I have an IHttpHandler returning a file. When the response stream is compressed, either automatically using Telerik RadCompression or by explicitly setting a filter using
context.Response.Filter = new GZipStream(context.Response.Filter, CompressionMode.Compress);
the response returned to the browser is correct but at the end of the response is some HTML. The HTML contains the exception:
[HttpException (0x80004005): Invalid use of response filter] System.Web.HttpResponseStreamFilterSink.VerifyState() +3928894 System.Web.HttpResponseStreamFilterSink.Write(Byte[] buffer, Int32 offset, Int32 count) +28 System.IO.Compression.DeflateStream.Dispose(Boolean disposing) +363 System.IO.Stream.Close() +28 System.IO.Compression.GZipStream.Dispose(Boolean disposing) +63 System.IO.Stream.Close() +28 System.IO.Compression.DeflateStream.Dispose(Boolean disposing) +595 System.IO.Stream.Close() +28 System.IO.Compression.GZipStream.Dispose(Boolean disposing) +63 System.IO.Stream.Close() +28 System.Web.HttpWriter.FilterIntegrated(Boolean finalFiltering, IIS7WorkerRequest wr) +754 System.Web.HttpResponse.FilterOutput() +121 System.Web.CallFilterExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute() +174
If I make sure the response is NOT compressed, the response contains no exception.
What's up with that?
Make sure you're not preventing buffering. Apparently, to compress the response, you have to allow it to be buffered.
I.e., one of the following two lines needs to be removed:
It seems to be related to this: GZipStream is cutting off last part of XML
I removed the compression module and it seem to be all good now.