PATCH: added new Test::Unit compiler plugin

Doug Kearns djkea2 at
Mon May 10 23:56:56 EDT 2004

On Thu, May 06, 2004 at 01:01:03PM +0100, Hugh Sasse Staff Elec Eng wrote:
> I'm coming in to the middle of this, caveat lector (!Hannibal)
> On Thu, 6 May 2004, Doug Kearns wrote:
> > On Wed, May 05, 2004 at 09:43:27AM -0400, Sam Roberts wrote:
> > > Wrote Doug Kearns <djkea2 at>, on Wed, May 05, 2004 at 05:02:42PM +1000:
> >
> > <snip>
> >
> > > > Well, why not set it to the genuine article?
> > >
> > > But I had it set to the genuine article, then the ruby support files
> > > changed it!
> >
> > The 'support files' change all sorts of options that you might have
> > previously set - :help after-directory
> >
> Can this be set in individual files with modeline    vim:makeprg=...
> sort of thing?



> > You seem to be missing my multiple previous statements that I agree that
> > the 'efm' needs to be fixed so that it supports errors as well as
> > failures, and it will be when I get to it, or sooner if someone else
> > does... :-)
> I think I'm the person who created the first version of this, but I
> must add that having looked at it again (a while ago) I did not
> understand how to improve it.

We're referring to the new, partially complete, rubyunit compiler plugin
I cobbled together which currently only works for test failures... As
for the main ruby.vim compiler, I've never had a problem with it and use
it pretty much every day.

I've just had a quick look and my assumption that errors were passed
through unaltered was incorrect. Test:Unit seems to do some reformatting
of errors which is not consistent with the failure format. So, the 'efm'
from the ruby.vim compiler isn't going to help here.


> > > c - you're editing the "main" file of a ruby script, and the script
> > > is runnable with no arguments
> >
> > No, see above. Why does it have to be runnable with no arguments?
> See The Selfish Class:

An interesting read. Is it just an orthogonal musing or did I
misunderstand Sam's point? :-) I read it to mean that the compiler was
not usable with scripts that accepted arguments...

> i.e. it should just work, in the ideal case.

In the case of this Test::Unit compiler, what do you think that means?


> > Ah, a fellow aegis user perhaps?
> Not Aegis as in Apollo, surely?

No, to paraphrase the mutt slogan; all SCM systems suck, but aegis sucks
less than most...


More information about the vim-ruby-devel mailing list