FreeTDS Development Group

["James K. Lowden" <jklowden AT freetds.org>]

On Tue, 1 Oct 2013 04:01:40 +0000
Chris Kings-Lynne <Chris.Kings-Lynne AT navitas.com> wrote:

> Surely you guys have got releasing down to 'make dist' by now? :)

You remind me of a boss I once had who in meetings would ask of some
task, "Who's going to push the button?"   

We actually "make dist" every night; it's been automated for years.  A
release though -- if you're serious about it -- requires 

1.  a review of the changes
2   a summary of them in the release notes
3.  a review of the documentation (which usually involves some fixes and
improvements)
4.  branching in the source code repository
5.  changes to the nightly build (because version changes affect
directory names) 
6   advertisements
7.  updates to the website

I've never done it in less than two days.  

To make it worthwhile, I prefer there to be some major addition to the
feature set, a reason for people to drop what they're doing and upgrade
their FreeTDS installation.  At present, ISTM full support for TDS 7.3
would be a good milestone, but so far the only 7.3 support we have is
in the ODBC library.  

I've been at the helm of this project for about 10 years.  I worked on
testing, error handling, iconv, and low-level communcations issues.
Frediano and I are both fanatical about writing standards-compliant,
safe code, and I would like to think that the care we took has made
FreeTDS reliable and robust.  

We may actually have done too good a job.  FreeTDS may suffer
from the "good enough" syndrome: it could be improved, but it functions
well enough that no one is particularly motivated to hack on it.
Perhaps bugs are better motivators than features?  

I'm sure I've spent more hours answering questions and writing
documentation than I have writing code.  I'm prepared to keep doing
that, but I don't have a lot of interest anymore in tinkering with
DB-Library.  Someone else is going to have to carry that ball.  

I continue to hope someone will, but so far no one has.  Consequently
there isn't much to release, and not much reason to push the button.  

--jkl