Re: [Libreoffice] windows.h inclusions (was: Re: Review request for 3-4-2 build fix for Windows)

On Fri, 2011-07-22 at 03:56 +0200, Fridrich Strba wrote:
> I actually dare to contend that the mdds fix is much cleaner and more
> robust for following reasons:

        Heh ;-) I think it is pretty clear that including windows.h is a
shortcut for rapid insanity. I'd strongly prefer to write readable code,
rather than tangled / non-obvious code that just happens to compile with
windows.h present. Also - by going this route we just build more
un-detected & pointless pain into our Windows compiling minority.

        If boost includes windows.h pollution in an un-controlled way, then we
should re-consider using whatever parts of it do that IMHO; it is not a
good system abstraction that forces system-specific compile breakage on
all its users. Can we excise whatever nonsense we are including there
with a more specific include sub-set ?

> Thanks for reading until here

        Hey ;-) Solaris has/had something quite similar they loved to define
macros that would mangle struct members with glibc names like 'read' or
'write' IIRC.

        Hey ho,

                Michael.

-- 
 michael.meeks@novell.com  <><, Pseudo Engineer, itinerant idiot