Michael Schuster writes: > this is a question about coding style/best practice, not about use. > > According to our plan, the (initial?) implementation of "ilbadm > show-statistics" will in fact consist of a thin shim of check > arguments-style work in ilbadm, and a ksh script that calls kstat(1M) and > does some munging of the output before presenting it to the caller.
Why wouldn't you call the libkstat(3LIB) functions to read the values directly? > If we strictly follow the precedent all other ilbadm subcommands have set, > we'd have to pass the show-statistics command from ilbadm through libilb > into ilbd, which would then fork a process to run the script and pipe the > output all the way to the caller. While this is certainly feasible, it > looks like a non-trivial amount of overhead that we actually don't need - > I'd advocate that ilbadm directly exec() the script in question and be done. > > I'd like to solicit your opinion on whether this sounds like an acceptable > approach, and if not, what the objections are. Sure; that seems reasonable, as long as libilb itself doesn't really need to be complete. -- James Carlson, Solaris Networking <james.d.carlson at sun.com> Sun Microsystems / 35 Network Drive 71.232W Vox +1 781 442 2084 MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677
