Re: setupapi/tests: Remove an unneeded cast.

[Hans Leidekker]

On Thursday 08 January 2009 00:05:01 Francois Gouget wrote:
>  memset(buffer, 0x55, sizeof(buffer));
> - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, (LPDWORD)0);
> - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, NULL);
> ok(retval != ~0u, "failed find 'UILEVEL' in string table\n");
> ok(memcmp(buffer, &data, 4), "unexpected data\n");

This test is there to show that the prototype of StringTableLookUpStringEx
is broken; the last parameter is treated as an integer instead of a pointer
and this notation serves to document that fact. But I can see how this will
repeatedly trigger janitorial scripts, so I do not object to this patch
going in.

-Hans

setupapi/tests: Remove an unneeded cast.

[Henri Verbeet]

2009/1/8 Hans Leidekker :
> On Thursday 08 January 2009 00:05:01 Francois Gouget wrote:
> >  memset(buffer, 0x55, sizeof(buffer));
> > - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, (LPDWORD)0);
> > - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, NULL);
> > ok(retval != ~0u, "failed find 'UILEVEL' in string table\n");
> > ok(memcmp(buffer, &data, 4), "unexpected data\n");
>
> This test is there to show that the prototype of StringTableLookUpStringEx
> is broken; the last parameter is treated as an integer instead of a pointer
> and this notation serves to document that fact. But I can see how this will
> repeatedly trigger janitorial scripts, so I do not object to this patch
> going in.

Using 0 instead of NULL should work just as well, if that makes it any clearer.

setupapi/tests: Remove an unneeded cast.

[Alexandre Julliard]

Hans Leidekker  writes:
> On Thursday 08 January 2009 00:05:01 Francois Gouget wrote:
> >  memset(buffer, 0x55, sizeof(buffer));
> > - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, (LPDWORD)0);
> > - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, NULL);
> > ok(retval != ~0u, "failed find 'UILEVEL' in string table\n");
> > ok(memcmp(buffer, &data, 4), "unexpected data\n");
>
> This test is there to show that the prototype of StringTableLookUpStringEx
> is broken; the last parameter is treated as an integer instead of a pointer
> and this notation serves to document that fact. But I can see how this will
> repeatedly trigger janitorial scripts, so I do not object to this patch
> going in.

Where does the prototype come from?  AFAICS it's not in the PSDK. Maybe
it's just wrong.

setupapi/tests: Remove an unneeded cast.

[Hans Leidekker]

On Thursday 08 January 2009 11:22:59 Alexandre Julliard wrote:
> > This test is there to show that the prototype of StringTableLookUpStringEx
> > is broken; the last parameter is treated as an integer instead of a pointer
> > and this notation serves to document that fact. But I can see how this will
> > repeatedly trigger janitorial scripts, so I do not object to this patch
> > going in.
>
> Where does the prototype come from?  AFAICS it's not in the PSDK. Maybe
> it's just wrong.

I don't know where the prototype came from, there was a stub there already
which entered Wine with commit 8196e2d0d0dbd04325201e14d95f88cad3cfd00f.

So we might as well change the prototype, given its demonstrated behavior.

-Hans

setupapi/tests: Remove an unneeded cast.

[Francois Gouget]

On Thu, 8 Jan 2009, Hans Leidekker wrote:
[...]
> I don't know where the prototype came from, there was a stub there already
> which entered Wine with commit 8196e2d0d0dbd04325201e14d95f88cad3cfd00f.

I did not find it in my SDKs either. And a web search turned no useful 
result.
> So we might as well change the prototype, given its demonstrated behavior.

Works for me. I'm sending a patch...

setupapi/tests: Remove an unneeded cast.

[Steven Edwards]

On Thu, Jan 8, 2009 at 5:22 AM, Alexandre Julliard  wrote:
> Where does the prototype come from?  AFAICS it's not in the PSDK. Maybe
> it's just wrong.

When Eric first implemented those functions they were mostly
undocumented. I believe I checked the prototypes verses some VB or
Pascal code I found on google where someone had reversed the interface
but I can't find the exact source I used to check things when writing
the first round of tests.

According to this link from google search, the prototypes are in MASM
include files going back to 2001. You might want to check the others
as well.

http://www.cecs.csulb.edu/~hill/cecs325/MASM32/INCLUDE/SETUPAPI.INC

setupapi/tests: Remove an unneeded cast.

[Steven Edwards]

On Thu, Jan 8, 2009 at 12:29 PM, Steven Edwards  wrote:
> According to this link from google search, the prototypes are in MASM
> include files going back to 2001. You might want to check the others
> as well.
>
> http://www.cecs.csulb.edu/~hill/cecs325/MASM32/INCLUDE/SETUPAPI.INC

Heh ignore that part of the spam...I didn't see that the MASM include
was not really helpful at all for the prototype...

setupapi/tests: Remove an unneeded cast.

[Paul TBBle Hampson]

On Thu, Jan 08, 2009 at 10:05:28AM +0100, Hans Leidekker wrote:
> On Thursday 08 January 2009 00:05:01 Francois Gouget wrote:
> >  memset(buffer, 0x55, sizeof(buffer));
> > - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, (LPDWORD)0);
> > - retval = pStringTableLookUpStringEx(table, uilevel, ST_CASE_SENSITIVE_COMPARE, buffer, NULL);
> > ok(retval != ~0u, "failed find 'UILEVEL' in string table\n");
> > ok(memcmp(buffer, &data, 4), "unexpected data\n");
> This test is there to show that the prototype of StringTableLookUpStringEx
> is broken; the last parameter is treated as an integer instead of a pointer
> and this notation serves to document that fact.

Wouldn't that actually be different type sizes under Win64?