On Tue, 12 Oct 2010, Mark D. Nagel wrote: > I am looking at how to handle a case where rulesets will be applied to > multiple input sources (logical or actual) via jump, but those rulesets > may create and use context names and variables that would conflict > across those input sources. How I have handled this prior to jump is > with multiple instances of SEC. This works, of course, but with some > scalability problems depending on the number of partitions needed. I > hoped to use jump to reduce that scalability problem, but this naming > conflict is not solved by jump. What I think may be a good solution > would be to add a namespace option to the jump rule:
why can't you make the context names unique? you can add whatever you would use as the namespace as part of the context name. David Lang > # somewhat contrived jump rule - logically split input source > type=jump > pattern=^([^-]+)-\S+ > namespace=$1 > cfset=some-rules > > In this case, the examined log line would have <source>-xxx at the start > of each line, so would set the namespace to <source> and jump to > some-rules, which might set variables, contexts, etc., but they would > all be implicitly scoped to <source>:<variable/context name>. The > namespace would be valid until the next namespace definition is > encountered (i.e., dynamically scoped). The namespace would have to be > pushed and popped to deal with a jump that includes continue=takenext so > that the previous namespace would be valid after the jump target > finishes processing. There would probably also need to be a way to > reference the global namespace, similar to $::<var> in Perl, and may as > well allow references to other namespaces, though my primary purpose > here is to hide the idea of namespaces from called rulesets. > > Does this seem like way too complex of an idea? I can make it work as I > have already with distinct instances, or by perhaps by manually setting > up prefixes to variables and contexts similar to this. It's just that > it would be complicated and error-prone, and as soon as I thought of > doing that, it occurred to me that a namespace solution would be much > more elegant and robust. Thoughts? I'm sure I'm forgetting something > that would make this horribly problematic. > > Thanks, > Mark > > ------------------------------------------------------------------------------ Beautiful is writing same markup. Internet Explorer 9 supports standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. Spend less time writing and rewriting code and more time creating great experiences on the web. Be a part of the beta today. http://p.sf.net/sfu/beautyoftheweb _______________________________________________ Simple-evcorr-users mailing list Simple-evcorr-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/simple-evcorr-users
