Hi Graeme, Pascal, forgot CC to list - A. Quoting [Andre Merzky] (May 08 2006):
I checked all usages of NULL in the spec, and all but two can be safely removed, and replaced by empty strings etc.
The two places are:
1) wait on a stream_server returns a stream if successfull, and NULL otherwise. I don't see any nice way to get rid of that NULL - I think that should be dealt with in the language bindings.
2) remove_callback on a metric returns the callback removed. If that cb was removed earlier (or removed itself after execution), the method returns NULL. Same here I think: check in the language bindings.
Does that sound ok?
Cheers, Andre.
Quoting [Andre Merzky] (Apr 27 2006):
Date: Thu, 27 Apr 2006 21:52:23 +0200 From: Andre Merzky
To: G.E.POUND@soton.ac.uk Cc: Andre Merzky Subject: Re: ISSUE 131 Ah, ok - I retract the Java argument then :-)
Thanks, Andre.
Quoting [G.E.POUND@soton.ac.uk] (Apr 27 2006):
Andre
'null' is a valid value in Java (in most instances). It is not a valid value in FORTRAN.
Graeme
Quoting Andre Merzky
: Quoting [Andre Merzky] (Apr 20 2006):
131) clearify what a NULL string/object is we sometimes return - OPEN, URGENT
iObvious again. Opinions? Should that even go into the API spec, or into the language bindings? What should we say in the spec?
No opinion.
I think NULL does not make sense for Java and Fortran (Is that true?) so I guess it should stay out of the general spec.
I changed all occurences to either throw, or empty string returns etc. There are only a couple of places where we used NULL anyway.
Volonteers?
Cheers, Andre
Cheers, Andre.
-- "So much time, so little to do..." -- Garfield
-- "So much time, so little to do..." -- Garfield
participants (1)
-
Andre Merzky