Re: HvMROMETA

: I think the best course of action to remedy this is to
treat it like
: HvAUX() (that is, add a comment in hv.h that it's the
responsibility
: of the caller to verify SvOOK before calling).  I looked
around and
: most places where it's used it's already being verified by
the calling
: code (usually indirectly via HvNAME_get).  I'll whip up a
patch that
: fixes the spots that aren't checking and takes the
SvOOK/NULL thing
: out of the macro, unless someone has a better idea.

Sounds like a plan to me.

On 5/26/07, Jarkko Hietaniemi <jhiiki.fi> wrote:
> : I think the best course of action to remedy this is
to treat it like
> : HvAUX() (that is, add a comment in hv.h that it's the
responsibility
> : of the caller to verify SvOOK before calling).  I
looked around and
> : most places where it's used it's already being
verified by the calling
> : code (usually indirectly via HvNAME_get).  I'll whip
up a patch that
> : fixes the spots that aren't checking and takes the
SvOOK/NULL thing
> : out of the macro, unless someone has a better idea.
>
> Sounds like a plan to me.

Attached is a patch for this.

-- Brandon