From 52d1dfb9acf71bd962cb3fe836818efb09dc1cc3 Mon Sep 17 00:00:00 2001
From: Vsevolod Stakhov <vsevolod@rspamd.com>
Date: Mon, 12 Feb 2024 14:59:52 +0000
Subject: [PATCH] [Feature] Lua_url: Add `to_http` method

---
 src/lua/lua_url.c | 87 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 86 insertions(+), 1 deletion(-)

diff --git a/src/lua/lua_url.c b/src/lua/lua_url.c
index 913469f6d..20f0a7121 100644
--- a/src/lua/lua_url.c
+++ b/src/lua/lua_url.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2023 Vsevolod Stakhov
+ * Copyright 2024 Vsevolod Stakhov
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -74,6 +74,7 @@ LUA_FUNCTION_DEF(url, lt);
 LUA_FUNCTION_DEF(url, eq);
 LUA_FUNCTION_DEF(url, get_order);
 LUA_FUNCTION_DEF(url, get_part_order);
+LUA_FUNCTION_DEF(url, to_http);
 
 static const struct luaL_reg urllib_m[] = {
 	LUA_INTERFACE_DEF(url, get_length),
@@ -101,6 +102,7 @@ static const struct luaL_reg urllib_m[] = {
 	LUA_INTERFACE_DEF(url, get_flags_num),
 	LUA_INTERFACE_DEF(url, get_order),
 	LUA_INTERFACE_DEF(url, get_part_order),
+	LUA_INTERFACE_DEF(url, to_http),
 	{"get_redirected", lua_url_get_phished},
 	LUA_INTERFACE_DEF(url, set_redirected),
 	{"__tostring", lua_url_tostring},
@@ -343,6 +345,89 @@ lua_url_tostring(lua_State *L)
 	return 1;
 }
 
+/***
+ * @method url:to_http()
+ * Get URL suitable for HTTP request (e.g. by trimming fragment and user parts)
+ * @return {string} url as a string
+ */
+static gint
+lua_url_to_http(lua_State *L)
+{
+	LUA_TRACE_POINT;
+	struct rspamd_lua_url *url = lua_check_url(L, 1);
+
+	if (url != NULL && url->url != NULL) {
+		if (url->url->protocol == PROTOCOL_MAILTO) {
+			/* Nothing to do here */
+			lua_pushnil(L);
+		}
+		else {
+
+			if (url->url->userlen > 0) {
+				/* We need to reconstruct url :( */
+				gsize len = url->url->urllen - url->url->fragmentlen + 1;
+
+				/* Strip the # character */
+				if (url->url->fragmentlen > 0 && len > 0) {
+					while (url->url->string[len - 1] == '#' && len > 0) {
+						len--;
+					}
+				}
+				gchar *nstr = g_malloc(len);
+				gchar *d = nstr, *end = nstr + len;
+				memcpy(nstr, url->url->string, url->url->protocollen);
+				d += url->url->protocollen;
+				*d++ = ':';
+				*d++ = '/';
+				*d++ = '/';
+
+				/* Host part */
+				memcpy(d, rspamd_url_host(url->url), url->url->hostlen);
+				d += url->url->hostlen;
+
+				int port = rspamd_url_get_port_if_special(url->url);
+
+				if (port > 0) {
+					d += rspamd_snprintf(d, end - d, ":%d/", port);
+				}
+				else {
+					*d++ = '/';
+				}
+
+				if (url->url->datalen > 0) {
+					memcpy(d, rspamd_url_data_unsafe(url->url), url->url->datalen);
+					d += url->url->datalen;
+				}
+
+				if (url->url->querylen > 0) {
+					*d++ = '?';
+					memcpy(d, rspamd_url_query_unsafe(url->url), url->url->querylen);
+					d += url->url->querylen;
+				}
+
+				g_assert(d < end);
+				lua_pushlstring(L, nstr, d - nstr);
+			}
+			else {
+				gsize len = url->url->urllen - url->url->fragmentlen;
+
+				/* Strip the # character */
+				if (url->url->fragmentlen > 0 && len > 0) {
+					while (url->url->string[len - 1] == '#' && len > 0) {
+						len--;
+					}
+				}
+				lua_pushlstring(L, url->url->string, len);
+			}
+		}
+	}
+	else {
+		lua_pushnil(L);
+	}
+
+	return 1;
+}
+
 /***
  * @method url:get_raw()
  * Get full content of the url as it was parsed (e.g. with urldecode)
-- 
2.39.5