@upper[] and @lower[] don't like the accented characters

[mstef]

If a string contains accented characters, @upper[] does not write them in capital letters:
F:\>echo %@upper[Les élèves français apprennent où placer les caractères accentués] LES éLèVES FRANçAIS APPRENNENT Où PLACER LES CARACTèRES ACCENTUéS
instead of:
LES ÉLÈVES FRANÇAIS APPRENNENT OÙ PLACER LES CARACTÈRES ACCENTUÉS

And lower[] has the opposite problem:
F:\>echo %@lower[LES ÉLÈVES FRANÇAIS APPRENNENT OÙ PLACER LES CARACTÈRES ACCENTUÉS] les ÉlÈves franÇais apprennent oÙ placer les caractÈres accentuÉs

@caps[] is affected too:
F:\>echo %@caps[cet étudiant vit à Paris] Cet étudiant Vit à Paris

Is it possible to fix these functions?

Tested with TCMD 26.02.43 and TCC-RT 29.00.17.

 

[Charles Dye]

It looks like these functions only know ASCII. Handling all the alpha characters in Unicode would not be reasonable. But I'm guessing you only really care about Latin-1, right?

Replacing @UPPER and @LOWER with a plugin would be trivial. @CAPS might be a little more interesting; @CAPS has some funky idiosyncratic parsing to handle the optional separator argument.

 

[mstef]

Indeed, the Latin-1 charset (or better the Latin-9 to have the French ligature œ) is sufficient for me. And I need the @UPPER and @LOWER only, not @CAPS. I had tested it just to see if it was also affected.

Thanks for the Plugins suggestion, it's a good idea. I just downloaded the SDK, I will now study the documentation.

 

[vefatica]

Maybe you could roll your own (one character at a time). These ought to be pretty easy to use in a in a "DO /C" loop in a subroutine.

v:\> echo %@char[%@winapi[user32.dll,CharLower,%@unicode[É]]]
é

v:\> echo %@char[%@winapi[user32.dll,CharUpper,%@unicode[é]]]
É

 

[Juppycmd]

Thanks for the Plugins suggestion, it's a good idea. I just downloaded the SDK, I will now study the documentation.

Unless you really want to wite a plugin, you could try replacing each possible accented char (there can't that be that many, can there?) with set upper=%@REPLACE[%@CHAR[...],%@CHAR[...],%lower]

 

[Charles Dye]

I've been playing with something like this:

wchar_t UpperChar( wchar_t ch )
{
    if ( ( ch >= 'a' ) && ( ch <= 'z' ) )
        return ch - 32;

    if ( ( ch >= 0x00e0 ) && ( ch <= 0x00fe ) )
        if ( ch != 0x00f7 )
            return ch - 32;

    if ( ch == 0x0153 )
        return 0x0152;

    return ch;
}


wchar_t LowerChar( wchar_t ch )
{
    if ( ( ch >= 'A' ) && ( ch <= 'Z' ) )
        return ch + 32;

    if ( ( ch >= 0x00c0 ) && ( ch <= 0x00de ) )
        if ( ch != 0x00d7 )
            return ch + 32;

    if ( ch == 0x0152 )
        return 0x0153;

    if ( ch == 0x1e9e )
        return 0x00df;

    return ch;
}

This is a pretty naïve approach, though. It wouldn't work for, say, Greek, where sigma has two different lowercase forms. Or for uppercasing the German ß to SS.

 

[vefatica]

If you use CharUpper/CharLower on a single character, it returns the converted character; that's what I did in post #4. But CharUpper and CharLower can do a whole string in one call. That's done in-place and the function returns the same pointer_to_string that you supplied as its argument. I couldn't figure out a way to wrap that up in @WINAPI. Did I miss something?

 

[Charles Dye]

If you use CharUpper/CharLower on a single character, it returns the converted character; that's what I did in post #4. But CharUpper and CharLower can do a whole string in one call. That's done in-place and the function returns the same pointer_to_string that you supplied as its argument. I couldn't figure out a way to wrap that up in @WINAPI. Did I miss something?

I like this approach better than my own. @UPPER and @LOWER become practically one-liners. @CAPS is still kind of a mess, though.

 

[Joe Caverly]

Here's the PureBasic code that I used in a plugin for a new @upper function;

ProcedureCDLL.i f_Upper(*lpszString)
  Static theString.s
  Static theUpperString.s
  
  theString = PeekS(*lpszString)
  
  If Len(theString) < 1
    WriteConsoleN("USAGE: @Upper[" + "Les élèves français apprennent où placer les caractères accentués" + "]")
    ProcedureReturn #Null
  Else
    theString = Trim(theString)
  EndIf
  
  theUpperString = PeekS(CharUpper_(theString))
  
  PokeS(*lpszString,theUpperString)
  
  ProcedureReturn #Null
EndProcedure

This produces;

E:\...\plugin>echo %@upper[Les élèves français apprennent où placer les caractères accentués]
LES ÉLÈVES FRANÇAIS APPRENNENT OÙ PLACER LES CARACTÈRES ACCENTUÉS

Joe

     _x64: 1
   _admin: 1
_elevated: 1

TCC  29.00.17 x64   Windows 10 [Version 10.0.19044.2604]

 

[Joe Caverly]

Please find attached a plugin, TCCUtils, which has the modified @upper function using the code from post #9.

plugin /i TCCUtils.dll for more info.

Joe

 

[Charles Dye]

I'm uploading my version here: CaseFix Plugin

 

[vefatica]

Tough ones ... eh? These work for me.

INT WINAPI f_UCASE(LPWSTR psz)
{
    CharUpper(psz);
    return 0;
}

INT WINAPI f_LCASE(LPWSTR psz)
{
    CharLower(psz);
    return 0;
}

 

[mstef]

Thank you to all! Your solutions work well. Charles et Joe, thanks for your plugins.
And all your answers have taught me a lot about the possibilities of plugins and @WINAPI.