Read on Twitter
Getting back up to speed with modern C++ after a couple years writing rust really makes me appreciate what a nicely designed language rust is. Some thoughts as I go…
& #39;At the application level, abstractions relying on such tagged unions are common and useful. The use of “naked” unions is best minimized.& #39;
Strong agreement there, but I only just learned that C++ stdlib *has* a tagged union type (equivalent to rust& #39;s enum).
Strong agreement there, but I only just learned that C++ stdlib *has* a tagged union type (equivalent to rust& #39;s enum).
It& #39;s apparently called "variant", so for the declaration:
variant<Node*,int> v;
you& #39;d check it like:
if (holds_alternative<int>(v)) { … }
which is fine, I guess, but wow: "holds_alternative"? That& #39;s such an awful name.
variant<Node*,int> v;
you& #39;d check it like:
if (holds_alternative<int>(v)) { … }
which is fine, I guess, but wow: "holds_alternative"? That& #39;s such an awful name.
Making enumerations have significantly different semantics with type safety according to whether they& #39;re declared with `enum` or `enum class` seems like a tidy encapsulation of what I find most troubling abut C++.
Extreme lol at the concept of proper modules being introduced in C++20.
Then again, better late than never.
Then again, better late than never.
A choice quote in the chapter on error handling and exceptions:
"In well-designed code try-blocks are rare."
I agree, but probably not for the reason the author thinks.
"In well-designed code try-blocks are rare."
I agree, but probably not for the reason the author thinks.
Reading about C++& #39;s function overloading and default arguments reminds me how little I& #39;ve missed those features in rust. Why do they seem necessary in languages like C++ and Java?
"The return type of a function can be deduced from its return value"
I had recently learned of the addition of `auto` for variable declarations; didn& #39;t think about using it for functions. This seems like maybe not such a good idea, though…
I had recently learned of the addition of `auto` for variable declarations; didn& #39;t think about using it for functions. This seems like maybe not such a good idea, though…
"…should be used carefully because a deduced type does not offer a stable interface: a change to the implementation of the function (or lambda) can change the type"
Ahhhh, thats the C++ I expect.
Ahhhh, thats the C++ I expect.