Make test-all vs Rubicon

[Booker C. Bense]

_ Is anyone collecting "make test-all" results?
How does this relate to rubicon?

thanks,

_ Booker C. Bense

 

[gabriele renzi]

Booker C. Bense ha scritto:

_ Is anyone collecting "make test-all" results?
How does this relate to rubicon?

rubycon was the name of a test suite for ruby wich was external to the
ruby distribution. Now the test suite is integrated in the interpreter
source and it superceded rubycon. The old rubycon become a rubyforge
project but it should be outdated (i.e. broken) and I'm not sure it has
been updated.

it seem that the RRR page for rubicon still works:
http://www.rubygarden.org/triple-r/
but if you get problems with tests in the test suite in the ruby
distribution you should report them as bugs.

 

[Charles O Nutter]

Rubicon is now the RubyTests project in Rubyforge, and it's under
active development. I just joined recently, and it's my impression
that things had been pretty quiet for some time. Recently, however,
we've done some good work.

Hopefully Rubicon will see more use now that it's actively being
developed again; we've already found a couple issues in core Ruby and
many in JRuby (and cleaned up several bugs in Rubicon) so things are
looking good.

On that note, one question that's come up recently is about the lack
of a formal specification for Ruby...The JRuby team and the Rubicon
team often have to consult many resources, as well as test by hand, in
order to determine the "correct" behavior of a given method. This is
error-prone and certainly not reliable. Does anyone know if a formal
specification of Ruby exists (in any form)? Wouldn't this be something
worth having?

- Charlie

 

[gabriele renzi]

Charles O Nutter ha scritto:

thanks for the precisation

Does anyone know if a formal
specification of Ruby exists (in any form)? Wouldn't this be something
worth having?

I think this is 'no' or the first question, but matz said he'd bless
people that would like to keep such a thing up.