[Lazarus] TProcess, UTF8, Windows

[Michael Van Canneyt]

On Fri, 13 Apr 2012, Jürgen Hestermann wrote:

> Felipe Monteiro de Carvalho schrieb:
>> It is better to document this issues once in 1 single place.
>
> Definitely not! Why is it such a problem so write the encoding type for a 
> function like CopyFile into the documentation for it? It belongs there. Of 
> course, there could also be a link to somewhere but it must be reachable from 
> the documentation of the function. The first information available is this 
> entry. When I have a question to any of these functions the first place to go 
> is google. And then I am stuck because there is no further information about 
> encoding of the string parameters.
>
>
> And we already have a large documentation about all of
>> this. Just read here:
>> http://wiki.freepascal.org/LCL_Unicode_Support
>
> Then this link has to be added to each of the affected functions.
> How else should someone get there from the documentation of CopyFile for 
> example?
> And it is only a very crude way to answer the simple question:
> What encoding is expected to the string paramters of CopyFile (or FindFirst)?
>
>
>> About FindFirst, you should know that there is FindFirstUTF8:
>> http://lazarus-ccr.sourceforge.net/docs/lcl/fileutil/findfirstutf8.html
>
> Yes, but why do I get this information only after lengthy discussions in this 
> list? Why does no link appear unter "See also" in the FindFirst documenation? 
> Aren't you annoyed by all these questions here in the list? If this would be 
> answered in the documenation then it would not come up here (and would not 
> annoy and displease so many new joiners to Free Pascal). If each and every 
> parameter would have the same encoding then it would not be such a problem 
> but in this case where each function has its own encoding it is essential.

Because FindFirst is part of the RTL, and FindFirstUTF8 is part of Lazarus.
FPC cannot refer to the lazarus docs.

Also:
Lazarus happened to choose UTF-8 as the encoding of their LCL. 
By contrast, MSEGui chose UTF-16 as it's encoding.

Work is in progress to make the RTL Unicode Aware, and as soon as it is so,
the encoding issues will be documented.

Michael.