What's wrong with using the VALUE() BIF? -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 עַם יִשְׂרָאֵל חַי נֵ֣צַח יִשְׂרָאֵ֔ל לֹ֥א יְשַׁקֵּ֖ר
________________________________________ From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> on behalf of Schmitt, Michael <michael.schm...@dxc.com> Sent: Thursday, January 9, 2025 5:19 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: ISREDIT CAPS ON OFF External Message: Use Caution I rewrote a 3719 source code development system CLIST in REXX. It was spaghetti code: 308 GOTOs, routines that return to labels set as variables, etc. One particular CLIST construct was trouble. In various places it did this: SET FIRSTLIB = &STR(&NODE2..&&FXNODE3) Or this: SET FIRSTLIB = &STR(&NODE2..&&NODE3) And then later in many places it would just reference that variable, such as: SET FOUNDLIB = &STR(&@UWPFX..&@CHAR1T&CUSTPFX..&FIRSTLIB) So how is &FIRSTLIB resolved? For those that don't know, in CLIST &something is a variable. But &&something is a double variable; it is a variable whose value is a variable. This means that the resolution of the inner variable is deferred until the outer variable is used. So let's say at the time FIRSTLIB was set, &NODE2 was MIKE, &FXNODE3 was IBMMAIN, &NODE3 was ZOS, and it was one of the statements that set it to &&FXNODE3. That means that FIRSTLIB was set to MIKE.&FXNODE3. And then when FOUNDLIB is set, it is fully resolved to MIKE.IBMAIN. Nope, that's wrong. It &FXNODE3 to its *current* value. What it was set before doesn't matter. So let's say the *current* value of &FXNODE3 is JANUARY and &NODE3 is YEAR2024. Then &FIRSTLIB's value is going to be MIKE.JANUARY or MIKE.YEAR2024, but which one it is depends on whether the last assignment to FIRSTLIB was one that used &&FXNODE3 or &NODE3. Yikes! How do you do that in REXX? Keeping in mind that all of these variables are set and used in multiple places. I did solve it, of course. But it was one of the tricker CLIST syntax to convert. -----Original Message----- From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> On Behalf Of Rupert Reynolds Sent: Thursday, January 9, 2025 3:42 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: ISREDIT CAPS ON OFF Back when CLIST was written, CAPS ON was pretty standard for almost everything, apart from SCRIPT docs. Even in the 1980s, writing in mixed case was the sign of rebellion or bravery. I was the rebel, to the surprise of nobody :-) I had to use CLIST for some TEST subcommand stuff, and I remember battling with CLIST all the way, desperately sprinkling &STR() and &NRSTR() almost at random to make it just *Do What I Mean* :-) At least 20 years since my last CLIST, I can't flush this from from my tortured memory, something like PROC 0 DEBUG CONTROL NOCAPS NOCONLIST NOSYMLIST NOLIST IF &STR(&DEBUG) EQ DEBUG CONTROL CONLIST SYMLIST LIST Or was it CONTROL ASIS? (shiver!) When I heard of TSO/E Rexx being available, I was so pleased I forgot to break for lunch. Roops On Thu, 9 Jan 2025 at 19:49, Paul Gilmartin < 0000042bfe9c879d-dmarc-requ...@listserv.ua.edu> wrote: > On Thu, 9 Jan 2025 16:06:22 +0000, Schmitt, Michael wrote: > > . > >CLIST has no syntax for quoted text. It is just text. In this case, CLIST > is passing the string to ISREDIT, and /that's/ where the quoting it, i.e. > it is part of the Edit macro syntax. > > ... > >This lll is why I hate coding in CLIST. It is a language that fights > against the programmer at every opportunity. For example, CLIST is trying > to evaluate everything that's in the quoted string, so if you have & or > something that looks like an operator, it will fail. > > > As you said, "CLIST has no syntax for quoted text." And, apparently, no > way to escape special characters. > > > >I will rewrite the ILISTS as REXX before making changes to them, but that > is easier said than done. CLISTs tend to be spaghetti code. And it doesn't > help that REXX gives uninitialized variables their own name as a value. > > > Use SIGNAL ON NOVALUE to prevent that. It's how I begin every REXX > program. > > My worst case: Once, as an experiment or accidentally I coded > "do" instead of "DO". It failed with a message like: > ILLEGAL STATEMENT: "DO". > > That's wrong two ways: > i) "DO" is not illegal. > i) "DO" is not what I coded. > > CLIST has a mortal terror of lower case. > > -- > gil > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
