I know how to deal with german Umlauts but I would like to understand.

Think of a *.csv file that contains strings like "Grünberg" and that has to be opened in Excel ( Office 2003 ).

It works if i use PWideChar(UTF8Decode(String)). Got this tipp here in the mailinglist!

But UTF8ToAnsi(String) fails.

Why is UTF8ToAnsi(String) not enough?

LHelp tells me about UTF8ToAnsi():
"Utf8ToAnsi converts an utf8-encode unicode string to an ansistring. It converts the string to a widestring and then converts the widestring to an ansistring."

That is NOT the same as PWideChar(UTF8Decode(String))?

Tipps are welcome.

Thank you!!

Obviously, because one produces an 8-bit string and the other a 16-bit string. How do you compare the result?

What compiler version?
Are you using LazUTF8? What is your DefaultSystemCodePage?

Mattias

Mattias Gaertner wrote

That is NOT the same as PWideChar(UTF8Decode(String))? Obviously, because one produces an 8-bit string and the other a 16-bit string. How do you compare the result?

I compare the results by open the *.csv file in Excel 2003 and see no correct "Grünberg"
Note: my Editor "TextPad" shows "PC UTF8" as file encoding if i use UTF8ToAnsi.

What compiler version?

Sorry, forgot that:

Lazarus 1.7 latest trunk FPC 3.0.0 i386-win32-win32/win64

Windows XP

Are you using LazUTF8?

Where to look for that?
"Alle Dateien haben bereits die richtige Kodierung" -> UTF8

What is your DefaultSystemCodePage? Mattias

in Systemsettings language its: "Automatisch ( oder Englisch)"

So the answer seems to be:
"one produces an 8-bit string and the other a 16-bit "

Did you start the csv file with the UTF8 BOM, so that Excel knows that the file is UTF-8 encoded?

UTF8BOM = #$EF#$BB#$BF

LazUTF8 is a unit, that sets the global variable
DefaultSystemCodePage to CP_UTF8. DefaultSystemCodePage defines the default encoding of AnsiString, so UTF8ToAnsi creates UTF-8. LazUTF8 is used by LCL applications.
If unsure use writeln or use the debugger to find out the value of DefaultSystemCodePage.

And PWideChar is a pointer, while UT8ToAnsi creates a String.

Mattias

Mattias Gaertner wrote

Did you start the csv file with the UTF8 BOM, so that Excel knows that the file is UTF-8 encoded? UTF8BOM = #$EF#$BB#$BF

In

https://en.wikipedia.org/wiki/Comma-separated_values

it says:

"Also, many regional versions of Excel will not be able to deal with Unicode in CSV."

and

"One simple solution when encountering such difficulties is to change the filename extension from .csv to .txt; then opening the file from an already running Excel instance with the "Open" command, where the user can manually specify the delimiters, encoding, format of columns, etc."

Thanks guys for your effort ... i stay with
PWideChar(UTF8Decode(String)) that works for me :-))

At the moment there is no need to know the details of character encoding history in computer science.

Landmesser John wrote

But UTF8ToAnsi(String) fails.

Maybe Utf8ToWinCP(AString) helps?
It's in unit LazUtf8.

Bart