SCI_POSITIONRELATIVE

[Neil Hodgson]

A new API, SCI_POSITIONRELATIVE, was implemented to optimize navigating by characters.

A document encoded in UTF-8 or DBCS may use multiple bytes for some characters. An application or platform may work with characters instead of bytes and there is a need to translate between the character-oriented view and Scintilla's byte-oriented view. For example, a compiler may report a warning for characters 12 to 18 inclusive on line 5. Previously, the SCI_POSITIONAFTER or SCI_POSITIONBEFORE methods could be called multiple times to navigate from the start of line 5 to character 12 on that line and then to (the end of) character 18. SCI_POSITIONRELATIVE avoids the loop and calling overhead.

Several platforms that Scintilla runs on use UTF-16 and treat non-BMP characters as 2 characters. It may simplify client code on these platforms to provide another API similar to SCI_POSITIONRELATIVE but which treats non-BMP characters as 2 characters to match the platform.

Here's an example interfacing to a spell checker that reports ranges in terms of UTF-16 code units where the non-BMP horse character 𐂃 confuses the position of the second mistake.



Neil

[Mike Lischke]

Hey Neil,

Post by Neil Hodgson
A new API, SCI_POSITIONRELATIVE, was implemented to optimize navigating by characters.
A document encoded in UTF-8 or DBCS may use multiple bytes for some characters. An application or platform may work with characters instead of bytes and there is a need to translate between the character-oriented view and Scintilla's byte-oriented view. For example, a compiler may report a warning for characters 12 to 18 inclusive on line 5. Previously, the SCI_POSITIONAFTER or SCI_POSITIONBEFORE methods could be called multiple times to navigate from the start of line 5 to character 12 on that line and then to (the end of) character 18. SCI_POSITIONRELATIVE avoids the loop and calling overhead.

Thanks for that new API. I can certainly use this to simply my code completion code. What I'd also appreciate is an API to convert text position <-> column/row pair (respecting tabs).

Mike

[Neil Hodgson]

Post by Mike Lischke
What I'd also appreciate is an API to convert text position <-> column/row pair (respecting tabs).

SCI_LINEFROMPOSITION, SCI_GETCOLUMN, SCI_FINDCOLUMN

Neil