https://bugs.kde.org/show_bug.cgi?id=427146

--- Comment #4 from Mark Wielaard <[email protected]> ---
(In reply to Mark Wielaard from comment #2)
> (In reply to Sam James from comment #1)
> > I was looking over stale copies of valgrind.h and noticed this, been fixed
> > on the glib side in
> > https://gitlab.gnome.org/GNOME/glib/-/commit/
> > 4aa051cafb389c5b0896217abb9740af8bac8495
> 
> Thanks, so at least it works somewhere.
> But that is a pretty stale version of valgrind.h. Would be good to get that
> upgraded.
> 
> So using stdint.h does require C99. But that might be reasonable to require
> these days.
> And it is only used for PLAT_x86_win32 with _MSC_VER. See bug #210935
> Which is a little odd, since that doesn't have a valgrind port.
> So this is only for building.
> 
> It also means we can put the #include <stdint.h> under #if defined(_MSC_VER).

I am reading the patch wrong.
We already use uintptr under _MSC_VER (without ever including stdint.h).
This patch also uses uintptr for the hunk under:

#if defined(PLAT_amd64_linux)  ||  defined(PLAT_amd64_darwin) \
    ||  defined(PLAT_amd64_solaris) \
    ||  defined(PLAT_amd64_freebsd) \
    ||  (defined(PLAT_amd64_win64) && defined(__GNUC__))

So that is win64, not win32. I assume mingw64?

I think replacing unsigned long int by uintptr_t is correct for all those 64bit
platforms.
As long as we indeed include stdint.h

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to