Skip to content

Commit

Permalink
Improve toLowerCase and toUpperCase functions. (jerryscript-proje…
Browse files Browse the repository at this point in the history 
…ct#1575)

Language-sensitive mappings are not processed now.

Fixes jerryscript-project#323

JerryScript-DCO-1.0-Signed-off-by: Robert Sipka [email protected]
  • Loading branch information
@robertsipka @LaszloLango
robertsipka authored and LaszloLango committed Feb 15, 2017
1 parent 025a99c commit 1b5f839
Show file tree
Hide file tree
Showing 5 changed files with 1,163 additions and 5 deletions.
257 changes: 253 additions & 4 deletions jerry-core/lit/lit-char-helpers.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,10 @@
#include "lit/lit-unicode-ranges.inc.h"
#include "lit-strings.h"

#ifndef CONFIG_DISABLE_UNICODE_CASE_CONVERSION
#include "lit-unicode-conversions.inc.h"
#endif /* !CONFIG_DISABLE_UNICODE_CASE_CONVERSION */

#define NUM_OF_ELEMENTS(array) (sizeof (array) / sizeof ((array)[0]))

/**
Expand Down Expand Up @@ -458,6 +462,184 @@ lit_char_is_word_char (ecma_char_t c) /**< code unit */
|| c == LIT_CHAR_UNDERSCORE);
} /* lit_char_is_word_char */

#ifndef CONFIG_DISABLE_UNICODE_CASE_CONVERSION

/**
* Check if the specified character is in one of those tables which contain bidirectional conversions.
*
* @return the mapped character sequence of an ecma character, if it's in the table.
* 0 - otherwise.
*/
static ecma_length_t
search_in_bidirectional_conversion_tables (ecma_char_t character, /**< code unit */
ecma_char_t *output_buffer_p, /**< [out] buffer for the result characters */
bool is_lowercase) /**< is lowercase conversion */
{
/* 1, Check if the specified character is part of the jerry_character_case_ranges table. */
int number_of_case_ranges = NUM_OF_ELEMENTS (jerry_character_case_ranges);
int conv_counter = 0;

for (int i = 0; i < number_of_case_ranges; i++)
{
if (i % 2 == 0 && i > 0)
{
conv_counter++;
}

int range_length = jerry_character_case_range_lengths[conv_counter];
ecma_char_t start_point = jerry_character_case_ranges[i];

if (start_point > character || character >= start_point + range_length)
{
continue;
}

int char_dist = character - start_point;

if (i % 2 == 0)
{
output_buffer_p[0] = is_lowercase ? (ecma_char_t) (jerry_character_case_ranges[i + 1] + char_dist) : character;
}
else
{
output_buffer_p[0] = is_lowercase ? character : (ecma_char_t) (jerry_character_case_ranges[i - 1] + char_dist);
}

return 1;
}

/* 2, Check if the specified character is part of the character_pair_ranges table. */
int bottom = 0;
int top = NUM_OF_ELEMENTS (jerry_character_pair_ranges) - 1;

while (bottom <= top)
{
int middle = (bottom + top) / 2;
ecma_char_t current_sp = jerry_character_pair_ranges[middle];

if (current_sp <= character && character < current_sp + jerry_character_pair_range_lengths[middle])
{
int char_dist = character - current_sp;

if ((character - current_sp) % 2 == 0)
{
output_buffer_p[0] = is_lowercase ? (ecma_char_t) (current_sp + char_dist + 1) : character;
}
else
{
output_buffer_p[0] = is_lowercase ? character : (ecma_char_t) (current_sp + char_dist - 1);
}

return 1;
}

if (character > current_sp)
{
bottom = middle + 1;
}
else
{
top = middle - 1;
}
}

/* 3, Check if the specified character is part of the character_pairs table. */
int number_of_character_pairs = NUM_OF_ELEMENTS (jerry_character_pairs);

for (int i = 0; i < number_of_character_pairs; i++)
{
if (character != jerry_character_pairs[i])
{
continue;
}

if (i % 2 == 0)
{
output_buffer_p[0] = is_lowercase ? jerry_character_pairs[i + 1] : character;
}
else
{
output_buffer_p[0] = is_lowercase ? character : jerry_character_pairs[i - 1];
}

return 1;
}

return 0;
} /* search_in_bidirectional_conversion_tables */

/**
* Check if the specified character is in the given conversion table.
*
* @return the mapped character sequence of an ecma character, if it's in the table.
* 0 - otherwise.
*/
static ecma_length_t
search_in_conversion_table (ecma_char_t character, /**< code unit */
ecma_char_t *output_buffer_p, /**< [out] buffer for the result characters */
const ecma_char_t *array, /**< array */
const uint8_t *counters) /**< case_values counter */
{
int end_point = 0;

for (int i = 0; i < 3; i++)
{
int start_point = end_point;
int size_of_case_value = i + 1;
end_point += counters[i] * (size_of_case_value + 1);

int bottom = start_point;
int top = end_point - size_of_case_value;

