[Geany-devel] [CODE REVIEW] Changes to encodings

[Lex Trotman]

> That's what I meant for set_text_full(). Basically we don't need to show
> the real SCI api here, just what's useful to us. If it means wrapping to
> calls, I'd say it's just fine.
> And having such a wrapper will help user (here meaning "client" code) to
> properly give a length parameter where set_text() used to be used.
>

The sciwrapper functions that just wrap scintilla messages are
documented by the scintilla docs.

If functions are added to sciwrapper that don't just wrap Scintilla
messages or change the names then they need to be documented.  Since
we use doxygen comments for API generation the only option for
wrappers that are not part of the API is a plain source code comment
and hope people see it :-(

So if possible the sciwrapper functions should have the same name as
the message.  In this particular instance backward compatibility means
that we can't do that, but IIUC the new function is part of the API
anyway.


> Hum, yeah, that must be the worst sentence I written since years :D
> I just mean I don't want to break the ABI now, and again in two days (or
> whatever short delay) because we figure out we have to change some other
> API to support NULs in another place, and again and again.
>

Good point.

Cheers
Lex