BUG/MINOR: http-htx: Support default path during scheme based normalization

As stated in RFC3986, for an absolute-form URI, an empty path should be
normalized to a path of "/". This is part of scheme based normalization
rules. This kind of normalization is already performed for default ports. So
we might as well deal with the case of empty path.

The associated reg-tests was updated accordingly.

This patch should fix the issue #2573. It may be backported as far as 2.4 if
necessary.

diff --git a/reg-tests/http-messaging/h1_host_normalization.vtc b/reg-tests/http-messaging/h1_host_normalization.vtc
index 48174b819c805a24c66b34b5046708c3cbf1b50f..909d5250942c7e3ae70b6a579bf6f02457584f3b 100644 (file)
--- a/reg-tests/http-messaging/h1_host_normalization.vtc
+++ b/reg-tests/http-messaging/h1_host_normalization.vtc
@@ -175,22 +175,62 @@ syslog S1 -level info {
 
     # C32
     recv
-    expect ~ "^.* uri: GET http:// HTTP/1.1; host: {}$"
+    expect ~ "^.* uri: GET http:/// HTTP/1.1; host: {}$"
     barrier b1 sync
 
     # C33
     recv
-    expect ~ "^.* uri: GET https:// HTTP/1.1; host: {}$"
+    expect ~ "^.* uri: GET https:/// HTTP/1.1; host: {}$"
     barrier b1 sync
 
     # C34
     recv
-    expect ~ "^.* uri: GET http:// HTTP/1.1; host: {}$"
+    expect ~ "^.* uri: GET http:/// HTTP/1.1; host: {}$"
     barrier b1 sync
 
     # C35
     recv
-    expect ~ "^.* uri: GET https:// HTTP/1.1; host: {}$"
+    expect ~ "^.* uri: GET https:/// HTTP/1.1; host: {}$"
+    barrier b1 sync
+
+    # C36
+    recv
+    expect ~ "^.* uri: GET http://hostname/ HTTP/1.1; host: {hostname}$"
+    barrier b1 sync
+
+    # C37
+    recv
+    expect ~ "^.* uri: GET http://hostname/ HTTP/1.1; host: {hostname}$"
+    barrier b1 sync
+
+    # C38
+    recv
+    expect ~ "^.* uri: GET http://hostname/ HTTP/1.1; host: {hostname}$"
+    barrier b1 sync
+
+    # C39
+    recv
+    expect ~ "^.* uri: GET https://hostname/ HTTP/1.1; host: {hostname}$"
+    barrier b1 sync
+
+    # C40
+    recv
+    expect ~ "^.* uri: GET https://hostname/ HTTP/1.1; host: {hostname}$"
+    barrier b1 sync
+
+    # C41
+    recv
+    expect ~ "^.* uri: GET https://hostname/ HTTP/1.1; host: {hostname}$"
+    barrier b1 sync
+
+    # C42
+    recv
+    expect ~ "^.* uri: GET http://hostname:81/ HTTP/1.1; host: {hostname:81}$"
+    barrier b1 sync
+
+    # C43
+    recv
+    expect ~ "^.* uri: GET https://hostname:444/ HTTP/1.1; host: {hostname:444}$"
 
 } -start
 
@@ -759,4 +799,108 @@ client c35 -connect ${h1_fe_sock} {
     expect resp.status == 200
 } -run
 
+# Wait matching on log message
+barrier b1 sync
+
+client c36 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "http://hostname" \
+      -hdr "host: hostname"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c37 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "http://hostname:80" \
+      -hdr "host: hostname"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c38 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "http://hostname:" \
+      -hdr "host: hostname"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c39 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "https://hostname" \
+      -hdr "host: hostname"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c40 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "https://hostname:443" \
+      -hdr "host: hostname"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c41 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "https://hostname:" \
+      -hdr "host: hostname"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c42 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "http://hostname:81" \
+      -hdr "host: hostname:81"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
+# Wait matching on log message
+barrier b1 sync
+
+client c43 -connect ${h1_fe_sock} {
+    txreq \
+      -req "GET" \
+      -url "https://hostname:444" \
+      -hdr "host: hostname:444"
+
+    rxresp
+    expect resp.status == 200
+} -run
+
 syslog S1 -wait

diff --git a/src/http_htx.c b/src/http_htx.c
index 405b688b6b25543cd505249623b7e0242895c8a8..36356edf021f9f35e3d4a2d16c3954afb41356e5 100644 (file)
--- a/src/http_htx.c
+++ b/src/http_htx.c
@@ -1828,8 +1828,9 @@ int http_scheme_based_normalize(struct htx *htx)
 {
        struct http_hdr_ctx ctx;
        struct htx_sl *sl;
-       struct ist uri, scheme, authority, host, port;
+       struct ist uri, scheme, authority, host, port, path;
        struct http_uri_parser parser;
+       int normalize = 0;
 
        sl = http_get_stline(htx);
 
@@ -1846,14 +1847,21 @@ int http_scheme_based_normalize(struct htx *htx)
 
        /* Extract the port if present in authority */
        authority = http_parse_authority(&parser, 1);
+       path = http_parse_path(&parser);
        port = http_get_host_port(authority);
-       if (!isttest(port)) {
-               /* if no port found, no normalization to proceed */
-               return 0;
+       if (!isttest(port) || !http_is_default_port(scheme, port))
+               host = authority;
+       else {
+               host = isttrim(authority, istlen(authority) - istlen(port) - 1);
+               normalize = 1;
+       }
+
+       if (!isttest(path)) {
+               path = ist("/");
+               normalize = 1;
        }
-       host = isttrim(authority, istlen(authority) - istlen(port) - 1);
 
-       if (http_is_default_port(scheme, port)) {
+       if (normalize) {
                /* reconstruct the uri with removal of the port */
                struct buffer *temp = get_trash_chunk();
                struct ist meth, vsn;
@@ -1869,8 +1877,8 @@ int http_scheme_based_normalize(struct htx *htx)
                /* reconstruct uri without port */
                chunk_memcat(temp, uri.ptr, authority.ptr - uri.ptr);
                chunk_istcat(temp, host);
-               chunk_memcat(temp, istend(authority), istend(uri) - istend(authority));
-               uri = ist2(temp->area + meth.len + vsn.len, host.len + uri.len - authority.len); /* uri */
+               chunk_istcat(temp, path);
+               uri = ist2(temp->area + meth.len + vsn.len, host.len + path.len + authority.ptr - uri.ptr); /* uri */
 
                http_replace_stline(htx, meth, uri, vsn);