Test file numbers

[Michael G Schwern]

Ovid wrote:
> --- Andy Lester <andy at petdance.com> wrote:
> 
>> I've always hated numbering tests, period.  The argument of "There's 
>> no point in running all these other tests if the 000 one doesn't  
>> pass" is the worst kind of worrying about nothing.
> 
> I confess that I'm the guy who set up the numbered tests and I'm quite
> ambivalent about them.  If 000 doesn't pass then perhaps we need it to
> bail out.

I'm fine with obvious numbering like that.  Explicit start/end actions would
be hoopy, too.  Might be something as simple as special filenames, alterable
in the parser.


> Speaking of TH3, what are the chances that we can ship by this weekend?
>  I know we need another test run on VMS.  What else?

I'm going to reserve sweating over VMS until we're feature complete again,
don't want to declare everything on VMS working only to have to do it all over
again because we changed the guts.  That means waiting for the parallel stuff.


-- 
Stabbing you in the face for your own good.