This is the mail archive of the guile@sourceware.cygnus.com mailing list for the Guile project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: kill, again


Russell McManus <russell.mcmanus@msdw.com> writes:

> Oops.  After reading the code, I guess that you can use guile's kill
> for the purpose I wanted.  It's kind of icky to need the 'catch',
> though.  Sorry for useless traffic.

Yep, given the common usage of kill for testing existence of a process,
it does seem a little unfortunate to have to catch an exception.  I
could argue it's not an exceptional circumstance that the process not
exist if the signal is 0, and that perhaps we should return either #t,
throw an exception if the signal is not 0 and the process does not
exist, or #f if the signal is 0 and the process does not exist.  This,
of course, complicates the description of kill, but simplifies a common
case dramatically.  The same simplification could be done in scheme
with an appropriate abstraction, so it remains to decide whether it's
worth changing the primitive to treat the 0 signal specially.  Perhaps
the answer for now is just that there are bigger fish to fry...

Greg

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]