On 2015-09-23 16:11, Charles Mills wrote:
>
> 5. Yes, I find make $( ls *.[cC] | sed 's/\.[cC]$/.o/' ) to be scary.
> Actually, I can sort of parse it: list every file ending in .c or .C and pipe
> it into sed and then do something with it.
>
Pretty much so; for illustration:
user@OS/390.24.00: ( set -x; ls -1; make $( ls *.[cC] | sed 's/\.[cC]$/.o/'
) )
With input data:
-sh:0+ ls -1
bar.C
foo.c
as you inferred, list all the source files:
-sh:0+ ls bar.C foo.c
"sed" changes any ".c" or ".C" suffix to ".o"
-sh:0+ sed s/\.[cC]$/.o/
and the "make $( ... )" uses the output of "sed" as parameters to "make".
-sh:0+ make bar.o foo.o
> 6. My use of JCL versus make has nothing to do with efficiency. I would
> assume they are equivalent in that regard. It has entirely to do with
> inertia. Been doin' compiles with JCL for nigh on forty years now. If JCL was
> good enough for me pappy it's good enough for me. JCL *is* a language for big
> room-sized computers with blinking lights. You got a problem with that? <g>
>
My wonderment at JCL is how it can use so many words to
accomplish so little.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
