Re^3: Passing undef to XSUB

Would int* also work?

No, but int would:

int
foo(ap,bp)
    int ap
    int bp
    PREINIT:
      int a, b;
    CODE:
      a = ap ? ap : -1;
      b = bp ? bp : -1;
      RETVAL = a * b;
    OUTPUT:
      RETVAL
[download]

Though, if warnings are enabled, that produces a "Use of uninitialized value in subroutine entry..." for each undef that's passed to it - no such warning with educated_foo's rendition.

Cheers,
Rob

Comment on Re^3: Passing undef to XSUB Select or Download Code

Replies are listed 'Best First'.
Re^4: Passing undef to XSUB by ikegami (Patriarch) on May 08, 2008 at 05:00 UTC
It (presumably) compiles and runs, but I wouldn't say it works. Not only does it introduce a warning, it gives incorrect results for arguments equal to zero.	[reply]
Re^5: Passing undef to XSUB by syphilis (Archbishop) on May 08, 2008 at 07:32 UTC
it gives incorrect results for arguments equal to zero Yes ... I had missed the important point that 0 ought to be evaluated as 0 (and not -1). Now I understand what your speculative code was driving at. Cheers, Rob	[reply]
Re^4: Passing undef to XSUB by kyle (Abbot) on May 08, 2008 at 04:39 UTC
I don't know much about XS, but I wonder if this code is able to distinguish between when it's passed undef and when it's passed `0`. That is, can it see the difference between `foo( undef, undef )` and `foo( 0, 0 )`? It seems to me the "uninitialized value" warning is `undef` getting turned into `0`.	[reply] [d/l] [select]
Re^5: Passing undef to XSUB by syphilis (Archbishop) on May 08, 2008 at 07:57 UTC
It seems to me the "uninitialized value" warning is undef getting turned into 0 Yes, I think that's right: `use warnings; use Inline C => Config => BUILD_NOISY => 1; use Inline C => <<'EOC'; void foo(int a) {} void bar(SV * a) {} void baz(SV * a) { int i = (int)SvIV(a); } EOC $u = undef; foo($u); # warning bar($u); # no warning baz($u); # warning` [download] Cheers, Rob	[reply] [d/l]