hebrew bug fix

[gnudles]

under the file bgl.cc
under class
BglArticleRequest
I added a method:

Code: Select all

void fixHebString(string & hebStr);

and implemented

Code: Select all

void BglArticleRequest::fixHebString(string & hebStr)
{
    wstring hebWStr=Utf8::decode(hebStr);
    for (unsigned int i=0; i<hebWStr.size();i++)
    {
        if (hebWStr[i]>=224 && hebWStr[i]<=250)
            hebWStr[i]+=1488-224;
    }
    hebStr=Utf8::encode(hebWStr);
}


under the method:

Code: Select all

void BglArticleRequest::run()

I moved the line:

Code: Select all

static Language::Id hebrew = LangCoder::code2toInt( "he" );

to the begining of the method (before the next loop) and in this loop:

Code: Select all

for( unsigned x = 0; x < chain.size(); ++x )

after the line:

Code: Select all

wstring headwordStripped =
      Folding::applySimpleCaseOnly( Utf8::decode( removePostfix( headword ) ) );

I added:

Code: Select all

displayedHeadword= displayedHeadword.size() ? displayedHeadword : headword;
    if (dict.idxHeader.langFrom == hebrew)
    {
        fixHebString(articleText);
        fixHebString(displayedHeadword);
    }


and few lines after, I changed the insertation into:

Code: Select all

mapToUse.insert( pair< wstring, pair< string, string > >(
      Folding::applySimpleCaseOnly( Utf8::decode( headword ) ),
      pair< string, string >( displayedHeadword,
        articleText ) ) );


in the same method I changed the code in the loops of "mainArticles" and "alternateArticles". instead of

Code: Select all

      result += "<h3>";
      result += postfixToSuperscript( i->second.first );
      result += "</h3>";


I did:

Code: Select all

      if (dict.idxHeader.langFrom == hebrew)
          result += "<h3 style=\"text-align:right;direction:rtl\">";
      else
          result += "<h3>";
      result += postfixToSuperscript( i->second.first );
      result += "</h3>";


in both of the loops... and this should fix most of the problems in hebrew
note that the rtl fix should be applied to all rtl languages, and not only hebrew

[gnudles]

to fix the problem with the last capital letter in the definition I added another method to the class:

Code: Select all

void fixHebArticle(string & hebArticle);

and implemented it like that:

Code: Select all

void BglArticleRequest::fixHebArticle(string & hebArticle)
{
    int nulls=hebArticle.size()-1;
    while ((hebArticle[nulls]<=32 && hebArticle[nulls]>=0) || (hebArticle[nulls]>=65 && hebArticle[nulls]<=90))//special chars and A-Z
    {
        nulls--;
    }
    hebArticle.erase (hebArticle.begin()+nulls+1, hebArticle.end());
}


the call to the method comes right after the call to

Code: Select all

fixHebString(articleText);

so we actually have this code:

Code: Select all

if (dict.idxHeader.langFrom == hebrew)
    {
        fixHebString(articleText);
        fixHebArticle(articleText);
        fixHebString(displayedHeadword);
    }


instead of

Code: Select all

if (dict.idxHeader.langFrom == hebrew)
    {
        fixHebString(articleText);
        fixHebString(displayedHeadword);
    }


in the last fix...............

[ikm]

Please make proper patches (use the "diff" tool to generate them).

Could you explain what exactly is going on here:

if (hebWStr[i]>=224 && hebWStr[i]<=250)
hebWStr[i]+=1488-224;

[gnudles]

the unicode characters between:
224 -250
are not hebrew characters.
they are the values of hebrew characters in other encodings like windows-1255 or ISO-8859-8.
the value of the first letter in hebrew is 1488 (the letter alef) and the value of the letter alef in windows-1255 or ISO-8859-8 is 224.
the last letter in hebrew is tav and is represented by 250 or 1504 in unicode.

btw,
I'll make a patch....

[gnudles]

there it is...

[nit_ar]

Thanks for the fix.

Is this fix/patch will be formally applied to GD ?

I would like to try and test it, but I have never tried to build a binary by myself.

[mehdishahini]

How to use this patch in dictionary

[nit_ar]

This patch has been already applied to the master trunk of the source code by Goldendict author.

You can either:
wait for a new binary release
or:
pull/download recent master tree of the source code + compilation tools and compile the binary by yourself.

[Tvangeste]

You can either:
wait for a new binary release
or:
pull/download recent master tree of the source code + compilation tools and compile the binary by yourself.

Or, alternatively, you could get an early-access version, compiled for Windows, from here:
https://github.com/goldendict/goldendic ... or-Windows