[asterisk-dev] Re: Pthread wrapper updates

[Andrew Kohlsmith]

On Tuesday 31 October 2006 14:50, Tim Panton wrote:
> Sure, so a 'new' call fails (possibly horribly), but any in-progress
> calls will continue until they
> need to access the problematic resource (possibly never). If you
> abort, then all in progress
> calls fail needlessly.

Yes, and if it does fail, nothing is logged so when the box crashes strangely 
and unreproducibly, there is no data trail.

> (As someone said to me once - "We are doing software engineering
> here, not computer science." )

While I agree with the sentiment of the comment, common sense must prevail and 
at the very least, a message emitted to the logs (and NOT a debug log, a 
regular old message log!) -- There is absolutely no value in logging this 
kind of low-frequency occurence to a resource which is only active when 
you're trying to fix a problem.  

-A.