app/uri: add url encode and decode methods

Lua FFI bindings of curl_easy_escape() [1] and curl_easy_unescape() [2].

1. https://curl.se/libcurl/c/curl_easy_escape.html
2. https://curl.se/libcurl/c/curl_easy_unescape.html

@TarantoolBot document
Title: Document a new methods to encode and decode URL's

New methods "uri:encode()" and "uri:decode()" have been introduced.
First one allows to escape symbols in a string and second one to
unescape symbols in a string. Escaping and unescaping symbols
implemented using CURL functions curl_easy_escape() and
curl_easy_unescape() and conforms to RFC 3986. Maximum length of string
is limited with CURL_MAX_INPUT_LENGTH (8 MB).

```
tarantool> require('uri').encode('тарантул')
---
- '%D1%82%D0%B0%D1%80%D0%B0%D0%BD%D1%82%D1%83%D0%BB'
...

tarantool>
```

Fixes tarantool#3682

changelogs/unreleased/gh-3682-uri-encode-decode.md

@@ -0,0 +1,13 @@
+## feature/lua/uri
+
+* Added functions for encoding and decoding URL string according to
+  (gh-3682).
+
+  ```lua
+  tarantool> require('uri').encode('тарантул')
+  ---
+  - '%D1%82%D0%B0%D1%80%D0%B0%D0%BD%D1%82%D1%83%D0%BB'
+  ...
+
+  tarantool
+  ```

src/lua/uri.lua

@@ -4,6 +4,8 @@ local ffi = require('ffi')
 local buffer = require('buffer')
 local uri = require('uri')
 
+local url_decode_outlength = ffi.new("int[1]")
+
 ffi.cdef[[
 struct uri_param {
     const char *name;
@@ -45,6 +47,13 @@ uri_set_destroy(struct uri_set *uri_set);
 
 int
 uri_format(char *str, size_t len, struct uri *uri, bool write_password);
+
+typedef void CURL;
+CURL *curl_easy_init(void);
+void curl_easy_cleanup(CURL *handle);
+char *curl_easy_escape(CURL *handle, const char *string, int length);
+char *curl_easy_unescape(CURL *handle, const char *string, int length, int *outlength);
+void curl_free(void *p);
 ]]
 
 local builtin = ffi.C;
@@ -169,8 +178,66 @@ local function format(uri, write_password)
     return str
 end
 
+--- Encodes the given string
+-- Uses Curl function curl_easy_escape(),
+-- see description in https://curl.haxx.se/libcurl/c/curl_easy_escape.html
+-- @function url_encode
+-- @string buf - string to be encode
+-- @returns result string or nil, err
+local function url_encode(buf)
+    if type(buf) ~= 'string' then
+        error("Usage: uri.encode(string)")
+    end
+
+    local handle = ffi.C.curl_easy_init()
+    if not handle then
+        return nil, 'curl_easy_init error'
+    end
+
+    local escaped_str = ffi.C.curl_easy_escape(handle, buf, #buf)
+    ffi.C.curl_easy_cleanup(handle)
+    if escaped_str == nil then
+        return nil, 'curl_easy_escape error'
+    end
+
+    local res = ffi.string(escaped_str)
+    ffi.C.curl_free(escaped_str)
+
+    return res
+end
+
+--- Decodes the given string
+-- Uses Curl function curl_easy_unescape(),
+-- see description in https://curl.haxx.se/libcurl/c/curl_easy_unescape.html
+-- @function url_decode
+-- @string buf - encoded URL
+-- @returns string or nil, err
+local function url_decode(buf)
+    if type(buf) ~= 'string' then
+        error("Usage: uri.decode(string)" .. type(buf))
+    end
+
+    local handle = builtin.curl_easy_init()
+    if not handle then
+        return nil, 'curl_easy_init error'
+    end
+
+    local unescaped_str = builtin.curl_easy_unescape(handle, buf, #buf, url_decode_outlength)
+    builtin.curl_easy_cleanup(handle)
+    if unescaped_str == nil then
+        return nil, 'curl_easy_unescape error'
+    end
+
+    local res = ffi.string(unescaped_str, url_decode_outlength[0])
+    builtin.curl_free(unescaped_str)
+
+    return res
+end
+
 return {
     parse_many = parse_many,
     parse = parse,
     format = format,
+    encode = url_encode,
+    decode = url_decode,
 };

test/app-tap/uri.test.lua

@@ -743,12 +743,95 @@ local function test_parse_invalid_uri_set_from_lua_table(test)
     test:is(tostring(error), expected_errmsg, "error message")
 end
 
