On Thu, 11 Dec 2014 22:48:26 -0600, John McKown wrote: > > ... !!! HOWEVER!!! the character set that the C regexp processor expects >when in TSO is the stupid IBM-037 character set. So the [ and ] at code >points 0xAD and 0xBD are the wrong hex values (the default CCSID is 037 and >so it wants 0xBA and 0xBB for [ and ]). So now the question is "how to I >set the locale to the "C" language locale in normal TSO?" I have the ISPF >terminal set to 3278L1 for the C language. But that is not affecting the C >run time locale. Echoing git: "I hate EBCDIC!". > You and Guido van Rossum and me, all.
>... And I'm not too damn fond >of TSO either. I tried changing the CEEPRM to set the LC_ALL and LANG >environment variables. But that didn't work. > Can the host determine the terminal's CCSID by such as a 3270 WSF Query? It may be as bad as having the terminal, the file being edited, and the EDIT macro in 3 diifferent character sets. Autoconversion? We should be so lucky. I believe UNIX files can be tagged with CCSID; PDS(E)s can't. There some arguments here for the user's keeping data in UNIX files, and for IBM's supporting HFS directories in the SYSEXEC concatenation (it works, sorta, but just not supported.) -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
