Re: [pdf-devel] Status of PDF_ERROR versus PDF_EBADDATA

[jemarch]

Hi Jeff.

   Nearly all (if not all) of the pdf-types functions return PDF_ERROR if
   response to a required but missing parameter. For example,

       pdf_i64_t a, b;
       pdf_status_t status;
       pdf_i64_add(NULL, a, b, &status)

   would return PDF_ERROR even though PDF_EBADDATA specifies bad
   arguments into a function.

   Other sources, such as pdf-time, usually returns PDF_EBADDATA. Should
   the various functions in pdf-types respond with PDF_EBADDATA in this
   circumstance?

I pushed a patch fixing this, making the bignum functions to return
PDF_EBADDATA if a NULL pointer is passed and to return PDF_EDIVBYZERO
in division and modulus functions.

I also updated the documentation.  It was not up to date anyway.

Thanks.

--
Jose E. Marchesi  <address@hidden>
                  http://www.jemarch.net
GNU Project       http://www.gnu.org