On Tue, 4 Oct 2005, Manuel Freire wrote:

>>> Oh, and seeing that the test framework is ruby-based... joy!
>
> Apart of Curt's idea about an extension language, if people is not
> confortable with Ruby it's very easy to add Python or Perl support now
> and rewrite the tests. This way we could take the Ruby thing out of
> the build.

I don't think it's Ruby that's the issue; I think the issue is that it's
not C, shell, or awk. (Or one of the other primitive languages that are
distributed with our base system.)

I'd be happy to be wrong, though. I've no real objections to switching
from, say, Ruby to Python, though I understand that the C interface for
Ruby is somewhat easier. I'd like to avoid Perl if Python or something
similar is an acceptable alternative, since Perl appears to have no real
advantages over any other scripting language, and has a rather more
hairy syntax.

Anyway, I'd suggest the best resolution to this is to keep on as we are
with Ruby for the moment (I will be glad to provide aid to anybody who
needs it, since I'm pretty experienced with it) and, after we've got a
test system and toolkit, and everybody sees what kind of functionality
we can get from Ruby, people start proposing replacements, and then we
can compare.

My biggest fear is that we're going to compromise our ability to
use this as a toolkit (for debugging, quick prototyping and so on)
because people completely reject the idea of writing anything but
lowest-common-denominator C.

cjs
-- 
Curt Sampson  <cjs@cynic.net>   +81 90 7737 2974   http://www.NetBSD.org
      Make up enjoying your city life...produced by BIC CAMERA