On Mon, 11 Dec 2000, Greg Stein wrote: > I'm not sure where this semantic was introduced. (I'll blame Ryan :-) > > The intent isn't to create *another* brigade. It is to split a bucket > *within* a brigade. The specific operation that we want to provide for is > within the byterange filter: > > 1) split a bucket at points P1 and P2 > 2) copy buckets from P1 through P2 and add these to the output brigade > > We do not want to destructively modify the original brigade. We are using > split() as a precursor to copying a subrange of the input brigade.
This should not be implemented by a utility function, because it is only really useful for Apache. This should be implemented by Apache itself. Plus, we don't really care if we modify the original data, as long as we don't modify the meaning of the data. Specifically, we don't care if we add two or three buckets to the brigade, as long as the response represented by the brigade is not changed at all. I do not see how the function you are asking for is useful for non-Apache applications. This is something that Apache needs, so Apache should implement it. Cliff is implementing a function that is generally useful. They are not equivalent, nor do they need to be. Ryan _______________________________________________________________________________ Ryan Bloom [EMAIL PROTECTED] 406 29th St. San Francisco, CA 94131 -------------------------------------------------------------------------------
