[Omake] race condition in omake?

[Nathaniel Gray]

Nicholas Kidd wrote:

> It would seem to me that the dependency analysis between the "clean" and 
> "main" targets is not correct if it sometimes fails. Doesn't the fact 
> that the failure occurs only part of the time imply that the (arbitrary) 
> ordering is missing some dependency?

Yes, definitely.  This is because your "clean" rules almost certainly 
don't describe their side-effects or dependencies to omake.  (Omake 
doesn't try to deduce the semantics of rule bodies like "rm -f *.cm*".) 
  As far as omake is concerned, a "clean" rule's body can be executed at 
an arbitrary time because it has no dependencies and no effects!  Of 
course we know that's not quite true, but we didn't bother to tell omake...

Also, I don't know if omake's dependency analysis engine has a way of 
coping with the idea of files being *removed* rather than created. 
Perhaps it could be done using value dependencies?

Cheers,
-n8

-- 
 >>>-- Nathaniel Gray -- Caltech Computer Science ------>
 >>>-- Mojave Project -- http://mojave.cs.caltech.edu -->