On Thu, Aug 4, 2016 at 4:29 PM, Yann Ylavic <ylavic....@gmail.com> wrote:
> On Thu, Aug 4, 2016 at 11:10 PM, William A Rowe Jr <wr...@rowe-clan.net> > wrote: > > On Thu, Aug 4, 2016 at 3:52 PM, Yann Ylavic <ylavic....@gmail.com> > wrote: > >> > >> On Thu, Aug 4, 2016 at 9:33 PM, William A Rowe Jr <wr...@rowe-clan.net> > >> wrote: > >> > > >> > It seems correcting the table is the correct way to go, by direct > >> > observation, and placing great faith that other than 0x15/0x37, > >> > the discrepancies between ASCII <> EBCDIC C0 mappings do > >> > not vary widely between EBCDIC mapping choices. > >> > >> Maybe to be sure we could compare the current 'ucharmap' with some > >> result of apr_xlate_conv_byte() for each byte? > > > > > > Perhaps a VALIDATE_TABLE define for the builder, especially when > > --with-maintainer-mode is given? > > Maybe a small temporary main() run once would be enough, it's not as > if it could change anytime soon, no? > Then we may as well build the table at runtime as a register_hooks callback... which is a possibility. The downside is that the table would be more difficult to validate, not less.