Mark Atwood, checking in, and requesting checkin

Eric S. Raymond esr at
Mon Feb 29 19:57:16 UTC 2016

Mark Atwood <fallenpegasus at>:
> My apologies for dropping the ball last week, this week I will tag and
> release 0.9.3, probably on Wednesday.

The delay was not a bad thing.  We had a couple good documentation fixes
and a build cleanup land- nothing major, but nice.  
> I am now managing the open source engagement for the OpenSwitch project,
> and I plan on arranging to have ntpsec be the implementation that
> OpenSwitch uses.

> I notice that Coverty scan has found two warnings.  Can someone open bugs
> for them, and then fix them?

Already been done - Matt Selsky developed a fix and I merged it.  I'd
have done a fix myself sooner, but the code that was generating the
warnings is going to go away when replay lands.  It was
instrumentation of mine, not timekeeping logic.

> Eric:
> - what is the status of testframe
> - what else have you been working on this week?

Testframe is proceeding, slowly.  Every time I think I'm converging on a
final version (so far) I trip over some internal cohesion I have to separate;
it's tricky work to do that without breaking existing functionality.

It has been a busier week than the last several for dealing with
(other) merge requests and issues.  Nothing serious, just a lot of
details to track and manage.

Last weekend I co-wrote and released a HOWTO for people who need to
forward-port Python 2 programs that handle binary data to Python 3.
Because of the big string-to-Unicode change this is a bit tricky.

Practical Python porting for systems programmers:

NTPsec utilities are exactly the use case the HOWTO addresses,
actually, and that's no coincidence. While I tested the methodology on
SRC and reposurgeon, in the near future we're going to want to move the
archaic Perl in our scripts directories to something maintainable, and
polymorphic Python (that is, Python code that auto-adapts to 2 or 3)
is the best candidate.

Now I not only know how to write polymorphic Python that is 8-bit-clean for
binary streams, I've written the technique down so we can delegate the utility
		<a href="">Eric S. Raymond</a>

More information about the devel mailing list