functional IF

[Warfley]

I've been working on something similar for the SharpBASIC compiler, but instead of using IF in an expression, a LET statement is used allowing for one or more 'on' branches:

Code: Text  [Select][+][-]

1. let s
2.   on x < y is "Foo";
3.   on x > y is "Bar" else "Nothing";
4. end
5. print(s);

It differs from a case/switch block in that it's purely an assignment statement. This construct clearly separates assignment from boolean evaluation.

This wouldn't work for multiple reasons, first, is is already defined as an operator, meaning it can't be used to follow up on an expression. Especially in this case where it would introduce additional ambiguity:

Code: Pascal  [Select][+][-]

1. on MyObject = OtherObject is TMyClass

In this case OtherObject is TMyClass is a perfectly normal expression.
Second in Pascal you can't define variables in the code section meaning such a let construct is also not possible.
Lastly the FPC team does not like the introduction of new keywords, which means adding a new keyword for this is also not an option.

Looking at the keywords available, I don't really see a good syntax option here.

Aside from that, if it's only for assignment it's quite limiting, as one of the main usecases for this is using it in function parameters if you don't want to introduce a new variable. If you just do a grep for IfThen in the Lazarus sources, you'll find that around half the use cases of the current IfThen function is not as an assignment.

[munair]

Second in Pascal you can't define variables in the code section meaning such a let construct is also not possible.

In SharpBASIC definitions are also not allowed in the code section, similar to Pascal, so s must be defined first:

Code: Text  [Select][+][-]

1. dim s:str;
2. dim x, y:int;
3. main do
4.   let s
5.     on x < y is "Foo";
6.     on x > y is "Bar" else "Nothing";
7.   end
8.   print(s);
9. end
10.  

Like I said, it's an assignment, but I get the idea of having something like this implemented as an intrinsic.

[avk]

Delphi ternary operator: https://blogs.embarcadero.com/coming-in-rad-studio-13-a-conditional-ternary-operator-for-the-delphi-language/

[Thaddy]

I understand from the comments on the bugtracker that a commit may be pending.....
https://gitlab.com/freepascal.org/fpc/source/-/merge_requests/1037
The comments seem to suggest that it is by now more than a MR.

[bytebites]

Pascaldragon once implemented it, but it was rejected. Sad for double work.

Principle, which does cause harm is not principle.

[PascalDragon]

Pascaldragon once implemented it, but it was rejected. Sad for double work.

Principle, which does cause harm is not principle.

Nowadays I'd implement as if <expr> then <expr> else <expr> instead of a magical IIF(<expr>, <expr>, <expr>) intrinsic (as I had done back then) anyway.

[jamie]

Still waiting for it! 

Jamie

[marcov]

I liked the IIF() syntax more, but anything is better than yet another redundant syntax, so then I'm fine with functional if then else.

[MarkMLl]

I liked the IIF() syntax more, but anything is better than yet another redundant syntax, so then I'm fine with functional if then else.

Speaking personally, I don't like conflating functions and control structures.

MarkMLl

[Gustavo 'Gus' Carreno]

Hey Y'All,

According to what Michael commented on the MR, since it's being adopted by Delphi, at least the if..then..else is being accepted.

I'm still unsure about the inclusion of the case and the try..except, for which I left a comment for Michael to clarify their inclusion. I'll let him voice his opinion on the MR.

But I would really like that all three would be merged, that's for sure !!

Cheers,
Gus