Initial implementation of EUC-JP

6 years ago · 8dfb1ba984
6 changed files with 337 additions and 1 deletions
--- a/lib/Encoding/EUCJP.php
+++ b/lib/Encoding/EUCJP.php
--- a/tests/cases/Encoding/TestEUCJP.php
+++ b/tests/cases/Encoding/TestEUCJP.php
--- a/tests/phpunit.xml
+++ b/tests/phpunit.xml
@ -23,6 +23,7 @@
        <file>cases/Encoding/TestUTF16BE.php</file>
        <file>cases/Encoding/TestSingleByte.php</file>
        <file>cases/Encoding/TestXUserDefined.php</file>
+        <file>cases/Encoding/TestEUCJP.php</file>
        <file>cases/Encoding/TestGB18030.php</file>
        <file>cases/Encoding/TestBig5.php</file>
        <file>cases/Encoding/TestEUCKR.php</file>
--- a/tools/mkindex.php
+++ b/tools/mkindex.php
@ -1,7 +1,7 @@
 <?php
 $labels = [
    'big5'                => "big5",
-    //'euc-jp'              => "eucjp",
+    'euc-jp'              => "eucjp",
    'euc-kr'              => "euckr",
    'gb18030'             => "gb18030",
    'ibm866'              => "single_byte",
@ -130,6 +130,13 @@ function euckr(string $label) {
    echo "const TABLE_CODES = $codes;\n";
 }

+function eucjp(string $label) {
+    $jis0208 = make_decoder_point_array(read_index("jis0208", "https://encoding.spec.whatwg.org/index-jis0208.txt"));
+    $jis0212 = make_decoder_point_array(read_index("jis0212", "https://encoding.spec.whatwg.org/index-jis0212.txt"));
+    echo "const TABLE_JIS0208 = $jis0208;\n";
+    echo "const TABLE_JIS0212 = $jis0212;\n";
+}
+
 // generic helper functions

 function read_index(string $label, string $url): array {
--- a/tools/mktest.php
+++ b/tools/mktest.php
@ -55,6 +55,12 @@ function make_test(string $label, string $url): array {
        $code = hexdec($match[1]);
        if ($label=="gb18030" && $bytes=="A8BC") { // this test is incorrect or out of date; both Vivaldi and Firefox yield code point 7743
            $code = 7743;
+        } elseif ($label=="euc-jp") { // these tests are out of date
+            if ($bytes == "5C") {
+                $code = 92;
+            } elseif ($bytes == "7E") {
+                $code = 126;
+            }
        }
        // convert the code point to decimal
        $out[] = $code;
--- a/tools/test-eucjp.html
+++ b/tools/test-eucjp.html
@ -0,0 +1,16 @@
+<!DOCTYPE html>
+<meta charset=euc-jp>
+<script>
+var sampleStrings = {
+    'empty string': "",
+    // valid single characters
+    'sanity check': "40",
+    'former ASCII deviations': "5C 7E",
+    'problem': "A1DD",
+};
+var sampleCharacters = {
+};
+var seekCodePoints = [
+];
+</script>
+<script src="test.js"></script>