/*  -*- C++ -*-

    SPDX-FileCopyrightText: 1998 Netscape Communications Corporation <developer@mozilla.org>

    SPDX-License-Identifier: MIT

*/

// for S-JIS encoding, observe characteristic:

// 1, kana character (or hankaku?) often have high frequency of appearance

// 2, kana character often exist in group

// 3, certain combination of kana is never used in japanese language

#ifndef nsEUCJPProber_h__

#define nsEUCJPProber_h__

#include "CharDistribution.h"

#include "JpCntx.h"

#include "nsCharSetProber.h"

#include "nsCodingStateMachine.h"

#include <memory>

namespace kencodingprober

{

class KCODECS_NO_EXPORT nsEUCJPProber : public nsCharSetProber

{

public:

    nsEUCJPProber();

    ~nsEUCJPProber() override = default;

    nsProbingState HandleData(const char *aBuf, unsigned int aLen) override;

    const char *GetCharSetName() override

    {

        return "EUC-JP";

    }

    nsProbingState GetState(void) override

    {

        return mState;

    }

    float GetConfidence(void) override;

protected:

    std::unique_ptr<nsCodingStateMachine> mCodingSM;

    nsProbingState mState = eDetecting;

    EUCJPContextAnalysis mContextAnalyser;

    EUCJPDistributionAnalysis mDistributionAnalyser;

    char mLastChar[2] = {0};

};

}

#endif /* nsEUCJPProber_h__ */