A small inconsistency

[egsuh]

With for...in loop,

Code: Pascal  [Select][+][-]

1. var
2.     idx: array[0..10] of integer;
3.     i: integer;
4.  
5.     cfs : array[0..10] of TStringList;
6.     acfs: TStringList;
7.  
8. begin
9.     for i in idx do i := 0;   // this raises compile error.  resetting "for loop" index
10.     for acfs in cfs do acfs.free;   // this seems to work fine.
11. end;

The i in the previous case is not a loop index in practice. But only reading is possible, and writing is not possible. How do you think?

[marcov]

Yes, it is weird, but consistent. Note that in the second case you don't change the loop variable (the class reference, which is a pointer), but call a method on it, which is allowed. Secondary effects of that (freeing the class, turning the reference stale) can't be modelled by for ... in.

[gues1]

Look here: https://forum.lazarus.freepascal.org/index.php/topic,71721.msg560420.html#msg560420

[Thaddy]

Yes, finally you understood it. (or still not?)

[gues1]

Yes, finally you understood it. (or still not?)

It's for me ?

If it is, I understand like all other things, but may be I'm not "inline" with you.
This doesn't mean your ideas or solutions aren't good.
When they're worthwhile, I'm happy to share them.