«

I'm currently trying to use Visual Studio 2010 to build Shareaza, a
piece of open source software that theoretically should just build
from the C++ files without any problems. But when I attempted to
build it I got the following two error messages:

1>c:\<path-name>\shareaza\hashlib\utility.hpp(34: error C2664:
'_BitScanReverse' : cannot convert parameter 1 from 'uint32 *' to
'DWORD *'
1> Types pointed to are unrelated; conversion requires
reinterpret_cast, C-style cast or function-style cast
1>c:\<path-name>\shareaza\hashlib\utility.hpp(370): error C2664:
'_BitScanForward' : cannot convert parameter 1 from 'uint32 *' to
'DWORD *'
1> Types pointed to are unrelated; conversion requires
reinterpret_cast, C-style cast or function-style cast
1>
1>Build FAILED.

I took a look at the vicinity of lines 348 and 370 and got

345: inline uint32 highestBitSet(uint32 value)
346: {
347: uint32 index;
348: return _BitScanReverse( &index, value ) ? index : 0;
349: }

and

367: inline uint32 lowestBitSet(uint32 value)
368: {
369: uint32 index;
370: return _BitScanForward( &index, value ) ? index : 0;
371: }

and I looked <_BitScanReverse()> and <_BitScanForward()> up on the
Internet (at "http://msdn.microsoft.com/en-us/library/fbxyd7zd%28VS.
80%29.aspx" and "http://msdn.microsoft.com/en-us/library/wfd9z0bb%28VS.
80%29.aspx") and got:

unsigned char _BitScanReverse(
unsigned long * Index,
unsigned long Mask
);

and

unsigned char _BitScanForward(
unsigned long * Index,
unsigned long Mask
);

In view of these definitions, can anyone tell me (1) why this code is
generating these error messages, and (2) if it should be generating
these error messages, why it could _ever_ have compiled on _any_ C++
compiler. I remind you that I copied this code from an open source
repository; theoretically this code has already been compiled to
generate a working executable.

I look at this code and I think, all I really have to do is change the
type of each <index> variable from <unit32> to <unsigned long>. Is
that naive? Might I run into some unforeseen problems if I do that?
I'm a little hesitant to tweak the code, since I've been led to
believe that it was supposed to be working code before I downloaded
it. So what do you think? Should I tweak the code or should I try to
find some _other_ way to get this code working?

Kevin S

On 10/23/10 11:43 AM, KevinSimonson wrote:
> I'm currently trying to use Visual Studio 2010 to build Shareaza, a
> piece of open source software that theoretically should just build
> from the C++ files without any problems. But when I attempted to
> build it I got the following two error messages:
>
> 1>c:\<path-name>\shareaza\hashlib\utility.hpp(34: error C2664:
> '_BitScanReverse' : cannot convert parameter 1 from 'uint32 *' to
> 'DWORD *'
> 1> Types pointed to are unrelated; conversion requires
> reinterpret_cast, C-style cast or function-style cast
> 1>c:\<path-name>\shareaza\hashlib\utility.hpp(370): error C2664:
> '_BitScanForward' : cannot convert parameter 1 from 'uint32 *' to
> 'DWORD *'
> 1> Types pointed to are unrelated; conversion requires
> reinterpret_cast, C-style cast or function-style cast
> 1>
> 1>Build FAILED.
>
> I took a look at the vicinity of lines 348 and 370 and got
>
> 345: inline uint32 highestBitSet(uint32 value)
> 346: {
> 347: uint32 index;
> 348: return _BitScanReverse(&index, value ) ? index : 0;
> 349: }
>
> and
>
> 367: inline uint32 lowestBitSet(uint32 value)
> 368: {
> 369: uint32 index;
> 370: return _BitScanForward(&index, value ) ? index : 0;
> 371: }
>
> and I looked<_BitScanReverse()> and<_BitScanForward()> up on the
> Internet (at "http://msdn.microsoft.com/en-us/library/fbxyd7zd%28VS.
> 80%29.aspx" and "http://msdn.microsoft.com/en-us/library/wfd9z0bb%28VS.
> 80%29.aspx") and got:
>
> unsigned char _BitScanReverse(
> unsigned long * Index,
> unsigned long Mask
> );
>
> and
>
> unsigned char _BitScanForward(
> unsigned long * Index,
> unsigned long Mask
> );
>
> In view of these definitions, can anyone tell me (1) why this code is
> generating these error messages, and (2) if it should be generating
> these error messages, why it could _ever_ have compiled on _any_ C++
> compiler. I remind you that I copied this code from an open source
> repository; theoretically this code has already been compiled to
> generate a working executable.

