Re: Checking the loop variable after the loop has ended (Was: Loops (was Re: do { quit; } else { }))

On 18/04/2025 17:46, Janis Papanagnou wrote:

On 18.04.2025 17:57, bart wrote:

On 18/04/2025 16:24, Kaz Kylheku wrote:

On 2025-04-17, Lew Pitcher <lew.pitcher@digitalfreehold.ca> wrote:

[...]

[...]
    for (n = NUM_ENTRIES - 1; ...

>
That looks like another inconsistency in the syntax (and another error
opportunity).
>
Since the loop for upward and downward iterations between 0 and N-1
inclusve would be:
>
    for (i = 0;   i <  N; ++i)
    for (i = N-1; i >= 0; --i)
>
'N-1' is what I would have to write; it's odd that C people have write
it too. Unless perhaps you do this:
>
    for (i = N; i-- > 0;)
>
Here you now have a genuine 2-part loop!
>
I guess this is C being 'flexible', in being able to reinvent for-loops
each time.

 It is perfectly fine to write your loops in "C" as
      for (i = 1; i <= N; ++i)
     for (i = N; i >= 1; --i)
 (and access your array elements as [i-1] if you wish).

That's never going to happen. Only, possibly, if you have to port a 1-based algorithm to C, without risking the bugs that will creep in if you try and convert it to 0-based.
Sure, with 1-based, upward and downward counting loops can be more symmetrical.

 You should be aware that the problem you have here is *not* a "C"
"loop-problem", it's a problem of the low-level _"C" arrays_ you
have to use; they force you to define and use bounds of 0 and N-1
respectively if you want to use an array of N elements.

I can use 0-base arrays in my language, and yes, for loops to iterate over bounds could involve 0 and N-1:
     for i 0 .. N-1 do
but they can also look like this:
    for i in A.lwb .. A.upb do
or just:
    for i in A.bounds do
    for i, x in A do                # iterate over bounds and values
'0' and 'N-1' or 'A.len-1', don't appear at all.
So while 0-based can be more troublesome, it can still be a language problem rather a 0-based one.