On Tue, 2010-02-23 at 10:04 -0500, Robert Story wrote: > On Tue, 23 Feb 2010 06:58:43 +0100 Magnus wrote: > MF> > I'd like to change that, and search for excluded tokens when -Dall is > MF> > specified. e.g. -Dall,-trace,-wtest > MF> > MF> I'd love that. > MF> I assume that the implication of that is to treat ALL as the name of the > MF> empty string > > ALL currently sets a flag that doesn't do any token checking..
Yes, I know. > The patch I've been using simply adds a loop through tokens looking > for excluded tokens. > MF> -DALL,-foo,foobar > MF> shows all logs but those that start with foo and not foobar > > hmm... now that's a tricky one... I think that we'd want to maintain the > 'first match wins' rule, so that would have to be -DALL,foobar,-foo. Here I would like to disagree - I think 'best match wins' is a better rule to use. It also have the added advantage that it allows binary searching in the token set while first match kind of forces a linear search through the set. > My patch doesn't currently handle that, but it should be a simple change. Currently it is also possible to end up with multiple instances of the same token in the set of enablers. How do you handle foo,-foo (or -foo,foo)? (My patch makes that case impossible to end up in simply by enforcing uniqueness) > MF> Should this affect the mib for nsDebugOutputAll as well? > MF> > MF> Today the description is > MF> > MF> "Whether the agent is configured to display all debugging output > MF> rather than filtering on individual debug tokens. Nothing will > MF> be generated unless nsDebugEnabled is also true(1)" > MF> > MF> and the internal effect is to flip the same variable that -DALL flips. > MF> In order to not change the MIB I think we could keep the -DALL stuff in > MF> the code and only change the effect of the command line option, but on > MF> the other hand I see no real reason to not deprecate the managed > MF> variable as well. > > I don't think we need to deprecate it, but it should probably return 0 if > exclusion tokens are added. This variable is settable, what should the effect of setting it to True be? Should all disablers be disabled or something else? If the answer to the last question is yes, should all the enablers be disabled if it is set to False? /MF ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Net-snmp-coders mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
