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

bart <bc@freeuk.com> writes:

On 21/04/2025 22:06, Waldek Hebisch wrote:

bart <bc@freeuk.com> wrote:

>
t iteration goes over all elements in the hash table.

  BTW2: When looking at 'for' loop you are supposed to see pattern,
without need to track all steps.

>
That's one of the disadvantages of using the same, often
/inappropriate/ keyword for every kind of pattern.

You think it's inappropriate.

Would your objections go away if a different keyword were used?

This is the original C again:
>
    for(p=sqliteHashFirst(&pSchema->trigHash); p; p=sqliteHashNext(p)){
        sqlite3DeleteTrigger(db, (Trigger*)sqliteHashData(p));
    }
>
This is the version that my visualisation tool produced (where C
source is visualised in my syntax); posted earlier but it's worth
posting again:
>
    p := pSchema^.trigHash.first
    while p do
        sqlite3DeleteTrigger(db, ref Trigger(p^.data))
        p := p^.next
    od
>
While the C version looks intimidating, this looks much cleaner and
simpler (expanding those macros helped).

*You* find it much cleaner and simpler.  I don't.  What makes you right
and everyone else wrong?

Just once, can you acknowledge that others have different opinions than
you do?  I don't recall that you've ever done that (while I've done it
repeatedly).

The intent of 'for' is to iterate over some collection.  Each of
A, B, C is needed to know the collection.

>
Sure, but ALL ON THE SAME LINE? All within the same parentheses?
Putting multiple things on the same line is usually frowned upon. Look
at these ludicrous examples:
>
  for(p=pDel; p && (p->pLeft||p->pRight); p=(p->pLeft ? p->pLeft :
  p->pRight)){
>
  for (yoffset = coef->MCU_vert_offset; yoffset <
  coef->MCU_rows_per_iMCU_row; yoffset++) {
>
  for (bi = 0; bi < num_blocks; bi++, start_col +=
  compptr->DCT_h_scaled_size) {....}
>
I /think/ the last two are simple iterations!

I dislike the way the lines are split in the middle of nested
constructs.  I prefer the decision of where to split a long line
to reflect the structure of the expression or other construct.

All three would be IMHO clearer if each of the three clauses were on a
separate line.  The fact that you can write a badly formatted C-style
for loop is not an argument against C-style for loops.  "Doctor, it
hurts when I do this."

But others might just ignore the line break, reading the two lines as
one logical "line" and not caring where the line is broken.  Different
people perceive things differently, and have different preferences.

Notice that that last bit of the last example is nothing to do with
the mechanics of the iteration; it's to do with the body.
>

Especially A, which is only executed once, after which it's no longer
part of the loop.

Without A you do not know range of iteration.

>
I've seen plenty of for-loops that start with 'for (;...)'; that
doesn't seem to worry anyone! Which is not surprising; nobody seems to
care anything here.

"nobody seems to care anything here".

That's just bullshit, and I think you know it.

[...]

There is no excuse.

There doesn't need to be an "excuse".

[...]

And yet, even within the limitations of C, people favour writing
appalling code, when it could be much cleaner. So (1) they don't care;
and (2) try to justify that bad code as being better.

Are you under the impression that "appalling" is an objective term?

People write code that you don't like.

Here's that visualisation example above, converted into C:
>
    p = pSchema->trigHash.first;
    while (p) {
        sqlite3DeleteTrigger(db, ref Trigger(p->data));
        p = p->next;
    }
>
Compare with the original C at the top.

I have no particular problem with either.  The original C doesn't have
as much whitespace as I'd like, but that's not a huge deal.

But no, my opinion is invalid, and only yours is correct.  (That was
sarcasm.)

--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
void Void(void) { Void(); } /* The recursive call of the void */