Re: [kernel] add dummy HeapSetInformation()

[Robert Shearman]

dank(a)vroo.pair.com wrote:
> /***********************************************************************
> + *           HeapSetInformation   (KERNEL32.@)
> + * Change options for a given heap.
> + *
> + * NOTES
> + *	All options are ignored.
> + *
> + * RETURNS
> + *	TRUE: Success
> + *	FALSE: Failure
> + */
> +BOOL WINAPI HeapSetInformation(
> +              HANDLE heap, /* [in] Handle to the heap */
> +              DWORD class, /* [in] enum describing which option to set */
> +              LPCVOID poption, /* [in] Pointer to option value */
> +              DWORD optionlen  /* [in] Length of option value */
> +) {
> +    /* Only option available is Low Fragmentation Heap, which
> +     * is just a performance thing.  Allow but ignore; this should let
> +     * Lionhead Studios' Black & White 2 run.
> +     */
> +    return TRUE;
> +}
> +

You should make this function print a fixme and a minor nitpick that 
NOTES should probably be BUGS instead.

-- 
Rob Shearman

Continuum patch

Hi,

There is a patch in the AppDB entry for the game Continuum. (sorry I
don't have AppDB playing nice with DNS here yet so no link.) What needs
to be done to have this considered?

Regards,

-jan

Continuum patch

[Dan Kegel]

On 7/20/06, jan(a)janhurst.com <jan(a)janhurst.com> wrote:
> There is a patch in the AppDB entry for the game Continuum. (sorry I
> don't have AppDB playing nice with DNS here yet so no link.) What needs
> to be done to have this considered?

Here's the patch:

diff --git a/dlls/kernel/process.c b/dlls/kernel/process.c
index 33f9ee1..d50cb7d 100644
--- a/dlls/kernel/process.c
+++ b/dlls/kernel/process.c
@@ -2460,6 +2464,7 @@ HANDLE WINAPI OpenProcess( DWORD access,
     OBJECT_ATTRIBUTES   attr;
     CLIENT_ID           cid;

+if (access & PROCESS_VM_WRITE) return NULL;
     cid.UniqueProcess = (HANDLE)id;
     cid.UniqueThread = 0; /* FIXME ? */

That patch as written doesn't seem quite right.
I think what needs to happen is somebody needs to
write a conformance test for OpenProcess.
That will help everyone figure out what the right patch is.
- Dan

Continuum patch

[Vitaliy Margolen]

Thursday, July 20, 2006, 11:53:37 PM, Dan Kegel wrote:
> On 7/20/06, jan(a)janhurst.com <jan(a)janhurst.com> wrote:
> > There is a patch in the AppDB entry for the game Continuum. (sorry I
> > don't have AppDB playing nice with DNS here yet so no link.) What needs
> > to be done to have this considered?
> Here's the patch:
> diff --git a/dlls/kernel/process.c b/dlls/kernel/process.c
> index 33f9ee1..d50cb7d 100644
> --- a/dlls/kernel/process.c
> +++ b/dlls/kernel/process.c
> @@ -2460,6 +2464,7 @@ HANDLE WINAPI OpenProcess( DWORD access,
>      OBJECT_ATTRIBUTES   attr;
>      CLIENT_ID           cid;
> +if (access & PROCESS_VM_WRITE) return NULL;
>      cid.UniqueProcess = (HANDLE)id;
>      cid.UniqueThread = 0; /* FIXME ? */
> That patch as written doesn't seem quite right.
> I think what needs to happen is somebody needs to
> write a conformance test for OpenProcess.
> That will help everyone figure out what the right patch is.
> - Dan

The problem this patch is for is covered in this bug:
http://bugs.winehq.org/show_bug.cgi?id=4317

That patch is as hackish as it gets. And it won't ever make it into
Wine. It's made for one app only to "fix" the anti-cheating check.

Vitaliy.