archive-gh-me/jerryscript
1
0
Fork 0
mirror of https://github.com/jerryscript-project/jerryscript.git synced 2025-12-15 16:29:21 +00:00
Code Issues Projects Releases Wiki Activity

Support all literals in 'jerry_get_literals_from_snapshot' when C format is used. (#3472)

Browse Source 
The literals must be saved in hex format to support cases with special
characters, like `var s = 'hello",\n"world';`

JerryScript-DCO-1.0-Signed-off-by: László Langó lango@inf.u-szeged.hu
This commit is contained in:
László Langó 2020-01-10 14:57:10 +01:00 committed by Robert Fancsik
parent 08da8bc7aa
commit 2960dffb50
2 changed files with 41 additions and 94 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

106
jerry-core/api/jerry-snapshot.c
View File

@ -1551,70 +1551,6 @@ jerry_append_number_to_buffer (uint8_t *buffer_p, /**< buffer */
utf8_str_size);
} /* jerry_append_number_to_buffer */
/**
* Check whether the passed ecma-string is a valid identifier.
*
* @return true - if the ecma-string is a valid identifier,
* false - otherwise
*/
static bool
ecma_string_is_valid_identifier (const ecma_string_t *string_p)
{
ECMA_STRING_TO_UTF8_STRING (string_p, str_buffer_p, str_buffer_size);
const uint8_t *str_p = str_buffer_p;
const uint8_t *str_end_p = str_buffer_p + str_buffer_size;
while (str_p < str_end_p)
{
lit_code_point_t code_point = *str_p;
lit_utf8_size_t utf8_length = 1;
if (JERRY_UNLIKELY (code_point >= LIT_UTF8_2_BYTE_MARKER))
{
utf8_length = lit_read_code_point_from_utf8 (str_p,
(lit_utf8_size_t) (str_end_p - str_p),
&code_point);
#if ENABLED (JERRY_ES2015)
if ((code_point >= LIT_UTF16_HIGH_SURROGATE_MIN && code_point <= LIT_UTF16_HIGH_SURROGATE_MAX)
&& str_p + 3 < str_end_p)
{
lit_code_point_t low_surrogate;
lit_read_code_point_from_utf8 (str_p + 3,
(lit_utf8_size_t) (str_end_p - (str_p + 3)),
&low_surrogate);
if (low_surrogate >= LIT_UTF16_LOW_SURROGATE_MIN && low_surrogate <= LIT_UTF16_LOW_SURROGATE_MAX)
{
code_point = lit_convert_surrogate_pair_to_code_point ((ecma_char_t) code_point,
(ecma_char_t) low_surrogate);
utf8_length = 2 * 3;
}
}
#endif /* ENABLED (JERRY_ES2015) */
}
if (str_p == str_buffer_p)
{
if (!lit_code_point_is_identifier_start (code_point))
{
break;
}
}
else if (!lit_code_point_is_identifier_part (code_point))
{
break;
}
str_p += utf8_length;
}
ECMA_FINALIZE_UTF8_STRING (str_buffer_p, str_buffer_size);
return str_p == str_end_p;
} /* ecma_string_is_valid_identifier */
#endif /* ENABLED (JERRY_SNAPSHOT_SAVE) */
/**
@ -1667,14 +1603,7 @@ jerry_get_literals_from_snapshot (const uint32_t *snapshot_p, /**< input snapsho
{
ecma_string_t *literal_p = ecma_get_string_from_value (buffer_p[i]);
/* NOTE:
* We don't save a literal (in C format) which isn't a valid
* identifier or it's a magic string.
* TODO:
* Save all of the literals in C format as well.
*/
if (ecma_get_string_magic (literal_p) == LIT_MAGIC_STRING__COUNT
&& (!is_c_format || ecma_string_is_valid_identifier (literal_p)))
if (ecma_get_string_magic (literal_p) == LIT_MAGIC_STRING__COUNT)
{
literal_count++;
}
@ -1702,14 +1631,7 @@ jerry_get_literals_from_snapshot (const uint32_t *snapshot_p, /**< input snapsho
{
ecma_string_t *literal_p = ecma_get_string_from_value (buffer_p[i]);
/* NOTE:
* We don't save a literal (in C format) which isn't a valid
* identifier or it's a magic string.
* TODO:
* Save all of the literals in C format as well.
*/
if (ecma_get_string_magic (literal_p) == LIT_MAGIC_STRING__COUNT
&& (!is_c_format || ecma_string_is_valid_identifier (literal_p)))
if (ecma_get_string_magic (literal_p) == LIT_MAGIC_STRING__COUNT)
{
literal_array[literal_idx++] = literal_p;
}
@ -1743,7 +1665,29 @@ jerry_get_literals_from_snapshot (const uint32_t *snapshot_p, /**< input snapsho
for (lit_utf8_size_t i = 0; i < literal_count; i++)
{
lit_buf_p = jerry_append_chars_to_buffer (lit_buf_p, buffer_end_p, " \"", 0);
lit_buf_p = jerry_append_ecma_string_to_buffer (lit_buf_p, buffer_end_p, literal_array[i]);
ECMA_STRING_TO_UTF8_STRING (literal_array[i], str_buffer_p, str_buffer_size);
for (lit_utf8_size_t j = 0; j < str_buffer_size; j++)
{
uint8_t byte = str_buffer_p[j];
if (byte < 32 || byte > 127)
{
lit_buf_p = jerry_append_chars_to_buffer (lit_buf_p, buffer_end_p, "\\x", 0);
ecma_char_t hex_digit = (ecma_char_t) (byte >> 4);
*lit_buf_p++ = (lit_utf8_byte_t) ((hex_digit > 9) ? (hex_digit + ('A' - 10)) : (hex_digit + '0'));
hex_digit = (lit_utf8_byte_t) (byte & 0xf);
*lit_buf_p++ = (lit_utf8_byte_t) ((hex_digit > 9) ? (hex_digit + ('A' - 10)) : (hex_digit + '0'));
}
else
{
if (byte == '\\' || byte == '"')
{
*lit_buf_p++ = '\\';
}
*lit_buf_p++ = byte;
}
}
ECMA_FINALIZE_UTF8_STRING (str_buffer_p, str_buffer_size);
lit_buf_p = jerry_append_chars_to_buffer (lit_buf_p, buffer_end_p, "\"", 0);
if (i < literal_count - 1)

