This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: MI patch criteria
- To: Andrew Cagney <ac131313 at cygnus dot com>
- Subject: Re: MI patch criteria
- From: Andrew Cagney <ac131313 at cygnus dot com>
- Date: Thu, 28 Jun 2001 00:29:52 -0400
- Cc: GDB Discussion <gdb at sources dot redhat dot com>
- References: <3B291120.8090003@cygnus.com>
> Hello,
>
> Given two non Red Hat MI patches have appeared in the space of less than a week I'd better get my act together and open a discussion about the acceptance criteria for MI changes.
>
> So first a little history. When MI was being developed (within Cygnus) a lot of emphasis was put on testing and documentation. It was intended for a commercial quality product (1). To that end, there were pretty strict rules on when/what could go in: All new commands had to be documented and tested; all bug fixes and to be tested (where possible).
>
> Now that MI is out in the open and part of FSF GDB rather than a pet Cygnus project, I think it is time to table what were then internal to Cygnus criteria and open them up for public discussion.
>
> Personally I'd like to stick to the existing criteria vis:
>
> For new commands: doco + testsuite
> For bug fixes: testsuite (possibly doco)
>
> Andrew
No comment was the loud reply. I'm going to assume this is so. To be
honest I think this is significant and over all positive, MI offers a
robust way of testing LIBGDB as it evolves.
Andrew