Should the tests in LOCAL-REPLACE be part of configure?

McCall, Don (GSE-WTEC-Alpharetta) don.mccall at
Tue May 15 12:03:26 GMT 2007

Hi Andrew,
I think it's a good idea.  Especially on systems like HP-UX 11.11, where
if we don't expose a function
(like strtoull - it's buried as an internal function in __strtoull) it
MAY be because it's behavior is not strictly compliant...
On 11.23 it IS exposed, and if it isn't working correctly, we'd want to
Better to find this out early - it's a real pain to track it down once
it passes build.

-----Original Message-----
From: Andrew Bartlett [mailto:abartlet at] 
Sent: Tuesday, May 15, 2007 2:40 AM
To: samba-technical at
Cc: McCall, Don (GSE-WTEC-Alpharetta)
Subject: Should the tests in LOCAL-REPLACE be part of configure?

I've been chasing down the build farm issues in LOCAL-REPLACE, and in
particular on our HP-UX machines. 

It seems that our tests for strtoull behaviour are very, very strict,
and if we are going to test for such strictness later, we should also
replace systems that fail these tests in the first place.

Is there any reason not to try and include these into a runtime test?

Andrew Bartlett
Andrew Bartlett
Authentication Developer, Samba Team 
Samba Developer, Red Hat Inc.        

More information about the samba-technical mailing list