Hmmm ok, indeed i saw what the code was doing, but you are probably correct here when you are wondering what could be optimizable! I guess I was too worried about performance, and forgot how streams worked here, of course the stax reader will consume the gzip stream...
Anyway thanks a lot for your help! (And of course the same to Ian) Regards, On Mon, Aug 4, 2008 at 18:24, Daniel Kulp <[EMAIL PROTECTED]> wrote: > On Monday 04 August 2008 11:32:20 am Brice wrote: > > Considering the modification of the input interceptor. > > I cannot say I will have time to work on it indefinetly as I'm staffed on > > other things, but if I want to implement the same intelligent mecanism > used > > to write a gzipped stream. I have somehow planned as follows > > 1 - I think first should create an AbstractThresholdInputStream > > 2 - Optimize the GZipInInterceptor > > 3 - ? Well that's were I'm confused, is it possible to have some > guidance. > > > > Ummm... on the Input side of things, you don't need any type of threshold > thing. The stream of data coming in is either gzipped or not. The > headers > tell us that. If it's gzipped, we wrapp it with the GZipInputStream. > If > it's not, we don't. I'm not sure what would be optimizable here. > > > -- > Daniel Kulp > [EMAIL PROTECTED] > http://www.dankulp.com/blog > -- Bryce
