1 files changed, 29 insertions, 0 deletions
diff --git a/web/server/h2o/libh2o/srcdoc/configure/reproxy_directives.mt b/web/server/h2o/libh2o/srcdoc/configure/reproxy_directives.mt
new file mode 100644
index 00000000..3288a65e
--- /dev/null
+++ b/web/server/h2o/libh2o/srcdoc/configure/reproxy_directives.mt
@@ -0,0 +1,29 @@
+? my $ctx = $main::context;
+? $_mt->wrapper_file("wrapper.mt", "Configure", "Reproxy Directives")->(sub {
+
+<p>
+This document describes the configuration directives of the reproxy handler.
+</p>
+
+<?
+$ctx->{directive}->(
+    name    => "reproxy",
+    levels  => [ qw(global host path extension) ],
+    default => q{reproxy: OFF},
+    desc    => <<'EOT',
+A boolean flag (<code>ON</code> or <code>OFF</code>) indicating if the server should recognize the <code>X-Reproxy-URL</code> header sent from <a href="configure/proxy_directives.html#proxy.reverse.url">upstream servers</a>.
+EOT
+)->(sub {
+?>
+<p>
+If H2O recognizes the header, it fetches the contents of the resource specified by the header, and sends the contents as the response to the client.
+If the status code associated with the <code>X-Reproxy-URL</code> header is 307 or 308, then the method of the original request is used to obtain the specified resource.
+Otherwise, the request method is changed to <code>GET</code>.
+</p>
+<p>
+For example, an upstream server may send an URL pointing to a large image using the <code>X-Reproxy-URL</code> header stored on a distributed file system, and let H2O fetch and return the content to the client, instead of fetching the image by itself.
+Doing so would reduce the load on the application server.
+</p>
+? })
+
+? })