RE: Re: GCIC Fullscreen applications - Update
"Rocky" <[email protected]>
| Newsgroups | gmane.comp.emulators.hercules390.vm |
|---|---|
| Message-ID | <[email protected]> |
Joe The mainframe does not do any transalation. All the transaltion is on the PC side --- Any customization would have to be done there and that is out of scope. Thats the Way the real IND$FILE works. IND$FILE receives and sends EBCIDIC from and to the PC. Well - it's almost EBCIDIC - It has its own internal tables but they are fixed and all the PC emulators adjust to that table Actually it's 4 tables - and a Control character in the data string to tell you which table to use. Similar logic to the way the old teletype printers worked. In the case of IND$FILE the the actual character set that is transmitted is only seventy something characters. But the control table tells you which table to look the character up in so in theory you have 70 * 4 possibilites. (not all are used in all tables) But no ASCII. All is EBCIDIC. On the bright side - I added the HEX routine this morning to the EDITOR change function. Less work than I thought it would be So after I looked at the uploaded "C" program and saw that there were x'05' tab characters' I did CH /x'05/ / * Just the way I wrote only 1 apostrophe after the x' Worked fine - and the "C" code is very legible. Holiday here today, but I'll try to put it on the server this afternoon before we leave fo the family dinner. Appreciate if you gys can play aorung with it. Only uploads right now You get tot the upload screen from the FLIST -PF6 - screen and after it's finished you have to refresh the flist to see the new file. I did all my testing with the File Transfer routine in the Emulator that I use. Like to see some testing on other emulators to see if any problems I worked from the X3270 "C" program that I downloaded from their site and reversed most of the logic. But that was only about 30% of the work --- The rest was done with traces of the data stream And that might be emulator dependent to some extent. I'll put post as soon as it on the server Roc _____ From: [email protected] [mailto:[email protected]] On Behalf Of Joe Monk Sent: Wednesday, September 08, 2010 12:57 To: [email protected] Subject: RE: [H390-VM] Re: GCIC Fullscreen applications - Update Rocky, Why don't you make the translation table user overridable. Then they can specify a translation table when file transfer is invoked, that you can apply to the input filestream. Example: ASCII -> EBCDIC EBCDIC -> ASCII -0-1-2-3 -4-5-6-7 -8-9-a-b -c-d-e-f -0-1-2-3 -4-5-6-7 -8-9-a-b -c-d-e-f 0- 00010203 372d2e2f 1605250b 0c0d0e9f 0- 00010203 ec09ca7f e2d2d30b 0c0d0ea9 1- 10111213 b6b53226 18193f27 1c1d1e1f 1- 10111213 efc508cb 1819dcd8 1c1d1e1f 2- 405a7f7b 5b6c507d 4d5d5c4e 6b604b61 2- b7b8b9bb c40a171b cccdcfd0 d1050607 3- f0f1f2f3 f4f5f6f7 f8f97a5e 4c7e6e6f 3- d9da16dd dedfe004 e3e5e9eb b0b19e1a 4- 7cc1c2c3 c4c5c6c7 c8c9d1d2 d3d4d5d6 4- 20c98384 85a0f286 87a49b2e 3c282bb3 5- d7d8d9e2 e3e4e5e6 e7e8e9ba e0bbb06d 5- 26828889 8aa18c8b 8de12124 2a293baa 6- 79818283 84858687 88899192 93949596 6- 2d2fb28e b4b5b68f 80a57c2c 255f3e3f 7- 979899a2 a3a4a5a6 a7a8a9c0 6ad0a107 7- ba90bcbd bef3c0c1 c2603a23 40273d22 8- 68dc5142 43444748 52535457 56586367 8- c3616263 64656667 6869aeaf c6c7c8f1 9- 719c9ecb cccddbdd dfecfc4a b1b23eb4 9- f86a6b6c 6d6e6f70 7172a6a7 91ce920f a- 4555cede 49699a9b ab0f5fb8 b7aa8a8b a- e67e7374 75767778 797aada8 d4d5d6d7 b- 3c3d624f 64656620 21227023 727374be b- 5e9c9dfa 9f1514ac abfc5b5d e4febfe7 c- 76777880 24158c8d 8e410617 28299d2a c- 7b414243 44454647 4849e893 9495a2ed d- 2b2c090a acadaeaf 1b3031fa 1a333435 d- 7d4a4b4c 4d4e4f50 5152ee96 8197a398 e- 36590838 bc39a0bf ca3afe3b 04cfda14 e- 5cf05354 55565758 595afdf5 99f7f6f9 f- e18f4675 fdebeeed 90efb3fb b9eabdff f- 30313233 34353637 3839dbfb 9af4eaff Joe From: [email protected] [mailto:[email protected]] On Behalf Of Rocky Sent: Tuesday, September 07, 2010 11:26 PM To: [email protected] Subject: RE: [H390-VM] Re: GCIC Fullscreen applications - Update Hi Paul - Mr. BFN And everybody else as well. I implemented the CRLF and LF per your suggestion. Works fine. But ... I keep running into new issues - TAB character --- Normal files work fine. However the "C" program that I choose as one of the test cases has lots of TAB x''05' Ebcidic in the data. Our version of IND$FILE routine brings that to the MF as HEX '05' which is proper. But the EDIT routine treats that as a binary non-displayable character. My version of XEDIT works the same way. No tabs. Same for the BROWSE routine. Currenty I cannot do CH /x'05'/ / Forgot or more likely too Lazy and did not add the "x" funtion in the CH command. or the F command I have three options as I see it. Need a suggestion. A) leave the file transfer as is: and fix the change routine - so that it can be fixed via XEDIT change - That is doable. My version of X-EDIT works that way. B) Add a parameter to the FT transfer (on the host FT screen) TAB=nn and then during the transfer convert them (that would change the file from the original) I have neve seen that solution anywhere. Would only affect the HOST SIDE. But I would probably have to eventually fix the F and CH to work with ch/x' ' in any case as well. C) Add a set parameter to the EDIT & Browse function "TAB=n" and just display the Shift with out changing the data. That would be easy for the Browser but extremely complicated for the editior - Would affect SHIFT and TRUNC and all CH Funtions. Even simple line edition would be affected. Since what you see displayed would not actually be what is in the data. Also would have to limit that to variable Files. Very complicated. Easiest is Option "B" but a very unique solution. Next and what I favor is: Option "A" for the editor and Option "C" for the browser and a fix to find hex as well in the browser. Option "C" for the editor is an awful lot of work and not really necessary. Older versions of X-EDIT. all the ones I am familiar with. do not display it correctly either. But they allow OPTION "B". Not familiar with the latest X-EDIT but I think it has an internal translate table and you can define how to DISPLAY every character Suggestion Please Roc _____ From: [email protected] [mailto:[email protected]] On Behalf Of kerravon86 Sent: Monday, September 06, 2010 23:51 To: [email protected] Subject: [H390-VM] Re: GCIC Fullscreen applications - Update --- In [email protected] <mailto:H390-VM%40yahoogroups.com> , "Rocky" <rocsystems@...> wrote: > > Should I use the IBM logic or can I enhnace it by recognizing > an LF or CR as an end of line as well. I'd vote for recognizing both LF and CRLF as line terminators, the only two formats I've ever seen in the real world. In PDPCLIB for MSDOS/Windows I do the same thing, ie a lot of effort to optimize the search for the LF or CRLF terminator. (I used a LF sentinel so that I could just search for LF before checking for CR). BFN. Paul.