Re: base 64 encode/decode test

[Federico Montesino Pouzols]

        It also fails on GNU/Linux. Looking at the numbers down here,
it does not seem an endian issue. I have committed these changes.

On Tue, Jan 07, 2003 at 12:55:10PM -0800, Chad Yates wrote:
> Here is the test program I promised.  On my win32 system and vc++ 6.0 the
> second two tests (testPadding1Short, testPadding2Short) fail unless I change
> part of the b64Decode function like so:
> 
> if (c == '=')
> {
>   if (bits & 0x40000)
>   {
>     if (dstsize < 2) break;
>     *(pdst++) = (bits >> 10);  // changed from '(bits >> 16)'
>     *(pdst++) = (bits >> 2) & 0xff;  // changed from '(bits >> 8) & 0xff'
>     break;
>   }
>   if (bits & 0x1000 && dstsize)
>     *(pdst++) = (bits >> 4);  // changed from '(bits >> 10)'
>   break;
> }
> 
> What I want to know if this is system specific (endian related?) or an
> actual bug.
> 
> thanks,
> 
> Chad


> _______________________________________________
> Bug-commoncpp mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/bug-commoncpp