Re: vast numbers of unimplemented MI commands.

[Joel Brobecker <brobecker@gnat.com>]

> What is the best solution? ...

Somebody like you for instance (who understands the needs of GDB
frontends because he's developping one) who also invests time into
making GDB/MI better. See for instance one of the commands that
our GVD (our graphical front-end) needed: What we did was simply propose
a new command, discussed it, and then implemented it.

It's true that our example is a bit biased because the alternative way
is a real pain compared to what the new command provides, so one could
say there was no real work-around to the problem. But still, this is how
things moved forward for us.

Could it work for you?

My suggestion would be that you implement all the functions you need,
and also make suggestions when something is missing... I think the
current speed of MI development is partly because few of its users
send any suggestion or complaint - and the fact that one can "work
around" any missing command doesn't help. And even if somebody sends
suggestions, I'm not certain anybody will stand up and implement the
suggestion unless he has any incentive to do so: And my guess is that
the only people who could feel this incentive are the developpers
working on a GDB frontend. Again, people like you.

-- 
Joel