On Saturday 13 June 2009 08:34:49 pm Brad Hards wrote:
> pdf_time_span_negate (pdf_time_span_t *p_span)
> {
> pdf_status_t p_status = PDF_OK;
> + if ( p_span == NULL )
> + {
> + return PDF_ERROR;
> + }
> pdf_i64_neg(p_span, *p_span, &p_status);
> - return PDF_OK;
> + return p_status;
> }
On further thought, whether that is the right fix depends on whether
pdf_i64_neg() should assume that its argument might be NULL (and therefore
check) or not.
This query is generally applicable - should the pdf_types.h operations protect
against NULL pointers, or should they assume that their arguments are already
protected by higher level functions?
Brad
