Re: [te-code-users] Discrepancies between .Net documentation and .Net implementation

[Andrew S. T. <ato...@ei...>]

Hi T.J.,

I'm glad you're finding the parser useful.  I had forgotten that I
didn't make a release for pre3 of the .NET edition.

I've not had a lot of time to put into the project for a while, but a
couple of months ago, in my spare time, I started working on the Java
edition again to make sure adequate tests were in place and to perform
some minor refactoring/enhancements requested by a user.  If I remember
correctly, I didn't release the pre3 version because the Java and .NET
got somewhat out of sync as each was contributing refinements to the
other (as I ported bits, other bits ended up needing refactoring in the
original).  However, the existing command.* classes should be stable.

At the moment, I'm up to my eyeballs in some things for work, but in the
meantime, if you have access to a CVS client, CVS head has all of what
you're looking for in (as far as I remember) release condition.  If you
can't get to CVS, I can send you a snapshot bundle, but it would be
easiest for you to check it out yourself.

One thing someone noticed was that the build file isn't compatible with
the latest version of NAnt (which is annoying because they keep changing
the format of the file).  This has been fixed in CVS as well.  What's in
CVS compiles with 0.85-rc3 of NAnt and all of the unit tests currently
pass, except for testEndOfArgsNone for the command parser which after an
hour of messing around with making sure CVS would work for you, I can't
figure out what's busted (I should be working on my other stuff). 
Patches welcome... :)

Immediate next steps for me with the project at the moment are:

1) Finish the refactoring of the Java Edition's command line parser
2) Verify backward compatibility of pre4 with rest using the
still-a-work-in-progress regression tests in
http://cvs.sourceforge.net/viewcvs.py/te-code/functionalqa/.
3) Back-port all of the Java changes into the .NET version as it is
arguably the cleanest of the two
4) Extend the functionalqa test harness to know about compilation of
.NET executables
5) Verify functional compatibility between the Java and .NET editions
6) See what's left to do to sync up the functionality of at least the
command, common, config, filter, io and trace packages.
7) Remove build of unfinished/unstable packages from the core (so you
know which are which)
8) Release 3.0.0-rc1 of both editions.

Of course, there's actually a lot of work involved in all this.  Given
the rate that I'll be able to work on it in the near future (getting
married in December), it will likely be at least March before #8 gets
completed.  That being said, I am reasonably confident that what is
there is stable (there's only been one or two minor issues raised in the
last 12 months that I'm aware of, and they have workarounds).  The above
is more house-cleaning and synchronization between the two.  There
shouldn't be much, if any, major new functionality overall.  However,
.NET will get filters and enhancements to the configuration stuff
present in the Java edition as well as at least some solution for
dealing with the .NET configuration files, but I don't really consider
that major enhancements--it's stuff that was originally intended to be
there from the beginning.

If there's a desire for all this to happen sooner, I can't do it without
help.  I will certainly consider patches/additions from users in this
area (as well as others, but I want to get 3.0 released as priority #1).

Let me know how you get on, and I'm glad to know you're using the
library.  BTW, if you don't minding me asking, how did you find out
about it?

Cheers,

ast

On Fri, 2005-10-28 at 16:42, TJ Brown wrote:
> I downloaded the .Net version of the libraries
> (te-common.NET-3.0.0-pre2) to try the command line parsing classes. 
> I've been very pleased so far, however, I did notice that the
> implementation is missing several classes that are described in the
> documentation.  The missing classes include:  MutexOptionConstraint,
> OptionConstraint, RequiredOptionConstraint, and
> RequiresAnyOptionConstraint.  These classes would be very helpful for
> me.  Would it be possible to make the latest implementation available?
>  
> Thanks and good work.
> T.J.
-- 
Andrew S. Townley <ato...@ei...>
http://atownley.blogspot.com