Re: [Gnushogi-devel] [Patch 0/1] Switch to getopt cmdline parsing

[ydirson]

Hi Justin,

> I started using gnushogi to play shogi and noticed some slight
> issues, but looking at the commit history doesn't seem to be activly
> developed.

There are sure issues, and you're welcome to expose them :)


> It only seemed polite to send a patch when inquiring about the
> development status so here's a first pass at switching to getopt for
> argument parsing found in the todo.

As noted by others, yes there is not much activity around here :) I
wanted things to be as good as possible, which ended up with many
patches sleeping in branches (including contributions), esp. as I got
less and less time to advance them.

On another front I had tries to do some archeologic work on how the AI
was able to get from a state where it was apparently not bad (7th out
of 22 runners at WCSC in 1994, ie. probably with 1.2p02 or maybe an
intermediate between 1.2p02 and 1.2p03), until its current state.
It's not yet clear to me if it just stagnated while the rest of the
Shogi-AI world progressed, or if some changes to its engine caused
more problems than they solved.

It would really take time to make it better, and as HGM wrote, it's
not sure how much worth the effort would be.  It's been some time I
wanted to run a couple of games between those old versions and the
current code, never got around to it...


> The main difference I've noticed is that this patch produces and
> error & usage when '--' options are used where the current method
> seems to ignore invalid options and just produce the usage text.
>
> This hasn't been fully tested, just wanted to gauge the reaction
> before I spent time testing and cleaning it up.

Yes there are surely some difference in option syntax when switching
to getopt.  But then, as old syntax is flagged as error, and that it
does not impact latest versions of the various GUI (which boils down
AFAIK to xboard, xshogi, tagua), it should be fine.

Best regards,
-- 
Yann