Get real offset depending on line terminator

Hi Tobias,

Actually DocumentPosition does also indicate the Character, which is the offset relative to the line's start.

Okay, this helps. I might have a problem with unicode characters, I got to check that out.
Again, the other way (to translate the internal offest to the "real" offset) is not possible?

However I think the documentation and name is misleading. The documentation says "The character of this DocumentPosition", so I assumed it is the actual character with an integer representation (like 41 for A) and not the offset.

The same thing applies for unicode character with more than 1 byte too, right?
Simply put your offset counter only counts the visible characters and not the number of bytes in the text file.
So I need a second counter in my compiler that counts the number of visible characters the way you do (at the moment he only knows the number of bytes).

Hi Tobias,

Well if you know all your real line terminators end with \r\n, you could take our offset and add the line index to it to get the "real" offset in the source file.

For instance with this file, where the | is the caret and the target position...
A
B
C|

It's our offset 6 and line index 2. So the "real" offset if this text had \r\n ends would be 6 + 2 = 8.

The DocumentPosition is a line/character index pair, similar to the information you see in the Visual Studio status bar where it indicates Ln (Line) and Ch (Character).

Once any file is loaded into a .NET string, you can no longer track bytes since strings are Unicode. So all our offsets and text positions are relative to Unicode character counts. It's up to the encoding you load and save with to determine how to translate that from/to bytes.

Thanks for your answer.
Unfortunately we cannot guarantee the line terminators are \r\n unless we save the file with the correct encoding before we load it.
Our compiler currently works with the byte position and we don't want to parse the current line to look for characters with multiple bytes (mainly asian characters) to get the byte-offset. So we will implement a second counter in our compiler that will return the same offset as SyntaxEditor does.

[Modified at 09/09/2011 12:43 AM]