>>>>> On Mon, 27 Dec 2010 16:19:16 -0500, Bill Fenner <[email protected]> said:
BF> Whether or not regression testing is feasible is a different issue BF> than whether a regression is considered to be a ship-stopper. That's a fair point. I'll make sure we discuss it during the meeting. It is a slippery slope, still, however... Some regressions are likely when portions of code are rewritten (which happens to MIB code fairly frequently) and may be "minor". We need to balance the line between very slow releases because bugs are continually found (which is true of any software) and making sure they work to a minimal level, even given rewrites. [btw, worst case, you can actually take the code modules from the older release for a given table and stick them in the newer release and they *should* work, assuming they're not making use of new features] BF> (As an aside, we only found the agentx bug after deploying a box BF> with a new image in our own network, with cacti and observium BF> talking to it, and someone observant wondered why we got a log BF> message about reconnecting to the master agent every few days...) Yep. That's the worst kind of bug! The "it only fails under load" bug! BF> (I've qualified my criteria further here so that it's clear that, BF> e.g., the multihomed reply regression doesn't count as stop-ship BF> since although there's a potential fix, we're waiting to hear back BF> from Jan about whether my fix breaks his case). I appreciate the thoughts too. Coding policy for perfection is much worse than coding code for perfection :-P -- Wes Hardaker Please mail all replies to [email protected] ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ Net-snmp-coders mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
