[Mdate-devel] New Mdate

[<mda...@li...>]

Okkkk,

I hope someone is still reading this :)

I've totally rewritten Mdate in C++, folding in the last libmdate. In many
ways this is a return to the old mdate distribution with the following
improvements/changes:

1. The command-line parser was rewritten. I generated it with gengetopt and
hacked that. It works really well, and as a result future maintainence and
work will be *much* easier.

2. Autoconf/automake/libtool have been abandoned. I've had too many reports
from users who had too many difficulties with different versions, different
OS's and the size of it wasn't defensible. In future it will be a simple
Makefile with targets, and a config.h file that can be tweaked. A
configure-like script can be written to test only the most critical things,
like getopt(), drem() etc. I've discovered at least one issue with getopt() on
my Cygwin distribution.

3. Using C++ classes means better data protection, overloading. It's really
made not a lot of difference to the library functions (some of which have been
renamed to save bits). It has made things easier to add, like correlation
flags.

4. Optional correlation is implemented. The default is still Lounsbury, but
now users can input -g, --gmt for the GMT correlation or -c,
--correlation=[DOUBLE] for whatever tickles their fancy. There is at least one
other correlation I'm looking at which may get in.

I've had an offer of date/strftime option syntax, which I think is a great
idea, and I'm waiting to see it. I've also had a request to implement the
dreamspell version of dates, which is a bit more tricky, since the correlation
is witheld/uncertain.

I'd like to attic the current CVS trees and add this new version, but for now,
if you want to play with it it's here:

http://www.geocities.com/ewe2_au/mdate/mdate-1.2.9.tar.gz

Please give me some feedback, I want to redo the sourceforge site soon.

Sean
 
-- 
Sean Dwyer <ew...@ca...>
Web: http://www.geocities.com/ewe2_au/