Re: Loops (was Re: do { quit; } else { })

On 22.04.2025 00:39, Keith Thompson wrote:

Janis Papanagnou <janis_papanagnou+ng@hotmail.com> writes:

On 21.04.2025 23:21, Keith Thompson wrote:
[...]

C-style for loops have been used successfully for decades, and have
been adopted by other languages (including bash, which isn't
particularly C-like).

>
I have to disagree on that. First I'm positive that Bash adopted
the Ksh loops (but incompletely!), and not the "C" loops.

 
Bash has two kinds of for loops.  From the bash manual:
    for NAME [ [in WORDS ...] ; ] do COMMANDS; done

This is the historic and standard form in all POSIX shells.

and
    for (( EXPR1 ; EXPR2 ; EXPR3 )) [;] do COMMANDS ; done
 
I'm referring to the second form, as I'm sure you know.

Yes, of course.

I see that ksh has something very similar.

Ksh invented it, along with the "arithmetic command" ((...)), as
addition to the (standard) arithmetic substitution $((...)) .

I have no idea which shell
added the second form earlier, but both are clearly derived from C,
directly or indirectly.

Yes. - In the sense of having this three-part separation.

[...]
 

[...]

[...]  But you can still do things that
you can't do with a simple counted for loop, for example printing
powers of 3:
 
    for (( i = 1; i < 1000; i *= 3 )) ; do echo $i ; done

Indeed. Because of the three-part separation of building blocks
you can do similar things as in "C" (modulo FP in Bash).

It's something you cannot do, e.g., with Pascal- or "bart-loops".

Janis