On Thu, 7 Mar 2019 13:03:38 -0800 Ron Norman <rjn@xxxxxxxxxxxx> wrote: > making such massive changes before the code merge happens would create > a problem getting the merge completed. Agreed. > Its not a good idea for the code to assume the character set used is > ASCII as in rare cases, it could be EBCDIC. I'm not sure what that implies about *p < ' ' > Likewise, assuming a 32/64 bit platform may not be good either, as > Unisys 2200 and Honeywell 6000 systems have 36/72 bit words... Do we compile on such machines? There are definitely byte-swapping macros that depend on 32/64 bit integers. --jkl
