On 2014-09-24 15:29, Bill Godfrey wrote: > On Wed, 24 Sep 2014 07:57:21 -0400, Shmuel Metz (Seymour J.) wrote: > >> on 09/24/2014 >> at 01:14 AM, Paul Gilmartin said: >> >>> My source was IBM-1047; my destination was UTF-8. But Bill Godfrey >>> had the correct explanation >> >> Curiouser and curiouser. RFC 959 does not define 557, nor does it >> I thought implementers are given some latitude to invent new codes and full freedom to define message texts within any severity level
>> require an end-of-line sequence for data in stream mode. >> It's IBMthink. Every file must consist of a sequence of records. But the message should be more explicit -- multiple lines if needed. > On the off chance that this is worth trying to unravel, here is an excerpt > from Paul's post: > Barely worth it -- Shmuel has an incorrigible bad habit of trimming to where the citation is unidentifiable. But I'll trim some more. > (begin quote) > > ftp> quote site sbdataconn=(IBM-1047,UTF-8) > 200-Some characters cannot be translated between UTF-8 and IBM-1047 > FTP is just being cautious here; it can't anticipate whether I intend to do a GET from IBM-1047 to UTF-8 or a PUT from UTF-8 to IBM-1047. > 200 SITE command was accepted > I hate that. It says "200" even for the most invalid SITE command. I've argued with IBM about that. They say it means only that there was no transmission error passing the SITE command, not that it was valid. > ftp> get chars.a2e chars.UTF-8 > local: chars.UTF-8 remote: chars.a2e > 229 Entering Extended Passive Mode (|||11526|) > 125 Sending data set /u/user/chars.a2e > 0 0.00 KiB/s > 557 Data contains codepoints that cannot be translated > > Awww c'mon! what codepoint can possibly have no UTF-8 representation!? > I was able to explain the 451 message as being related to a missing > end-of-line. > > Paul mistakenly thought your quoting his "Awww c'mon" comment was about the > problem with the 451 message, but it wasn't. > > The 557 message is from a different transfer, and has a different > explanation, which is that UTF-8 translation was being attempted with SBCS. > Why is there a distinction between encoding=mbcs and encoding=sbcs? Isn't that implicit in the code pages being used, or is there a code page with a single ID that exists in both SB and MB variants? -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