while (bottom <= top)
{
int middle = (bottom + top) / 2;

middle -= ((middle - bottom) % (size_of_case_value + 1));

ecma_char_t current = array[middle];

if (current == character)
{
ecma_length_t char_sequence = 1;

switch (size_of_case_value)
{
case 3:
{
output_buffer_p[2] = array[middle + 3];
char_sequence++;
/* FALLTHRU */
}
case 2:
{
output_buffer_p[1] = array[middle + 2];
char_sequence++;
/* FALLTHRU */
}
default:
{
output_buffer_p[0] = array[middle + 1];
return char_sequence;
}
}
}

if (character < current)
{
top = middle - (size_of_case_value + 1);
}
else
{
bottom = middle + (size_of_case_value + 1);
}
}
}

return 0;
} /* search_in_conversion_table */
#endif /* !CONFIG_DISABLE_UNICODE_CASE_CONVERSION */

/**
* Returns the lowercase character sequence of an ecma character.
*
Expand All @@ -471,8 +653,6 @@ lit_char_to_lower_case (ecma_char_t character, /**< input character value */
ecma_char_t *output_buffer_p, /**< [out] buffer for the result characters */
ecma_length_t buffer_size) /**< buffer size */
{
/* TODO: Needs a proper lower case implementation. See issue #323. */

JERRY_ASSERT (buffer_size >= LIT_MAXIMUM_OTHER_CASE_LENGTH);

if (character >= LIT_CHAR_UPPERCASE_A && character <= LIT_CHAR_UPPERCASE_Z)
Expand All @@ -481,6 +661,41 @@ lit_char_to_lower_case (ecma_char_t character, /**< input character value */
return 1;
}

#ifndef CONFIG_DISABLE_UNICODE_CASE_CONVERSION

ecma_length_t lowercase_sequence = search_in_bidirectional_conversion_tables (character, output_buffer_p, true);

if (lowercase_sequence != 0)
{
return lowercase_sequence;
}

int num_of_lowercase_ranges = NUM_OF_ELEMENTS (jerry_lower_case_ranges);

for (int i = 0, j = 0; i < num_of_lowercase_ranges; i += 2, j++)
{
int range_length = jerry_lower_case_range_lengths[j] - 1;
ecma_char_t start_point = jerry_lower_case_ranges[i];

if (start_point <= character && character <= start_point + range_length)
{
output_buffer_p[0] = (ecma_char_t) (jerry_lower_case_ranges[i + 1] + (character - start_point));
return 1;
}
}

lowercase_sequence = search_in_conversion_table (character,
output_buffer_p,
jerry_lower_case_conversions,
jerry_lower_case_conversion_counters);

if (lowercase_sequence != 0)
{
return lowercase_sequence;
}

#endif /* !CONFIG_DISABLE_UNICODE_CASE_CONVERSION */

output_buffer_p[0] = character;
return 1;
} /* lit_char_to_lower_case */
Expand All @@ -498,8 +713,6 @@ lit_char_to_upper_case (ecma_char_t character, /**< input character value */
ecma_char_t *output_buffer_p, /**< buffer for the result characters */
ecma_length_t buffer_size) /**< buffer size */
{
/* TODO: Needs a proper upper case implementation. See issue #323. */

JERRY_ASSERT (buffer_size >= LIT_MAXIMUM_OTHER_CASE_LENGTH);

if (character >= LIT_CHAR_LOWERCASE_A && character <= LIT_CHAR_LOWERCASE_Z)
Expand All @@ -508,6 +721,42 @@ lit_char_to_upper_case (ecma_char_t character, /**< input character value */
return 1;
}

#ifndef CONFIG_DISABLE_UNICODE_CASE_CONVERSION

ecma_length_t uppercase_sequence = search_in_bidirectional_conversion_tables (character, output_buffer_p, false);

if (uppercase_sequence != 0)
{
return uppercase_sequence;
}

int num_of_upper_case_special_ranges = NUM_OF_ELEMENTS (jerry_upper_case_special_ranges);

for (int i = 0, j = 0; i < num_of_upper_case_special_ranges; i += 3, j++)
{
int range_length = jerry_upper_case_special_range_lengths[j];
ecma_char_t start_point = jerry_upper_case_special_ranges[i];

if (start_point <= character && character <= start_point + range_length)
{
output_buffer_p[0] = (ecma_char_t) (jerry_upper_case_special_ranges[i + 1] + (character - start_point));
output_buffer_p[1] = (ecma_char_t) (jerry_upper_case_special_ranges[i + 2]);
return 2;
}
}

uppercase_sequence = search_in_conversion_table (character,
output_buffer_p,
jerry_upper_case_conversions,
jerry_upper_case_conversion_counters);

if (uppercase_sequence != 0)
{
return uppercase_sequence;
}

#endif /* !CONFIG_DISABLE_UNICODE_CASE_CONVERSION */

output_buffer_p[0] = character;
return 1;
} /* lit_char_to_upper_case */
Loading

0 comments on commit 1b5f839

Please sign in to comment.