The code is a bit wanky, mixing fixed with types and standard integral
types is never a good idea. The relative widths of these types my vary
across platforms and there's certainly no guarantee their pointers will
be compatible. Although why the prototype parameter is "unsigned long*"
and the error says "DWORD*" is rather odd.

> I look at this code and I think, all I really have to do is change the
> type of each<index> variable from<unit32> to<unsigned long>. Is
> that naive? Might I run into some unforeseen problems if I do that?
> I'm a little hesitant to tweak the code, since I've been led to
> believe that it was supposed to be working code before I downloaded
> it. So what do you think? Should I tweak the code or should I try to
> find some _other_ way to get this code working?

Have a look and how they define "unit32" would be a good start.

--
Ian Collins

On Sat, 23 Oct 2010 11:53:16 +1300, Ian Collins <ian->
wrote:

>On 10/23/10 11:43 AM, KevinSimonson wrote:
>> I'm currently trying to use Visual Studio 2010 to build Shareaza, a
>> piece of open source software that theoretically should just build
>> from the C++ files without any problems. But when I attempted to
>> build it I got the following two error messages:
>>
>> 1>c:\<path-name>\shareaza\hashlib\utility.hpp(34: error C2664:
>> '_BitScanReverse' : cannot convert parameter 1 from 'uint32 *' to
>> 'DWORD *'
>> 1> Types pointed to are unrelated; conversion requires
>> reinterpret_cast, C-style cast or function-style cast
>> 1>c:\<path-name>\shareaza\hashlib\utility.hpp(370): error C2664:
>> '_BitScanForward' : cannot convert parameter 1 from 'uint32 *' to
>> 'DWORD *'
>> 1> Types pointed to are unrelated; conversion requires
>> reinterpret_cast, C-style cast or function-style cast
>> 1>
>> 1>Build FAILED.
>>
>> I took a look at the vicinity of lines 348 and 370 and got
>>
>> 345: inline uint32 highestBitSet(uint32 value)
>> 346: {
>> 347: uint32 index;
>> 348: return _BitScanReverse(&index, value ) ? index : 0;
>> 349: }
>>
>> and
>>
>> 367: inline uint32 lowestBitSet(uint32 value)
>> 368: {
>> 369: uint32 index;
>> 370: return _BitScanForward(&index, value ) ? index : 0;
>> 371: }
>>
>> and I looked<_BitScanReverse()> and<_BitScanForward()> up on the
>> Internet (at "http://msdn.microsoft.com/en-us/library/fbxyd7zd%28VS.
>> 80%29.aspx" and "http://msdn.microsoft.com/en-us/library/wfd9z0bb%28VS.
>> 80%29.aspx") and got:
>>
>> unsigned char _BitScanReverse(
>> unsigned long * Index,
>> unsigned long Mask
>> );
>>
>> and
>>
>> unsigned char _BitScanForward(
>> unsigned long * Index,
>> unsigned long Mask
>> );
>>
>> In view of these definitions, can anyone tell me (1) why this code is
>> generating these error messages, and (2) if it should be generating
>> these error messages, why it could _ever_ have compiled on _any_ C++
>> compiler. I remind you that I copied this code from an open source
>> repository; theoretically this code has already been compiled to
>> generate a working executable.
>
>The code is a bit wanky, mixing fixed with types and standard integral
>types is never a good idea. The relative widths of these types my vary
>across platforms and there's certainly no guarantee their pointers will
>be compatible. Although why the prototype parameter is "unsigned long*"
>and the error says "DWORD*" is rather odd.

In Windows, DWORD is a typedef for unsigned long.

>
>> I look at this code and I think, all I really have to do is change the
>> type of each<index> variable from<unit32> to<unsigned long>. Is
>> that naive? Might I run into some unforeseen problems if I do that?
>> I'm a little hesitant to tweak the code, since I've been led to
>> believe that it was supposed to be working code before I downloaded
>> it. So what do you think? Should I tweak the code or should I try to
>> find some _other_ way to get this code working?
>
>Have a look and how they define "unit32" would be a good start.

It's uint32, not unit32 but I suspect propagation of typos here. In
boost, uint32 is typedef of uint32_t and it is a typedef for unsigned
int, thus no conversion exists for unsigned long to unsigned int.