transactions using XA [Fwd: OTN Discussion Forum "Berkeley DB" has been updated]

Jeff Johnson jbj at jbj.org
Thu Sep 24 20:53:23 CEST 2009

This is kind of off-topic, but I thought I'd post A FYI anyways.

I'm the lead developer @rpm5.org, with almost a decade of
experience (and scars, sigh) dealing with Berkeley DB.

During that decade, pybsddb is _THE_ best embedding of
Berkeley DB I've seen.

I've almost shopped pybsddb into RPM on more than one occaision.
What stopped me has nothing to do with pybsddb, just, well, an rpmdb
is about the worst possible application usage case or any database.
Buggy code, NPTL stale locks, bare metal empty chroot's and more
all make intelligent use of Berkeley DB almost impossible.

But that ain't what I came here to say ...

I'm actively working on JavaScript bindings for Berkeley DB
much like what has already been achieved with pybsddb
in Python.

If there's any any interest in doing for JavaScript what pybsddb
does for Python, feel free to join in, either here, or (if JavaScript
is somehow OT here) on
	<rpm-devel at rpm5.org>

And if there is interest in a common binding approach for OSS to all
interpreted languages, with a common set of unit test harnesses, etc  
I'm more than willing to try. Note that @rpm5.org already embeds
and likely a few more languages. Ah yes
too ;-)

Yes I well known there is prior art with Berkeley DB everywhere, just,
well, art is always in the eye of the beholder, isn't it?

So my emphasis is on a common approach with equivalent functionality,
not reinventing sorta kinda round wheels. There's some seriously
good and new functionality in db-4.8.24, including db_sql(1) and
db->associate_foreignkey(3) that I'd like not to have to wait several
years for OSS to implement.

Also sent privately, I'm never sure if/when <jbj at jbj.org? mail can be  
sent ...

73 de Jeff

More information about the pybsddb mailing list