Release Note

1.0.3

Fixed wrong method calls in Big5Prober#handleData, EUCTWProber#handleData, GB18030Prober#handleData. Thanks, Lersh99.
Fixed build.xml to emit Java1.5-compatible bytecode. Thanks, lnezda.
Included TestDetector.java in the source package. Thanks, lnezda.

1.0.2

Canonicalized charset names which are returned by UniversalDetector.getDetectedCharset().
All charset names which juniversalchardet can detect are now defined in org.mozilla.universalchardet.Constants. They can be compared to the result of UniversalDetector.getDetectedCharset().
Added SWIG interface.

1.0.1

Fixed an array bound violation in ISO2022JPSMModel. Thanks, Kazutoshi Satoda.

1.0

Initial release.

Comment by leo.kuzn...@gmail.com, Sep 4, 2007

misspelled: Constatants

Comment by leo.kuzn...@gmail.com, Sep 4, 2007

public class HebrewProber extends CharsetProber {
...
    public int handleData(byte[] buf, int offset, int length)
...
        for (int i=offset; i<maxPos; ++i) {
            c = buf[i];
            if (c == SPACE) {
                if (this.beforePrev != SPACE) {
                    if (isFinal(this.prev)) {
                        ++this.finalCharLogicalScore;
                    } else if (isNonFinal(this.prev)) {
                        ++this.finalCharVisualScore;
                    }
                }
            } else {
assert c != SPACE; // :-)
                if ((this.beforePrev == SPACE) &&
                     isFinal(this.prev) &&
>>> always true:     c != SPACE) { // can be removed
                    ++this.finalCharVisualScore;
                }
            }
            this.beforePrev = this.prev;
            this.prev = c;
        }

Comment by project member taksc...@gmail.com, Jul 23, 2008

Thank you for your report! Missspelling is now fixed. The redundant check in HebrewProber? also resides in the original code, so I'll leave it as-is ;-)

► Sign in to add a comment