Not in & is not
lazarus / 64-FPC development

----------------------------------------------------------------------
bytebites · 2025-09-13 18:59

Delphi 13 introduced "not in" and "is not" operators, which are one
more pain for the parser.

----------------------------------------------------------------------
AlexTP · 2025-09-13 21:53

'not in' allows to remove brackets-level: not (a in v_set).
It is good idea.
Same for other one, maybe.

----------------------------------------------------------------------
Bart · 2025-09-13 23:52

What's next: not =?, not <? not is not in?

Bart

----------------------------------------------------------------------
Thaddy · 2025-09-14 05:23

Links: https://docwiki.embarcadero.com/RADStudio/Florence/en/What%27s_
New#Updates_to_the_Delphi_Compiler
https://docwiki.embarcadero.com/RADStudio/Florence/en/Expressions_(Del
phi)#The_is_not_and_not_in_Operators

They also finally copied our {$push}{$pop} with a slightly different
syntax of course....

I like the new option that lets you specify the long name of a
compiler define in {$ifopt}

----------------------------------------------------------------------
AlexTP · 2025-09-14 12:35

Quote from: Bart on September 14, 2025, 01:52:50 amWhat's next: not
=?, not <? not is not in?

'not =' is the same as '<>'.
'not <' is the same as '>='. In these cases 'not' is unneeeded.

----------------------------------------------------------------------
Bart · 2025-09-14 20:09

Quote from: AlexTP on September 14, 2025, 02:35:51 pm'not <' is the
same as '>='. In these cases 'not' is unneeeded.
I was aware of this.
I was just trying to express my dismay about this abomination.

Bart

----------------------------------------------------------------------
Leledumbo · 2025-09-14 21:17

Just a syntactic sugar, has no real value. However, I've been on sugar
free diet for almost 2 years now, never feel better, so I have no
support for the idea. Maybe if someone is interested in implementing,
keep it under {$mode delphi}.

----------------------------------------------------------------------
Warfley · 2025-09-14 21:44

The "is not" is certainly a breaking change, as the following is
completely valid pascal (even though if you ever write something like
this you should probably question your life choices):
Code: Pascal  [Select][+][-]window.onload = function(){var x1 =
document.getElementById("main_content_section"); if (x1) { var x =
document.getElementsByClassName("geshi");for (var i = 0; i type  TTest
=record end; function Test: TTest; begin end; operator not(x:TTest):
TClass;begin  Result:=TStrings;end; begin  if TStringList.Create is
not Test then    WriteLn('Oopsie');end.
The not-in on the other hand would be easily implementable, even
though it would be quite an unecessary change. If it can be overloaded
seperately from the in operator it could be used to implement some
more efficient lookups for hashsets or something something bloom
filter, but outside a few edge cases it wouldn't serve much purpose

Edit: Just fired up D12 and this also compiles as expected:
Code: Pascal  [Select][+][-]window.onload = function(){var x1 =
document.getElementById("main_content_section"); if (x1) { var x =
document.getElementsByClassName("geshi");for (var i = 0; i type  TTest
= record    class operator LogicalNot(x: TTest): TClass;  end; class
operator TTest.LogicalNot(x: TTest): TClass;begin 
Result:=TStrings;end; function Test: TTest; begin end; begin  if
TStringList.Create is not Test then    WriteLn('Oopsie');end.

----------------------------------------------------------------------
LeP · 2025-09-15 09:59

Quote from: Warfley on September 14, 2025, 11:44:58 pmEdit: Just fired
up D12 and this also compiles as expected:
This doesn't work anymore in D13.