MINOR: uri_normalizer: Add a `merge-slashes` normalizer to http-request normalize-uri

This normalizer merges adjacent slashes into a single slash, thus removing
empty path segments.

See GitHub Issue #714.

doc/configuration.txt

@@ -6011,6 +6011,19 @@ http-request early-hint <name> <fmt> [ { if | unless } <condition> ]
 
   See RFC 8297 for more information.
 
+http-request normalize-uri <normalizer> [ { if | unless } <condition> ]
+http-request normalize-uri merge-slashes [ { if | unless } <condition> ]
+
+  Performs normalization of the request's URI. The following normalizers are
+  available:
+
+  - merge-slashes: Merges adjacent slashes within the "path" component into a
+      single slash.
+
+      Example:
+      - //        -> /
+      - /foo//bar -> /foo/bar
+
 http-request redirect <rule> [ { if | unless } <condition> ]
 
   This performs an HTTP redirection based on a redirect rule. This is exactly

include/haproxy/action-t.h

@@ -102,7 +102,7 @@ enum act_timeout_name {
 };
 
 enum act_normalize_uri {
-	ACT_NORMALIZE_URI_PLACEHOLDER,
+	ACT_NORMALIZE_URI_MERGE_SLASHES,
 };
 
 /* NOTE: if <.action_ptr> is defined, the referenced function will always be

include/haproxy/uri_normalizer.h

@@ -14,8 +14,12 @@
 #ifndef _HAPROXY_URI_NORMALIZER_H
 #define _HAPROXY_URI_NORMALIZER_H
 
+#include <import/ist.h>
+
 #include <haproxy/uri_normalizer-t.h>
 
+enum uri_normalizer_err uri_normalizer_path_merge_slashes(const struct ist path, struct ist *dst);
+
 #endif /* _HAPROXY_URI_NORMALIZER_H */
 
 /*

reg-tests/http-rules/normalize_uri.vtc

@@ -0,0 +1,87 @@
+varnishtest "normalize-uri tests"
+#REQUIRE_VERSION=2.4
+
+# This reg-test tests the http-request normalize-uri action.
+
+feature ignore_unknown_macro
+
+server s1 {
+    rxreq
+    txresp
+} -repeat 10 -start
+
+haproxy h1 -conf {
+    defaults
+        mode http
+        timeout connect 1s
+        timeout client  1s
+        timeout server  1s
+
+    frontend fe_merge_slashes
+        bind "fd@${fe_merge_slashes}"
+
+        http-request set-var(txn.before) url
+        http-request normalize-uri merge-slashes
+        http-request set-var(txn.after) url
+
+        http-response add-header before  %[var(txn.before)]
+        http-response add-header after  %[var(txn.after)]
+
+        default_backend be
+
+    backend be
+        server s1 ${s1_addr}:${s1_port}
+
+} -start
+
+client c1 -connect ${h1_fe_merge_slashes_sock} {
+    txreq -url "/foo/bar"
+    rxresp
+    expect resp.http.before == "/foo/bar"
+    expect resp.http.after == "/foo/bar"
+
+    txreq -url "/foo//bar"
+    rxresp
+    expect resp.http.before == "/foo//bar"
+    expect resp.http.after == "/foo/bar"
+
+    txreq -url "/foo///bar"
+    rxresp
+    expect resp.http.before == "/foo///bar"
+    expect resp.http.after == "/foo/bar"
+
+    txreq -url "///foo///bar"
+    rxresp
+    expect resp.http.before == "///foo///bar"
+    expect resp.http.after == "/foo/bar"
+
+    txreq -url "///foo/bar"
+    rxresp
+    expect resp.http.before == "///foo/bar"
+    expect resp.http.after == "/foo/bar"
+
+    txreq -url "///foo///bar///"
+    rxresp
+    expect resp.http.before == "///foo///bar///"
+    expect resp.http.after == "/foo/bar/"
+
+    txreq -url "///"
+    rxresp
+    expect resp.http.before == "///"
+    expect resp.http.after == "/"
+
+    txreq -url "/foo?bar=///"
+    rxresp
+    expect resp.http.before == "/foo?bar=///"
+    expect resp.http.after == "/foo?bar=///"
+
+    txreq -url "//foo?bar=///"
+    rxresp
+    expect resp.http.before == "//foo?bar=///"
+    expect resp.http.after == "/foo?bar=///"
+
+    txreq -req OPTIONS -url "*"
+    rxresp
+    expect resp.http.before == "*"
+    expect resp.http.after == "*"
+} -run

src/http_act.c

@@ -215,8 +215,23 @@ static enum act_return http_action_normalize_uri(struct act_rule *rule, struct p
 		goto fail_alloc;
 
 	switch ((enum act_normalize_uri) rule->action) {
-		case ACT_NORMALIZE_URI_PLACEHOLDER:
-			(void) uri;
+		case ACT_NORMALIZE_URI_MERGE_SLASHES: {
+			const struct ist path = http_get_path(uri);
+			struct ist newpath = ist2(replace->area, replace->size);
+
+			if (!isttest(path))
+				goto leave;
+
+			err = uri_normalizer_path_merge_slashes(iststop(path, '?'), &newpath);
+
+			if (err != URI_NORMALIZER_ERR_NONE)
+				break;
+
+			if (!http_replace_req_path(htx, newpath, 0))
+				goto fail_rewrite;
+
+			break;
+		}
 	}
 
 	switch (err) {
@@ -277,8 +292,10 @@ static enum act_parse_ret parse_http_normalize_uri(const char **args, int *orig_
 		return ACT_RET_PRS_ERR;
 	}
 
-	if (0) {
+	if (strcmp(args[cur_arg], "merge-slashes") == 0) {
+		cur_arg++;
 
+		rule->action = ACT_NORMALIZE_URI_MERGE_SLASHES;
 	}
 	else {
 		memprintf(err, "unknown normalizer '%s'", args[cur_arg]);

src/uri_normalizer.c

@@ -10,9 +10,48 @@
  *
  */
 
+#include <import/ist.h>
+
 #include <haproxy/api.h>
 #include <haproxy/uri_normalizer.h>
 
+/* Merges adjacent slashes in the given path. */
+enum uri_normalizer_err uri_normalizer_path_merge_slashes(const struct ist path, struct ist *dst)
+{
+	enum uri_normalizer_err err;
+
+	const size_t size = istclear(dst);
+	struct ist newpath = *dst;
+
+	struct ist scanner = path;
+
+	/* The path will either be shortened or have the same length. */
+	if (size < istlen(path)) {
+		err = URI_NORMALIZER_ERR_ALLOC;
+		goto fail;
+	}
+
+	while (istlen(scanner) > 0) {
+		const char current = istshift(&scanner);
+
+		if (current == '/') {
+			while (istlen(scanner) > 0 && *istptr(scanner) == '/')
+				scanner = istnext(scanner);
+		}
+
+		newpath = __istappend(newpath, current);
+	}
+
+	*dst = newpath;
+
+	return URI_NORMALIZER_ERR_NONE;
+
+  fail:
+
+	return err;
+}
+
+
 /*
  * Local variables:
  *  c-indent-level: 8