Re: CVS debacle

[sml13 cornell edu]

On Tue, 15 Dec 1998, Jesse D. Sightler wrote:

> Note, however, that I do not believe that this applies to compile time
> issues.  Any developer throwing things into the code that are not compilable
> should be burnt at the stake.  Well ok, maybe that's SLIGHTLY too harsh, but
> really, a Mozilla style Tinderbox "blame" system would be both very nice.
> 

YES, YES, YES!!  Has anyone approached VA Research about maybe giving us 
a discount on a YMP system that we can use to set up a Tinderbox for 
constant builds so this 
haphazard, randomass comitting of non-compileable changes leads to some 
serious public humiliation (a la http://www.mozilla.org/tinderbox.html)?  
I agree that at least compilation-level (i.e., does this shit that I am 
about to commit cause the tree to fail to compile) regression tests are 
in order for _any_ project, free or not!

I think Tinderbox ought be implemented ASAP (hey, if my machine had the 
horsepower for it, I would gladly volunteer :-).

shane