Lee Burgess wrote:
> What I mean is, there are two points you make that we should all take
> to heart:
>
> 1) as progress continues on the client spec it grows away from being
> easily compatible with AutoOpt
>
> 2) the idea that we are plugging the subversion client into AutoOpt
>
> Remember, this is just an observation. Maybe I am just being petty
> over semantics; but I think it might be more appropriate to think in
> terms of using an option parser that we can plug into the subversion
> client, rather than the other way around.
Worthy point. I think the difficulty mostly lies in the number
of commands. Each command by itself is easy enough to do.
The problem mostly lies in making sure the Makefile sees all the
different commands. The other issue is that the main() routine
has hand-crafted calls to each implementation routine. That hand-crafted
code has to be hand-extracted and put into the implementation
modules. If the top level were already table driven, I would have
mechanical means to do what I needed. But it is not, so I raise
this issue :-).
Received on Sat Oct 21 14:36:14 2006