On Thu, 2 Dec 2021 13:38:17 -0800, Charles Mills wrote: >Okay, I get it, but this ship is not very ship-shape. > >Seeing as DSN=&MYPRM means what it means, then when IBM introduced variable >symbols they should have used && or % or something, not a single ampersand. >Yeah, yeah, that boat has departed the dock. > Alas, that boat was laden to the gunwales. In 1965(?) I could have coded any arbitrary string, e.g. PARM='FOO&SYSUID.BAR' and expected that literal string to be passed to my PGM. Symbol substitution had breakage: today that means PARM='FOOgilBAR'. Breakage could have been avoided only by using some construct previously not syntactically permitted. There was none such.
>The IEFC657I does not really do the job though, does it? If I have PROC >MYPARM='SYS1.FOO' and mistakenly code DSN=&MYPRM then I will not get an >error on it assuming I have also coded something=&MYPARM elsewhere in the >PROC. Right? > >I think I still say that two wrongs don't make a right. Flagging what (I >say) should be a non-error is not the answer to some other coding mistake. >If someone codes DSN=&MYPRM the error will show up somewhere, either as a >DSN not found or as a "can't catalog a temp DSN" or something. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
