AutoHotkey Community

I think not; however, the syntax to force an expression seems more awkward to me than "unclosed" variable derefs. I might suggest a new syntax to force expressions, but I don't have any better idea.

I think a better question is: Why have a binary ? operator? How would it differ from and in v2?
One-line IF is supported, but requires a delimiter to remove ambiguity:

It can be written like this:



I think that in many cases it is more convenient to have multiple explicit output vars, especially where short names like "x" and "y" are common. The code may also be easier to understand for anyone that hasn't memorized the order and names of the output parameters. Something else to keep in mind is that objects aren't supported by the command syntax. Even if and when this is changed, the syntax would be more verbose (e.g. %(x.y) vs %y or %y%).

Since I want to support %(expression) embedded in text anyway, that is a possibility.

Hmm maybe I'm in mistake, but ternary don't need to exist also, we can write this:So it's the "why not" have a binary because we have a ternary operator. Everybody doesn't will look into "logic gate" and "short-circuit" concepts... (as other people who use math operation instead of AND(*) or OR(+) for fun/art code, or other smart form) Sorry to still annoy you

_________________
"You annoy me, therefore I exist."

You're right; the ternary operator doesn't need to exist. It's been redundant from the very beginning.

However, there are important differences between a ? b : c and a && b || c:

• If a evaluates to true and b evaluates to false, a && b || c evaluates both b and c whereas a ? b : c evaluates only b. This would commonly be a problem when the true branch calls a function which has no return value.
• Order of evaluation. For example, a ? b := x : c := y works without parentheses whereas a && b := x || c := y would be evaluated as a && (b := (x || (c := y))).
• Readability. It is easier to see the boundaries of each branch in a ternary operation.

I could rephrase my previous question: Why have ?, and and && all do the same thing? Why not? It would increase code size/complexity and syntax redundancy. Having multiple ways to do something isn't always a good thing. Allowing ? without : may also complicate the rules for order of evaluation.

One final point. For any change, there are always reasons not to make it:

• My time is limited.
• My level of interest is limited.
• Existing users would need to adjust.

If there aren't good reasons to make a change, these reasons not to make it are usually enough.

I think a single ? operator without the following : might be good for conditionally calling functions, it doesn't need to support assignments this way IMO.

I didn't realise && could do the same thing; never mind.

First, thx for your time Lexikos !
I tested more with your explanation, the binary have similar false returned feature. I tested to convert binary and ternary:But I understand you haven't interest in binary opeartor.

_________________
"You annoy me, therefore I exist."

I don't find another binary operator interesting either, if that means anything. I'm interested in unary assignment!

But why? I dont care about the instring but the others? why?

...cuz they are legacy & not needed...

i only like the legacy because they are more logical. Yes the function based way is better. But the legacy way just makes sense.

I don`t understand this. Why does the command / legacy syntax makes more sense to you? The functions are connected with the If-Expression. The expression form is preferred in the future, because functions allow connection in a lot of situations. In contrast every command have to be written on its own line.

In the future with v2, it does not matter if the original feature is a function or command.

:arrow: http://www.autohotkey.com/forum/viewtopic.php?p=480577#480577

_________________
{1:"ahkstdlib", 2:"my libs", 3:"my apps", 4:"my license"}
--> Don't feed the troll! <--

v2.0-a027-7b2f825

• Fixed calling of variadic functions with command syntax.
• Changed GuiSize and GuiDropFiles to use A_EventInfo exclusively (not ErrorLevel).
• Removed If..is/in/contains (to be replaced with operators in a future update).
• Merged bug fixes from v1.1.05.02.

I think this thread has gotten too long to be useful. Anyone have ideas for cleaning it up?

Thanks for sharing your work.
What about to post new alpha releases in the Announcement section?

Im mostly ranting because of change, you can really just ignore what I said. I just dont want to have to update my crap. But i probably will anyway.