Re: Fix Sparse Warnings for server/

[Alexandre Julliard]

Robert Shearman <rob(a)codeweavers.com> writes:
> Changelog:
> Fix sparse warnings.

sparse should be taught that 0 is a perfectly fine value (and IMO
preferable to NULL) for handles.

-- 
Alexandre Julliard
julliard(a)winehq.org

Fix Sparse Warnings for server/

[Michael Stefaniuc]

Alexandre Julliard wrote:
> Robert Shearman <rob(a)codeweavers.com> writes:
> > Changelog:
> > Fix sparse warnings.
>
> sparse should be taught that 0 is a perfectly fine value (and IMO
> Linus wont agree to that. I didn't ask him but seen some of his posts to 
> lklm regarding this matter.
> preferable to NULL) for handles.
> Rob, I already talked to Alexandre about this at the Wineconf; that's 
> why i didn't send in such a patch.

bye
	michael
-- 
Michael Stefaniuc               Tel.: [phone]
Sr. Network Engineer            Fax.: [phone]
Red Hat GmbH                    Email: mstefani(a)redhat.com
Hauptstaetterstr. 58            http://www.redhat.de/
D-70178 Stuttgart

Fix Sparse Warnings for server/

[Robert Shearman]

Alexandre Julliard wrote:
> Robert Shearman <rob(a)codeweavers.com> writes:
> > Changelog:
> > Fix sparse warnings.
>
> sparse should be taught that 0 is a perfectly fine value (and IMO
> preferable to NULL) for handles.

I don't really understand this reasoning. Should gcc also be taught not 
to warn when we try to assign -1 to a handle? Are we using the wrong 
type for usr_handle_t then?

Rob

Fix Sparse Warnings for server/

[Alexandre Julliard]

Robert Shearman <rob(a)codeweavers.com> writes:
> Alexandre Julliard wrote:
> > sparse should be taught that 0 is a perfectly fine value (and IMO
> > preferable to NULL) for handles.
>
> I don't really understand this reasoning. Should gcc also be taught
> not to warn when we try to assign -1 to a handle? Are we using the
> wrong type for usr_handle_t then?

In a sense, yes. Handles are conceptually much closer to integers than
pointers, but they are pointers in the Windows API so that the
different handles can have distinct types. It's really just a
workaround for the fact that all integer types are freely
interchangeable in C.

-- 
Alexandre Julliard
julliard(a)winehq.org