29
tests/unit-core/test-snapshot.c
View File

@ -400,7 +400,7 @@ main (void)
static jerry_char_t literal_buffer_c[LITERAL_BUFFER_SIZE];
static uint32_t literal_snapshot_buffer[SNAPSHOT_BUFFER_SIZE];
static const jerry_char_t code_for_c_format[] = "var object = { aa:'fo o', Bb:'max', aaa:'xzy0' };";
static const jerry_char_t code_for_c_format[] = "var object = { aa:'fo\" o\\n \\\\', Bb:'max', aaa:'xzy0' };";
jerry_value_t generate_result;
generate_result = jerry_generate_snapshot (NULL,
@ -411,35 +411,37 @@ main (void)
literal_snapshot_buffer,
SNAPSHOT_BUFFER_SIZE);
TEST_ASSERT (!jerry_value_is_error (generate_result)
&& jerry_value_is_number (generate_result));
TEST_ASSERT (!jerry_value_is_error (generate_result));
TEST_ASSERT (jerry_value_is_number (generate_result));
size_t snapshot_size = (size_t) jerry_get_number_value (generate_result);
jerry_release_value (generate_result);
TEST_ASSERT (snapshot_size == 120);
TEST_ASSERT (snapshot_size == 124);
const size_t lit_c_buf_sz = jerry_get_literals_from_snapshot (literal_snapshot_buffer,
snapshot_size,
literal_buffer_c,
LITERAL_BUFFER_SIZE,
true);
TEST_ASSERT (lit_c_buf_sz == 200);
TEST_ASSERT (lit_c_buf_sz == 239);
static const char *expected_c_format = (
"jerry_length_t literal_count = 4;\n\n"
"jerry_char_t *literals[4] =\n"
"jerry_length_t literal_count = 5;\n\n"
"jerry_char_t *literals[5] =\n"
"{\n"
" \"Bb\",\n"
" \"aa\",\n"
" \"aaa\",\n"
" \"xzy0\"\n"
" \"xzy0\",\n"
" \"fo\\\" o\\x0A \\\\\"\n"
"};\n\n"
"jerry_length_t literal_sizes[4] =\n"
"jerry_length_t literal_sizes[5] =\n"
"{\n"
" 2 /* Bb */,\n"
" 2 /* aa */,\n"
" 3 /* aaa */,\n"
" 4 /* xzy0 */\n"
" 4 /* xzy0 */,\n"
" 8 /* fo\" o\n \\ */\n"
"};\n"
);
@ -452,9 +454,10 @@ main (void)
literal_buffer_list,
LITERAL_BUFFER_SIZE,
false);
TEST_ASSERT (lit_list_buf_sz == 30);
TEST_ASSERT (!strncmp ((char *) literal_buffer_list, "2 Bb\n2 aa\n3 aaa\n4 fo o\n4 xzy0\n", lit_list_buf_sz));
TEST_ASSERT (lit_list_buf_sz == 34);
TEST_ASSERT (!strncmp ((char *) literal_buffer_list,
"2 Bb\n2 aa\n3 aaa\n4 xzy0\n8 fo\" o\n \\\n",
lit_list_buf_sz));
jerry_cleanup ();
}