On Wed, 1 Aug 2007 10:53:43 +0100 Joe Orton <[EMAIL PROTECTED]> wrote:
> This bit is also broken - unsetting C-L here is wrong presumably > because the HTTP input filter no longer knows how to fetch the > request body. t/modules/deflate.t in the test suite triggers the > issue; removing that unset fixes it. It's a fair cop. That was indeed broken. The fix is to postpone unsetting C-L until after the first get_brigade from upstream, where it's required. Fixed in r563133 I've also hacked up a draft for another input filter that'll re-insert content-length if it's required - e.g. proxying to a 1.0 server. That's potentially expensive, because it means buffering the input in order to count it, so the filter takes care to run only when strictly necessary. Happy to include it in the core distribution. > Also simply removing content-range is a semantic change, so I can't > see how that is correct. If the filter cannot adjust content-range > appropriately it should either do nothing or fail, I'd say. Yep. That wants another patch, which I'll be happy to do. -- Nick Kew Application Development with Apache - the Apache Modules Book http://www.apachetutor.org/
