Manu Evans comments

Results 325 comments of


                                            Manu Evans

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

My imagine just doesn't light up at all with respect to the concern you raise. I can _kinda_ see your argument to prevent it applied to pointers and I'm sympathetic...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

> [...] then removing that allowance will break legitimate code. > > If, on the other hand, we disallow this kind of code at the beginning, then no valid code...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

I wouldn't say I'm bent out of shape, it was more of a meta-commentary on a broad issue that influences decision making on a detail like this. @WalterBright are you...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

An alternative rule might be that the target type has the same `.sizeof` as the source type... I think that could catch a similar but broader set of mistakes. Rather...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

Yes, like I said; "I think that could catch a similar but broader set of mistakes". I realise that `int` was arbitrary, but that demonstrates exactly why this suggestion works...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

> Yeah, that's a good usage. I tried seeing how to do it without reinterpret cast, and it's not as nice. > > ```d > myPointer = cast(typeof(myPointer))(cast(size_t)myPointer | bitStuffingFlags);...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

I think it's fair to expect people have at least basic knowledge of the language... D isn't meant to be javascript. The whole point of D from my perspective is...

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

I guess this will sit here forever? Should I click the merge button?

fix #20644 `cast(ref T)` as shorthand for `cast(T)&`

Oh dear. What a waste of everyone's time. It will just say exactly what he already wrote in the spec PR...

Will RTU over TCP be included?

I just came here intending to log this bug. I use ethernet serial servers, so I need RTU over TCP, which isn't an option here it seems...?