+local url_encode_test = {
+    { "a", "a" },
+    { "/", "%2F" },
+    { "a=b", "a%3Db" },
+    { "1/./0", "1%2F.%2F0" },
+    { "-._~!#%&", "-._~%21%23%25%26" },
+    { "тарантул", "%D1%82%D0%B0%D1%80%D0%B0%D0%BD%D1%82%D1%83%D0%BB" },
+}
+
+local url_decode_test = {
+    { "a", "a" },
+    { "a=b", "a%3Db" },
+    { "a=", "a%3D" },
+    { "1/./0", "1%2F.%2F0" },
+    { "-._~!#%&", "-._~%21%23%25%26" },
+    { "a", "%61" },
+    { "aa", "%61a" },
+    { "ab", "%61b" },
+    { "%6 1", "%6 1" },
+    { "%6%a", "%6%a" },
+    { "j", "%6a" },
+    { "%-2", "%-2" },
+    { "%FG", "%FG" },
+    { "/", "%2F" },
+    { "a", "%61" },
+    { "aa", "%61a" },
+    { "ab", "%61b" },
+    { "%6 1", "%6 1" },
+    { "%6%a", "%6%a" },
+    { "j", "%6a" },
+    { "%-2", "%-2" },
+    { "%FG", "%FG" },
+}
+
+local function test_url_encode(test)
+    local len = table.getn(url_encode_test)
+    test:plan(len)
+    for _, testcase in pairs(url_encode_test) do
+        local raw = testcase[1]
+        local encoded = testcase[2]
+        local desc = ('encoding of "%s"'):format(raw)
+        test:is(encoded, uri.encode(raw), desc)
+    end
+end
+
+local function test_url_decode(test)
+    local len = table.getn(url_decode_test)
+    test:plan(len)
+    for _, testcase in pairs(url_decode_test) do
+        local raw = testcase[1]
+        local encoded = testcase[2]
+        local desc = ('decoding of "%s"'):format(encoded)
+        test:is(raw, uri.decode(encoded), desc)
+    end
+end
+
+local function test_url_encode_decode_roundtrip(test)
+    test:plan(14)
+
+    test:is('hello', uri.encode(uri.decode('hello')), 'decoding and encoding of "hello"')
+    test:is('привет', uri.decode(uri.encode('привет')), 'encoding and decoding of "привет"')
+
+    test:is('%00', uri.encode(uri.decode('%00')), 'decoding and encoding of %00')
+    test:is('\0', uri.decode(uri.encode('\0')), 'encoding and decoding of \\0')
+
+    test:is('hello', uri.encode('hello'), 'correct encoding of eng')
+    test:is('%D0%BF%D1%80%D0%B8%D0%B2%D0%B5%D1%82', uri.encode('привет'), 'correct encoding of rus')
+    test:is('привет', uri.decode(uri.encode('привет')), 'encoding and decoding of "привет"')
+    test:is('\0', uri.decode(uri.encode('\0')), 'encoding and decoding of \\0')
+
+    test:is('hello', uri.decode('hello'), 'correct decoding of eng')
+    test:is('привет', uri.decode('%D0%BF%D1%80%D0%B8%D0%B2%D0%B5%D1%82'), 'correct decoding of rus')
+
+    test:is('hello', uri.encode(uri.decode('hello')), 'decoding and encoding of "hello"')
+    test:is('привет', uri.decode(uri.encode('привет')), 'encoding and decoding of "привет"')
+
+    test:is('%00', uri.encode(uri.decode('%00')), 'decoding and encoding of %00')
+    test:is('\0', uri.decode(uri.encode('\0')), 'encoding and decoding of \\0')
+end
+
 tap.test("uri", function(test)
-    test:plan(6)
+    test:plan(9)
     test:test("parse", test_parse)
     test:test("parse URI query params", test_parse_uri_query_params)
     test:test("parse URIs with query params", test_parse_uri_set_with_query_params)
     test:test("parse URIs from lua table", test_parse_uri_set_from_lua_table)
     test:test("parse invalid URIs from lua table", test_parse_invalid_uri_set_from_lua_table)
     test:test("format", test_format)
+    test:test("encode uri", test_url_encode)
+    test:test("decode uri", test_url_decode)
+    test:test("encode and decode uri (roundtrip)", test_url_encode_decode_roundtrip)
 end)