Wow, thanks for replying, that is interesting and I'd love to know if you get to any conclusion. Maybe a feature lack on the emulator you are using, the emulator I use doesn't even has IBM-930 so I can't even test it here but comes to show that support for that code page may be lacking overall.
Regards, Leo -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Cameron Conacher Sent: Thursday, November 28, 2019 9:24 AM To: [email protected] Subject: Re: Question - UUID Approach for Mainframes in Japan Hello Leonardo, I am sorry I took so long, but I wanted to get all my ducks in line. The link you provided shows lower case English letters with hex values x'62'-x'69', x'71'-x'78', x'8B', x'9B', x'AB', and x'B3'-x'B9'. I created a small file with all the 256 EBCDIC hex combinations, and browsed it after setting my emulator Host CodePage to IBM-930. (IBM PCOM) All of the above listed hex values show a blank character displayed. When I browse the same file after setting my PCOM emulator Host CodePage to EBCDIC IBM-037, I see lower case English Letters in hex positions x'81'-x'89', x'91'-x'99', and x'A2'-x'A9'. These hex positions are occupied by SBCS Katakana characters in the EBCDIC CodePage 930 display. Then I took my little file, and transformed it from IBM-930 to UTF-8. When I browse that file as a UTF-8 encoded file, I can see lower case English letters. I hope this all makes some sense to you. I tried to be as diligent as possible in gathering the information. What this actually means is a bit beyond me. Characters really are converted as per the chart, but never displayed... hmmmmm. I think that the emulator may be doing something, or perhaps something happens within the 3270 datastream. In any event,my emulator does not render any displayable characters for the hex positions indicated by the lower case English values, as if they do not exist. This is interesting. And I need to do some more thinking on this, but I already have an approach to deal with the UUID as per Charles and Attila. Thanks On Mon, Nov 25, 2019 at 4:05 PM Leonardo Vaz <[email protected]> wrote: > Hello Cameron! > > I see all lowercase letters on > http://demo.icu-project.org/icu-bin/convexp?conv=ibm-930. > > Why do you say there is no support? > > Regards, > Leo > > > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On > Behalf Of Cameron Conacher > Sent: Monday, November 25, 2019 3:43 PM > To: [email protected] > Subject: Question - UUID Approach for Mainframes in Japan > > Hello folks, > I am here with another question today. > We are a large international company with a market presence in Japan. > We store our mainframe EBCDIC data for these markets in EBCDIC CodePage > 930. > This CodePage has no support for lower case English letters. > > If I were a distributed platform and I generated a UTF-8 encoded UUID > value, and sent this value to the mainframe, it would be then transformed > into EBCDIC CodePage 930. > If the UUID were to be generated with any lower case English values ("a", > "b", "c", "d", "e", or "f") I would expect to encounter some issue at > conversion/transformation time, since the underlying EBCDIC CodePage cannot > support the value. > However, if upper case values were sent instead ("A", "B", "C", "D", "E", > or "F"), everything would flow and transform politely. > > So, my question is whether in the Japan world, mainframe application expect > Consumers to send only upper cased values, or if an intermediate step prior > to message transformation occurs close to the mainframe side of things to > force upper casing of the UUID. > Or some other technique? > Similarly, if a UUID were to be sent from the mainframe to the middle tier > somewhere, should I expect that the mainframe would only pas along upper > cased values in the UUID area? > > I believe I can handle things on the mainframe side by transforming the > entire message to UTF-16BE, and then upper casing the UUID, and then > transforming this updated UTF-16BE message area to EBCDIC CodePage 930. > Not sure if this is a "good" way, but it would work. > > Any thoughts? > > Thanks > > .......Cameron > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
