>> yeah, I wasn't quite sure what to do here. I didn't like the idea of >> length() returning an $rv and populating a $length argument, since it >> seemed >> more useful to simply have it return the length. but I'm deviating >> from the >> API here, so I dunno. > > > at least mark it with XXX, so we know it's an untied end. We will how to > deal with it better as this API will be used.
done >> hmm, the length passed ends up being allocated? I was just trying to >> simulate a user that used a big number since the size was unknown (and >> who >> didn't know about length()) > > > ok then. I'd add at least a comment that one should pass either a small > known size (if they want to get at most X bytes) or call the length > method. some people use tests as examples, while we don't have the > normal docs up. > >>> I don't think flatten is effective, as it has to go through all the >>> bucket twice - once because you need to figure out its length, second >>> time to slurp data. ok, I added some comments in the test > I think your current patch (plus a few comments) is fine to go in. Can > do pflatten on the next iteration. committed. I changed the test name to flatten.pm, though, since it seemed better than the generic brigade.pm. --Geoff --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
