On Tue, Jun 9, 2009 at 4:34 AM, Reinier Lamers <[email protected]>wrote:

> Hi all,
>
> On Tuesday 09 June 2009 12:55:04 Petr Rockai wrote:
> > Eric Kow <[email protected]> writes:
> > > Sorry to drag this out some more.  Just remember that I'm only here to
> > > offer alternative suggestions if they can break us out of a loop.
> >
> > All I can say now is patches welcome. I refuse to spend any more time on
> > this (in comparison to the size of the "undefined" subthread, the code in
> > question is utterly irrelevant... can I say bikeshed now?).
>
> This is not just about your code. If this supposed bikeshed reaches a
> conclusion, that conclusion will be a rule for all code that comes in the
> future. A general rule that there shall be no undefineds in darcs' source
> could be a useful result. Or a rule that every fromJust should have a
> comment
> explaining why it is really safe (perhaps haskell_policy already does that,
> not sure).


By the way, we have a macro, "impossible" for cases that should never be
reached.  Not my favorite because it doesn't take a description, but it does
let the reader know when something should not happen.  And it prints the
line number when darcs crashes.

Jason
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to