-Coders,
in my current effort to fix the net-snmp 5.2.2/5.3 build/startup issues
on IRIX I stumbled across something that I think needs general
discussion. As outlined in bug 1388562
http://sf.net/support/tracker.php?aid=1388562
IRIX (32-bit) and IRIX64 (64-bit) kernels need different handling
(nlist() vs. nslist64()), but there doesn't seem to be an existing
define (we currently assume IRIX64 which doesn't exist) to make this
decision at compile-time. Even if there was one, a binary compiled for
the N32 ABI can happily run on both IRIX32 and IRIX64 systems. So why
not handle this at runtime instead?
net-snmp already has netsnmp_os_prematch() which seems to allow us to
handle platform specifics at runtime. Naturally, there's a trade-off
between flexibility and performance.
Where do we draw the line between compile-time and runtime decisions? In
the above case, I'd lean towards a runtime decision, but I thought I'd
ask for your valuable comments first.
Cheers,
Thomas
--
Thomas Anders (thomas.anders at blue-cable.de)
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Net-snmp-coders mailing list
Net-snmp-coders@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
