Re: Testing with concurrent sessions

Hi,

Kevin Grittner wrote:
> I just got to the point of having what appears to be a working but
> poorly optimized version of serializable transactions,

Cool.

> so it is
> critical that I create a good set of tests to confirm correct
> behavior and monitor for regressions as I optimize. I see that
> you've been working on dtester recently -- should I grab what you've
> got here, stick with 0.1, or do you want to package something?

I'm preparing for a 0.1 release, yes. However, I'm trying to keep up
with the postgres-dtests branch, so that should continue to work with
the latest dtester code.

> If I
> should pull from your git, any hints on the best git statements to
> merge that in are welcome, I'm still rather new to git, and I tend
> not to get things right on a first try without some hints. :-/

I'd recommend merging postgres-dtests into your branch, so you'll get
all of the adjustments from there, if I change dtester itself.

Once 0.1 is out, I'm *trying* to remain backwards compatible in future
releases. However, that holds me from releasing 0.1 in the first place :-)

Independent of versions and releases: if you build your tests on top of
dtester, I'm more than willing to support you with that. If you keep
your code in a git repository, I could even provide patches, in case I
need (or want) to change the dtester interface.

Regards

Markus Wanner