From: Amaury Denoyelle <adenoyelle@haproxy.com>
Date: Tue, 6 Jul 2021 08:52:58 +0000 (+0200)
Subject: MINOR: http: use http uri parser for scheme
X-Git-Tag: v2.5-dev2~30
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=8ac8cbfd7219b5c8060ba6d7b5c76f0ec539e978;p=thirdparty%2Fhaproxy.git

MINOR: http: use http uri parser for scheme

Replace http_get_scheme by the http_uri_parser API. The new function is
renamed http_parse_scheme. A new http_uri_parser state is declared to
mark the scheme parsing as completed.
---

diff --git a/include/haproxy/http-t.h b/include/haproxy/http-t.h
index 140a9beedb..ddc0f2350e 100644
--- a/include/haproxy/http-t.h
+++ b/include/haproxy/http-t.h
@@ -134,6 +134,7 @@ enum http_etag_type {
 /* Indicates what elements have been parsed in a HTTP URI. */
 enum http_uri_parser_state {
 	URI_PARSER_STATE_BEFORE = 0,
+	URI_PARSER_STATE_SCHEME_DONE,
 };
 
 /* HTTP URI format as described in rfc 7230 5.3.
diff --git a/include/haproxy/http.h b/include/haproxy/http.h
index faf00a0345..60e101ebd6 100644
--- a/include/haproxy/http.h
+++ b/include/haproxy/http.h
@@ -36,7 +36,7 @@ extern const uint8_t http_char_classes[256];
 enum http_meth_t find_http_meth(const char *str, const int len);
 int http_get_status_idx(unsigned int status);
 const char *http_get_reason(unsigned int status);
-struct ist http_get_scheme(const struct ist uri);
+struct ist http_parse_scheme(struct http_uri_parser *parser);
 struct ist http_get_authority(const struct ist uri, int no_userinfo);
 struct ist http_get_path(const struct ist uri);
 int http_header_match2(const char *hdr, const char *end,
diff --git a/src/http.c b/src/http.c
index aff27f1c13..2899232149 100644
--- a/src/http.c
+++ b/src/http.c
@@ -470,25 +470,23 @@ const char *http_get_reason(unsigned int status)
 
 /* Parse the uri and looks for the scheme. If not found, an empty ist is
  * returned. Otherwise, the ist pointing to the scheme is returned.
+ *
+ * <parser> must have been initialized via http_uri_parser_init. See the
+ * related http_uri_parser documentation for the specific API usage.
  */
-struct ist http_get_scheme(const struct ist uri)
+struct ist http_parse_scheme(struct http_uri_parser *parser)
 {
 	const char *ptr, *start, *end;
 
-	if (!uri.len)
+	if (parser->state >= URI_PARSER_STATE_SCHEME_DONE)
 		goto not_found;
 
-	ptr = uri.ptr;
-	start = ptr;
-	end = ptr + uri.len;
-
-	/* RFC7230, par. 2.7 :
-	 * Request-URI = "*" | absuri | abspath | authority
-	 */
-
-	if (*ptr == '*' || *ptr == '/')
+	if (parser->format != URI_PARSER_FORMAT_ABSURI_OR_AUTHORITY)
 		goto not_found;
 
+	ptr = start = istptr(parser->uri);
+	end = istend(parser->uri);
+
 	if (isalpha((unsigned char)*ptr)) {
 		/* this is a scheme as described by RFC3986, par. 3.1, or only
 		 * an authority (in case of a CONNECT method).
@@ -512,9 +510,12 @@ struct ist http_get_scheme(const struct ist uri)
 		goto not_found;
 	}
 
+	parser->uri = ist2(ptr, end - ptr);
+	parser->state = URI_PARSER_STATE_SCHEME_DONE;
 	return ist2(start, ptr - start);
 
  not_found:
+	parser->state = URI_PARSER_STATE_SCHEME_DONE;
 	return IST_NULL;
 }
 
diff --git a/src/http_htx.c b/src/http_htx.c
index 399848db6e..5a62d0a21c 100644
--- a/src/http_htx.c
+++ b/src/http_htx.c
@@ -1740,6 +1740,7 @@ int http_scheme_based_normalize(struct htx *htx)
 	struct htx_sl *sl;
 	struct ist uri, scheme, authority, host, port;
 	char *start, *end, *ptr;
+	struct http_uri_parser parser;
 
 	sl = http_get_stline(htx);
 
@@ -1748,7 +1749,8 @@ int http_scheme_based_normalize(struct htx *htx)
 
 	uri = htx_sl_req_uri(sl);
 
-	scheme = http_get_scheme(uri);
+	parser = http_uri_parser_init(uri);
+	scheme = http_parse_scheme(&parser);
 	/* if no scheme found, no normalization to proceed */
 	if (!isttest(scheme))
 		return 0;