Issue 7555 - ddoc whitespace issues due to version tags

[9rnsr]

https://d.puremagic.com/issues/show_bug.cgi?id=7555

[ghost]

Thanks for the fix. LGTM.

[ghost]

Another comma operator found here. Note that I'm mentioning this since it might eventually become an error in D, and we're heading towards DDMD anyway.

[9rnsr]

Ok, fixed.

Honestly I have some negative opinion against disallowing comma operators. I doubt the necessity of the disallowing.

[ghost]

I doubt the necessity of their initial inclusion to any language. What's the point except to confuse programmers?

[9rnsr]

Why comma operator confuse programmers? I don't think so.

[mihails-strasuns]

It may not confuse as-is but it should definitely be reused for something more useful ;)

[ghost]

Why comma operator confuse programmers? I don't think so.

Because I have to look up the spec to see how it's implemented. Specifically the order of operations. And then I have to wonder whether the compiler follows the spec.

At the end of the day you could have just used ;, but wanted to hide two statements into one just to save on bracket count.

Also, wasn't there discussion about how tuple implementations require removing the comma operator?

[ghost]

Think of the comma operator as the is expression. It's a useful feature, but even advanced users may end up having to look up the documentation to remind themselves of its tricky semantics. Unless of course those experts practically implemented the feature. :)

[yebblies]

Honestly I have some negative opinion against disallowing comma operators. I doubt the necessity of the disallowing.

Spot the bug:

foreach(ch; chain(iota('a', 'я'+1)), iota('А','Я'+1))

Sure, foreach could take an AssignExp to prevent this, but there are very few legitimate uses that aren't clearer without the comma.

[yebblies]

Think of the comma operator as the is expression. It's a useful feature, but even advanced users may end up having to look up the documentation to remind themselves of its tricky semantics. Unless of course those experts practically implemented the feature. :)

The is expression is a lot more complex than the comma operator. The comma operator has the same precedence as the comma in function calls, so the problem is usually forgetting it exists rather than now it works.

The is expression also cannot be trivially rewritten in terms of other constructs. Comma can be: e1, e2 -> { e1; return e2; }().

[yebblies]

Aaaaanyway LGTM.

[yebblies]

Auto-merge toggled on