I haven't seen that constant either. Google finds me a message for gcc which seems to claim that it's a C99 thing for declaring floating point constants in hexadecimal. I found some discussion by example, which suggests to me that 0x1p63 could be 1E63.
I'm not sure why things get different between 32bit mode and 64bit mode here, but maybe your code never compiled under VC64bit either.
In reply to Re^9: supporting quads on 32 bit Perl
by Corion
in thread supporting quads on 32 bit Perl
by bulk88
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |