Re: kernel32.dll: implementation of GetVolumePathNamesForVolumeNameA/W

[Dmitry Timoshkov]

Tuomo Mattila  wrote:
> +BOOL WINAPI GetVolumePathNamesForVolumeNameA(LPCSTR volumename, LPSTR volumepathname, DWORD buflen, PDWORD returnlen)
> +{
> -
> - BOOL ret = FALSE;
> - WCHAR* volumenameW = 0;
> - LPWSTR volumepathnameW = 0;
> -
> - if (volumename == 0 || volumepathname == 0 || returnlen == 0)
> - {
> -
>    SetLastError(RPC_X_NULL_REF_POINTER);
> -
>    return FALSE;
> - }

It's a common practice to use NULL for pointers, and not 0. Something tells
me that RPC_xxxx error code has nothing to do here, do you have a test case
for that?
> - memset(volumepathname, 0, buflen);

Is there any reason that you zero out the whole 'volumepathname' here?
> - volumepathnameW = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, buflen*sizeof(WCHAR));
> - if (volumepathnameW == 0) return FALSE;

Same for 'volumepathnameW' here.
> - if ((volumenameW = FILE_name_AtoW(volumename, TRUE)) == 0) return FALSE;
> -
> - ret = GetVolumePathNamesForVolumeNameW(volumenameW, volumepathnameW, buflen, returnlen);
> - if (*returnlen > 0)
> -
>    FILE_name_WtoA(volumepathnameW, (*returnlen), volumepathname, (*returnlen));

There is no need to put '*returnlen' in braces. Besides, '*returnlen' may
differ for multibyte and unicode strings.