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

On 20/04/2025 15:34, Michael S wrote:

On Sun, 20 Apr 2025 14:53:54 +0100
bart <bc@freeuk.com> wrote:
 

On 20/04/2025 12:00, Janis Papanagnou wrote:

>
Exactly 19 days ago I'd even have written such a proposal, and,
fitting to the calendar date, of course I'd have used a distinct
name for the feature; 'for_losers(var,expr,expr[,expr])'.

>
This is quite telling in that:
>
(1) You regard the idea of desiring such a feature as a joke
>
(2) You consider those who'd like to use it as 'losers'
>
Just such a feature seems to be the primary style of 'for' in
languages such as Ada, Python, Modern Fortran, Odin, Ruby, Julia,
Rust, Matlab, Algol68, Euphoria, Ocaml, Logo, Nim, PL/I, Haxe, Lua,
....
>

 In majority of languages in your list 'for' loops iterates through all
elements of collection. If collection happens to be a range
(terminology shared by python and Rust) then 'for'  behaves as a
counting loop. If it isn't then it does not. Even Matlab, despite its
fortranic roots, belongs to that group.

I don't quite get your point. So, given a generic form like:
    for x in A
then some languages, mainly newer ones, will iterate through the values of A if it is not a range. If it is a range, then those values are the interpolated elements of the range.
In other words, in something like:
    for x in 1..10
x will take on the values 1, 2, ... up to 10 (depending on whether the range is closed or open in the language).
Isn't this exactly what some of us are trying to do? So why is it a joke to desire such a nifty form of loop, and why are we losers to do so?
There is nothing to stop any of those languages from also having a feature that does what C-for does, but they choose not to.
However there are also a great number that do have a C-style 'for', often as the primary means of iteration, and there it usually looks completely out of place. (For example, JavaScript.)
This was my opening observation in this sub-thread.

C++ achieve the same objectives [with typical C++ ugliness] by means
of std::foreach.
If I am not mistaken, all exception to that pattern are old languages.
 

>
Even Bash has it. All losers?
>

 Not all, just many.
 

Interestingly, so does BCPL: you may be aware that C was influenced
by B, which itself was supposed to be influenced by BCPL. So what the
hell happened?
>

 So, what happened?

That's what I'm asking! BCPL, an older and more primitive language than the other two, has such a 'for' as described above:
     FOR N = E1 TO E2 DO
It seems to have been dropped by B, which has only 'while', until C introduced a souped-up, 3-element version of 'while', named 'for'.
Clearly the author of both (Ken Thompson) had no great love for 'for'-loops; real ones.