Changing Netty 4 HTTP File Server example to use ChunkedStream instead of ChunkedFile -

i'm trying wrap head around netty 4 way of implementing http server serve httpresponses bodies using chunked transfer-encoding when total data size unknown.

as starting point, changed httpstaticfileserverhandler (found in https://github.com/netty/netty/tree/netty-4.0.0.cr1/example/src/main/java/io/netty/example/http/file) use chunkedstream instead of chunkedfile (they both chunkedbyteinputs).

i understand not ideal in original example use case use fileinputstream think makes example reusing known code.

so, here diff against httpstaticfileserverhandler class io.netty.example.http.file package (vs. 4.0.0.cr1):

diff --git a/example/src/main/java/io/netty/example/http/file/httpstaticfileserverhandler.java b/example/src/main/java/io/netty/example/http/file/httpstaticfileserverhandler.java index 904579b..0d3592f 100644 --- a/example/src/main/java/io/netty/example/http/file/httpstaticfileserverhandler.java +++ b/example/src/main/java/io/netty/example/http/file/httpstaticfileserverhandler.java @@ -27,13 +27,14 @@ import io.netty.handler.codec.http.fullhttpresponse;  import io.netty.handler.codec.http.httpheaders;  import io.netty.handler.codec.http.httpresponse;  import io.netty.handler.codec.http.httpresponsestatus; -import io.netty.handler.stream.chunkedfile; +import io.netty.handler.stream.chunkedstream;  import io.netty.util.charsetutil;   import javax.activation.mimetypesfiletypemap;  import java.io.file; +import java.io.fileinputstream;  import java.io.filenotfoundexception; -import java.io.randomaccessfile; +import java.io.inputstream;  import java.io.unsupportedencodingexception;  import java.net.urldecoder;  import java.text.simpledateformat; @@ -159,17 +160,15 @@ public class httpstaticfileserverhandler extends channelinboundmessagehandlerada              }          }  -        randomaccessfile raf; +        inputstream raf; // use inputstream instead of randomaccessfile          try { -            raf = new randomaccessfile(file, "r"); +            raf = new fileinputstream(file);          } catch (filenotfoundexception fnfe) {              senderror(ctx, not_found);              return;          } -        long filelength = raf.length();           httpresponse response = new defaulthttpresponse(http_1_1, ok); -        setcontentlength(response, filelength);          setcontenttypeheader(response, file);          setdateandcacheheaders(response, file);          if (iskeepalive(request)) { @@ -180,7 +179,7 @@ public class httpstaticfileserverhandler extends channelinboundmessagehandlerada          ctx.write(response);           // write content. -        channelfuture writefuture = ctx.write(new chunkedfile(raf, 0, filelength, 8192)); +        channelfuture writefuture = ctx.write(new chunkedstream(raf)); // use chunkedstream instead of chunkedfile           // decide whether close connection or not.          if (!iskeepalive(request)) { 

and here complete changed file: https://gist.github.com/eskatos/5311587

changes minimal: use fileinputstream instead of randomaccessfile , chunkedstream instead of chunkedfile. pipeline untouched.

to reproduce, apply changes netty example, run , try download file.

after change, directory listing works because reponses not chunked file downloads don't. client download file never finish download, hold connection , wait forever. i've tried several browsers curl, wget etc.. i've tried add bytelogginghandler pipeline , can see empty trailing chunk don't understand why browser still wait data.

any clue?

in order terminate chunked transfer of unkown size (content-length not known , not specified) send empty chunk last chunk. permits keep connection open in order support keepalives:

        ctx.write(new chunkedinputadapter(new chunkedstream(raf, 8192)));         channelfuture writefuture = ctx.writeandflush(lasthttpcontent.empty_last_content); 

here complete example: https://github.com/scireum/sirius/blob/develop/web/src/sirius/web/http/response.java#l649