Re: [PATCH] [gdb] Fix heap-buffer-overflow in args_complete_p

[Andrew Burgess <aburgess@redhat.com>]

Tom Tromey <tom@tromey.com> writes:

> Andrew> Maybe one day we'll move off the mailing list model.
>
> What would it take?

I cannot imagine a time in the near future when everyone involved in GDB
would agree to such a change.  But maybe if the majority of the largest
contributors wanted the change then we could figure something out?

>
> What I'd like ideally is that landing patches be done via the UI.
> That way we could have some guarantee that what lands is what is
> reviewed.  I don't think this is possible with the current setup though.

This would be my ideal too.  This opens the door real automated
pre-commit testing.  There are setups which do this sort of thing with a
mailing list (if I remember correctly), but it still relies on people
manually checking prior to pushing.

> For my part I think my main worry is that I'd miss patches that I want
> to be involved in.

True.  But at least for me, that's already the case as....

> Anyway it seems clear that the current approach isn't working extremely
> well.  I don't know about anyone else but I have >3500 "pending"
> messages in gdb-patches, some dating back years.  And probably more than
> half of those are actual unreviewed patches.

...I too have a huge backlog of unread patches.  I am trying to do more
reviews, but I just cannot read patches fast enough.  Anyway, I haven't
investigated, but my ideal system would be managed via a web UI, but
could still send users email when new patches are posted.  That way, if
folks find it easier to track their todo list via email, they could
still do that.

Thanks,
Andrew