Hi,

Is there a simple method of searching a string for a unicode substring and replacing it with the correct character?

At this stage, I'm not really worried about the unicode itself but I have substrings within my string like '\U+00B0' which I'd like to replace with chr(176).

Is there a simple way of doing this?

~Dave

Dave,

I've tried some regex units implemented in pascal but they are either not complete or buggy or slow.
The best I've found was PCRE which is kind of a standard in many scripting languages.
I'm not sure if there is a lazarus wrapper but I'm using it with delphi.

http://www.renatomancuso.com/software/dpcre/dpcre.htm

Maik

Hi Maik,

This is what I've got.

It works, but it's a bit tacky!

  i:=Pos('\U+',mPCharArray);
  while i>0 do
  begin
    subS:=Copy(mPCharArray,i+3,4);
    mainString:=mPCharArray;
    Delete(mainString,i,7);
    Insert(chr(Hex2Dec(subS)),mainString,i);
    mPCharArray:=PChar(mainString);
    i:=Pos('\U+',mPCharArray);
  end;

I was hoping that someone could suggest something a little more elegant!

The new Delphi XE also has regex units added, but it seems they go for a .NET compatible regex. But I don't the exact differences of MS' implementation with PCRE.

FPC come with a regex unit: packages/regexpr/src/regex.pp

-Michael

Maik Wojcieszak wrote

I've tried some regex units implemented in pascal but they are either not complete or buggy or slow.

Can you show me the bug-reports? That way it can be fixed. (I suppose you've also used the regex-unit which comes with fpc)

Joost

Hi,

I've tried some regex units implemented in pascal but they are either not complete or buggy or slow.

Did you try TRegExpr by Andrey V. Sorokin?

That is the best one.

Benito

Hi,

sorry for beeing too general about that. I was testing this some years ago. I've NEVER tested regex on lazarus.

> Did you try TRegExpr by Andrey V. Sorokin?

I've tested a delphi unit I found on the internet "TRegExpr library from Andrey V. Sorokin" which was
only partially implemented at that time (about 10yrs ago).

The main problem was the lack of some features (positive/negative lookahead) and differences in the syntax compared with perl regex at that point of the development because I am using python regex (perl regex) in some parts of
my sw.

Again, sorry for beeing too general about that.
Maik

I'm using that on Delphi and to a much lesser extent on Lazarus. It is adequate for my requirements, although to a large extent I'm able to tailor how I solve application-level problems to what underlying facilities I have available.