|
Post by pepe on Dec 2, 2020 13:35:44 GMT -5
I am trying to use an UATO hilite file for the JCLs, anybody has a JCL.AUTO file?
Thanks in advance.
|
|
|
Post by George on Dec 2, 2020 15:04:08 GMT -5
I looked in the distribution stuff and didn't see the JCL.AUTO file. I will add it certainly, but a browse seems to not have that many keywords, maybe I'll have a look at whether it needs a refresh.
But it's also in the old AUTO format, again time for a refresh.
I'll have a look Robert.
Just had a look, Wow, are there a lot of new parameters since either of us coded any JCL.
George
|
|
|
Post by George on Dec 2, 2020 16:44:35 GMT -5
If there were a list somewhere of all the valid keywords, whipping up an AUTO file wouldn't take long. But a quick search on the web doesn't find anything (at least for me).
If anyone can find one, or even find a syntax highlighter file for some other editor, we can massage it into AUTO format.
Barring that, it will take some enterprising user to come up with one.
George
|
|
|
Post by Stefan on Dec 2, 2020 17:28:40 GMT -5
I don't know how anyone writes this stuff any more, it's mind blowingly complex. LOL... The same way we used to.... Assuming you can't copy it from someone else's job, you write it once and then copy & adapt it for ever more!
I was fortunate to first encounter JCL when I had a year's sabbatical at IBM in '79 on OS/VS2. They sent me on a proper course for a week for JCL and then another week for 'utilities' (IEBCOPY, IEHPROGM, IEBGENER, etc.)
Afterwards, I think I was the only one in the office who knew that IEBGENER's SYSIN didn't always have to be DD DUMMY!
|
|
|
Post by Stefan on Dec 3, 2020 4:43:15 GMT -5
You could just cut/paste the JCL Manual's Table of Contents into a file and the massage it with SPFLite. That will get you a long way there.
I suppose the (for me) new style IF THEN ELSE stuff could be picked out from the humdrum JCL keywords. And you could pick out JESn statements, etc
Robert is right - there's plenty of opportunity for colourisation to get it wrong.
I guess if you don't specify '/' as a delimiter, you could avoid colourising DD names and JOB names as they would appear to start with // and hence //DD wouldn't match what's in the AUTO file. But I don't recall if '/' would be needed as a delimiter elsewhere in JCL syntax.
Were dataset names specified in quotes? If not, that's another possible colourisation issue in the making. Leaving out '.' from the delimiter list may help.
Same goes for all other literals, like the Account code field on the JOB statement, etc
Nostalgia!
Now I'm almost wishing I'd kept some JCL. I retained Assembler code, CLISTs and ISPF/PDF extensions I wrote during my time, but no JCL. But my main regret is not keeping the PL1 source listing of our (then epic) Startrek adventure program used during lunchtimes at IBM UK.
Augmented by a couple of assembler modules to handle 3270 input/output (TGET/TPUT), it's amazing how much addicting fun you can have with TSO and a 3270 VDU!
|
|
|
Post by nicc on Dec 3, 2020 6:38:59 GMT -5
According to my v2r3 JCL Language Reference there are only 20 JCL statements then some JECL statements. All the other gubbins are parameters - even if they now stand alone whereas, before, they were sub-parameters eg DCB=(LRECL=...,RECFM=...,BLKSIZE=...) can now be coded as LRECL=...,RECFM=...,BLKSIZE=... - no need for DCB=().
JCL writing is an art and is easier than program writing - just like programming you only need a subset of the language. I copy my programs and edit them to make new programs but I have read the PL/1 and Rexx Language references from front to back (mostly!). Now I am creating a Rexx program to write my the basic structural requirements for my new Rexx programs - it's a simple 'include' processor, reads a skeleton and outputs something that just needs a few tweaks and the business logic added. Keeps me off the street corners!
|
|
|
Post by mueh on Dec 3, 2020 9:24:47 GMT -5
George : There is already a zjcl.auto member in SPFLiteColorizeFiles . Comment still shows COBOL and it doesn't contain UNIT VOL LABEL colorization . Here a Picture of changes i made with Scheme display and result . ( PROC PEND UNIT as a Sample ) Thanks
|
|
|
Post by George on Dec 3, 2020 12:37:38 GMT -5
Well if you guys thought JCL was a nightmare, when I worked for a "Service Bureau" as they used to be called, I wrote a tool for updating JCL, it broke down the raw JCL into pieces, provided a macro language to allow altering all the pieces (like reformatting DSNAMES) and then re-build the result all back into working JCL. We needed this to migrate new customer's JCL into our shared processing environment where multiple users all ran on the same systems and were isolated from each other only by strict naming conventions.
The tool finally died when we could no longer keep up with the constant additions to JCL that IBM was making.
Back then I knew JCL intimately.
George
|
|
|
Post by George on Dec 3, 2020 15:24:09 GMT -5
Did a very quick 'massage' of the contents of the latest JCL manual. Sure glad I'm not writing JCL anymore, looks like a real zoo. No warranty, expressed etc. etc. George zjcl.auto (5.75 KB)
|
|
|
Post by Clive L on Jul 9, 2021 11:55:02 GMT -5
Yeah, I suspect that what you did originally was mostly for MVS/SP and just some initial z/OS stuff but not all of it. The z/OS JCL manual is gigantic. I don't know how anyone writes this stuff any more, it's mind blowingly complex. Just as a (very late) additional comment on this; if you think JCL has changed a lot, try looking at Principles of Operation and compare it to the old days. I've recently been very boggled at having to do this for real!
|
|