> BTW, shouldn't those tests be expecting a PDF_EBADDATA ?
> They're expecting a PDF_ERROR when they're giving a NULL parameter.
Yes, probably you're right. Time spans are stored as pdf_i64_t
internally, so probably we are facing the same issue here. Those
Time-Module unit tests should expect PDF_EBADDATA, as well as
every unit test testing wrong input arguments (like NULL pointers)
in the methods.
I am preparing a patch fixing that. Then I will apply Gerel's patch
as well.
--
Jose E. Marchesi [email protected]
GNU Project http://www.gnu.org
