LumeWeb
/
chardet
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

cleanup

This commit is contained in:
Dmitry Shirokov 2013-08-12 12:11:54 +10:00
parent 9fd949edf3
commit 26054b56e7
1 changed files with 18 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
encoding/mbcs.js
View File

@ -266,25 +266,26 @@ module.exports.big5 = function() {
this.nextChar = function(iter, det) { this.nextChar = function(iter, det) {
iter.index = iter.nextIndex; iter.index = iter.nextIndex;
iter.error = false; iter.error = false;
var firstByte;
firstByte = iter.charValue = iter.nextByte(det); var firstByte = iter.charValue = iter.nextByte(det);
if (firstByte < 0) {
if (firstByte < 0)
return false; return false;
}
if (firstByte <= 0x7f || firstByte == 0xff) { // single byte character.
// single byte character. if (firstByte <= 0x7f || firstByte == 0xff)
return true; return true;
}
var secondByte = iter.nextByte(det); var secondByte = iter.nextByte(det);
if (secondByte < 0) {
if (secondByte < 0)
return false; return false;
}
iter.charValue = (iter.charValue << 8) | secondByte; iter.charValue = (iter.charValue << 8) | secondByte;
if (secondByte < 0x40 ||
secondByte == 0x7f || if (secondByte < 0x40 || secondByte == 0x7f || secondByte == 0xff)
secondByte == 0xff) { iter.error = true;
iter.error = true;
}
return true; return true;
}; };
}; };
@ -343,14 +344,14 @@ function eucNextChar(iter, det) {
if (firstByte == 0x8f) { if (firstByte == 0x8f) {
// Code set 3. // Code set 3.
// Three byte total char size, two bytes of actual char value. // Three byte total char size, two bytes of actual char value.
thirdByte = iter.nextByte(det); thirdByte = iter.nextByte(det);
iter.charValue = (iter.charValue << 8) | thirdByte; iter.charValue = (iter.charValue << 8) | thirdByte;
if (thirdByte < 0xa1) { if (thirdByte < 0xa1) {
iter.error = true; iter.error = true;
} }
} }
} }
return (iter.done == false); return iter.done == false;
}; };
@ -479,7 +480,7 @@ module.exports.gb_18030 = function() {
break buildChar; break buildChar;
} }
} }
return (iter.done == false); return iter.done == false;
}; };
// TODO: This set of data comes from the character frequency- // TODO: This set of data comes from the character frequency-