httptunnel 3.3 doesn't work with Apache 1.3.33 mod_proxy

[Szabó Péter]

Hi!

Since Apache mod_proxy flushes the HTTP request bodies too
slowly, httptunnel doesn't work with it. See my patch to
mod_proxy attached.

However, I think that it is not mod_proxy that should be
patched, but httptunnel. That's because httptunnel should
be able to communicate in a proxied environment, where the
proxy buffers request bodies. Now httptunnel doesn't work
this way, so I think this is a bug.

This is not a new bug at all! It was reported in 2001
(http://marc.theaimsgroup.com/?l=httptunnel&m=100118684804708&w=2),
and it still persists.

Best regards,

Péter Szabó

--- mod_proxy/proxy_http.c.orig 2006-11-27 16:48:37.000000000 +0100
+++ mod_proxy/proxy_http.c      2006-11-27 16:48:49.000000000 +0100
@@ -386,20 +386,21 @@
      * if this happens we ignore the 100 continue status line and read the
      * response again.
      */
     {
         /* send the request data, if any. */
         ap_hard_timeout("proxy receive request data", r);
         if (ap_should_client_block(r)) {
             while ((i = ap_get_client_block(r, buffer, sizeof buffer)) > 0) {
                 ap_reset_timeout(r);
                 ap_bwrite(f, buffer, i);
+                ap_bflush(f);
             }
         }
         ap_bflush(f);
         ap_kill_timeout(r);
 
 
         /* then, read a response line */
         ap_hard_timeout("proxy receive response status line", r);
         result = ap_proxy_read_response_line(f, r, buffer, sizeof(buffer)-1, 
&backasswards, &major, &minor);
         ap_kill_timeout(r);
--- mod_proxy/proxy_util.c.orig 2006-11-27 16:49:13.000000000 +0100
+++ mod_proxy/proxy_util.c      2006-11-27 16:49:56.000000000 +0100
@@ -592,25 +592,27 @@
                 ap_log_rerror(APLOG_MARK, APLOG_ERR, c->req,
                               "proxy: error writing to %s", c->tempfile);
                 c = ap_proxy_cache_error(c);
             }
             else {
                 c->written += n;
             }
         }
 
         /* Write the block to the client, detect aborted transfers */
+        ap_bflush(con->client);
         while (!nowrite && !con->aborted && n > 0) {
             if (alternate_timeouts)
                 ap_soft_timeout("proxy send body", r);
 
             w = ap_bwrite(con->client, &buf[o], n);
+            ap_bflush(con->client);
 
             if (alternate_timeouts)
                 ap_kill_timeout(r);
             else
                 ap_reset_timeout(r);
 
             if (w <= 0) {
                 if (c != NULL) {
                     /*
                      * when a send failure occurs, we need to decide whether