ndyacclex Remove AnsiString and AnsiChar and set of char usage. Add TCharacter usage for unicode classes

I'll try do this.

See also: http://docwiki.embarcadero.com/VCL/en/Character.TCharacter

Aug 21 '14 12:08 Nashev

In http://roman.yankovsky.me/?p=577 you wrote that you do it yourself. Is there any progress?

Aug 21 '14 13:08 Nashev

Yes, this was my plan, but unfortunately there is no progress. I would very appreciate your help.

Aug 22 '14 07:08 RomanYankovsky

Did you have a chance to start this work?

Sep 25 '14 16:09 RomanYankovsky

Not yet, sorry. I've recognized that is not necessary for my job now.

Oct 06 '14 14:10 Nashev

I've looked at this task again and found at the current original UNIX-way yacc/lex same problem with the ANSI charset usage. There are widely used workaround like declare UTF-8 sequences as special text in lex file. And some feeling of doubt and qualm has gone to me.

see for example: http://compiler.su/nado-li-ispolzovat-yacc-lex-i-podobnye-instrumenty.php http://www.mkssoftware.com/support/kb/default.asp?article=54

But Python implementation can use unicode as native strings in puthon and allow specially enable some unicode support for used RegExps: http://www.dabeaz.com/ply/ply.html#ply_nn22 ("4.20 Miscellaneous Issues" section notes unicode)

Exists same implementations with unicode support declared: http://www.visoracle.com/download/freeware30/cssxml/sas.html

And documentation of GNU lexx implementation Bison declare environment dependences of encoding: see http://www.gnu.org/software/bison/manual/bison.html#Symbols

Conversion to sring and TCharacter must be performed.

May 05 '15 11:05 Nashev