| @@ 1105-1111 (lines=7) @@ | ||
| 1102 | if (0 == (0x80 & ($in))) { |
|
| 1103 | // US-ASCII, pass straight through. |
|
| 1104 | $mBytes = 1; |
|
| 1105 | } elseif (0xC0 == (0xE0 & ($in))) { |
|
| 1106 | // First octet of 2 octet sequence |
|
| 1107 | $mUcs4 = ($in); |
|
| 1108 | $mUcs4 = ($mUcs4 & 0x1F) << 6; |
|
| 1109 | $mState = 1; |
|
| 1110 | $mBytes = 2; |
|
| 1111 | } elseif (0xE0 == (0xF0 & ($in))) { |
|
| 1112 | // First octet of 3 octet sequence |
|
| 1113 | $mUcs4 = ($in); |
|
| 1114 | $mUcs4 = ($mUcs4 & 0x0F) << 12; |
|
| @@ 1117-1123 (lines=7) @@ | ||
| 1114 | $mUcs4 = ($mUcs4 & 0x0F) << 12; |
|
| 1115 | $mState = 2; |
|
| 1116 | $mBytes = 3; |
|
| 1117 | } elseif (0xF0 == (0xF8 & ($in))) { |
|
| 1118 | // First octet of 4 octet sequence |
|
| 1119 | $mUcs4 = ($in); |
|
| 1120 | $mUcs4 = ($mUcs4 & 0x07) << 18; |
|
| 1121 | $mState = 3; |
|
| 1122 | $mBytes = 4; |
|
| 1123 | } elseif (0xF8 == (0xFC & ($in))) { |
|
| 1124 | /* First octet of 5 octet sequence. |
|
| 1125 | * |
|
| 1126 | * This is illegal because the encoded codepoint must be either |
|
| @@ 1136-1142 (lines=7) @@ | ||
| 1133 | $mUcs4 = ($mUcs4 & 0x03) << 24; |
|
| 1134 | $mState = 4; |
|
| 1135 | $mBytes = 5; |
|
| 1136 | } elseif (0xFC == (0xFE & ($in))) { |
|
| 1137 | // First octet of 6 octet sequence, see comments for 5 octet sequence. |
|
| 1138 | $mUcs4 = ($in); |
|
| 1139 | $mUcs4 = ($mUcs4 & 1) << 30; |
|
| 1140 | $mState = 5; |
|
| 1141 | $mBytes = 6; |
|
| 1142 | } else { |
|
| 1143 | /* Current octet is neither in the US-ASCII range nor a legal first |
|
| 1144 | * octet of a multi-octet sequence. |
|
| 1145 | */ |
|