On Mon, May 3, 2010 at 10:03 AM, Mayan Moudgill <[email protected]> wrote:
> Apropos the discussion on scalar/string compression in the > https://issues.apache.org/jira/browse/THRIFT-110 > I'm curious: if a particular application would tend to compress better > using a different algo than the one(s) provided, what happens? There's not really any variety in the kinds of values that get "compressed" in the sense that the Compact Protocol provides. Small positive and negative numbers are stored efficiently, as are any of the inherently unsigned length/count fields that go with string, binary, list, set, and map. String values themselves remain completely untouched.
