On Mon, 22 May 2000, Dmitry Timoshkov wrote:
> James Sutherland <[EMAIL PROTECTED]> wrote:
>
> >> As pointed out by Patrik Stridvall, version of FormatMessage for debugger
> >> must use only Unix API. But ability of using of localized messages, messages
> >> from system, user defined message tables along with other very useful options
> >> make use of this API very handy and convenient.
> >
> >The function in question never goes anywhere near any Win32 code; any
> >localised messages are strings internal to Wine. Where would any "user
> >defined message tables" come in? If the user wants to customise Wine in
> >any way, they can - but this is nothing to do with any Win32 aspects.
> >
> >Why try to make what must be a Unix function, invisible to the Win32 side,
> >look like a Win32 call? It cannot use the same resources or code anyway -
> >trying to make one look like the other will only lead to confusion.
>
> 1. FormatMessage() API should be (or already) implemented regardless
> used it somewhere in Wine or not.
Yes - except it isn't a Win32 function we need.
> 2. Using Win32 API in Wine ensures that its implementation is bug free.
No, it just makes any bug in that function have knockon effects on other,
apparently unrelated, areas.
> 3. Using of Win32 API functions in Wine shows its real, possibly
> undocumented behavior, and serves as good documentation.
Except where the use is completely inappropriate :-)
> 4. Add your own...
There are places where Wine should use Win32 functions. However, this is
purely a Unix-side function - IMO, it should be kept entirely separate
from the Win32 implementation. No Win32 code will ever "see" this
function, and it IS really a Unix function - so why try to make it look
like something it isn't?
James.
