v0.7.1

[hsutter]

What's Changed: Highlights

• Add .. syntax to invoke only a member function (no UFCS), e.g., mystring..append("foo")
• Allow x: const = init; and x: * = init; without _ type wildcard
• Allow concatenated string literals, e.g., "Hello " "(name$)\n"
• Faster compile time when doing heavy reflection and code generation
• Ensure definite first use initialization cannot happen inside loops
• Better diagnostics for illegal UFCS calls
• Move-from-last-use only for moveable types
• Remove whitespace sensitivity for * and &
• Allow extra trailing commas in more places (assert)
• Support latest GCC and Clang
• Added -quiet
• Added -cwd

New Contributors: Thanks again, and welcome!

• @dutkalex made their first contribution in Create CITATION.cff #1044
• @matgla made their first contribution in Fixes for arm-none-eabi-g++ #1045
• @alvarogalloc made their first contribution in docs: fix typo #1081
• @sookach made their first contribution in prohibit semicolons in parameter list #1100
• @helmesjo made their first contribution in feat: Add command -cwd (change working directory) #1024

Full Changelog: v0.7.0...v0.7.1

---

This discussion was created from the release v0.7.1.

[TerensTare]

Congrats for the release! Hope we get to see UFCS in C++ as well someday!

Does cppfront use the -> syntax for members of pointers (as in foo->bar(baz))? If not, I think it would be better if the language used -> instead of .. for member-only calls. Confusion is obviously a problem, but I believe it's trivial since the compiler can provide a diagnostic of the form:

Member function `bar` not found. Did you mean to call the UFCS function `bar`: `foo.bar(baz)`?

[hsutter]

Thanks!

Cpp2 does not have -> because dereference is suffix, so *. is the equivalent. I currently have a special diagnostic check for -> to direct folks to *.... the message is error: '->' is not Cpp2 deference syntax - write '*.' instead. Currently -> is used exclusively for return types and I'd like to stick with that so that as much as possible one syntax has one meaning in the language. I chose .. because it's similar to . but I'm open to revisiting that if there's a usability reason to in the future... probably the biggest one would be ranges, if I wanted that syntax for something like 1..10 or first..last, but I think ... will work fine there as ... is already expressing a (variadic) series of things.

[TerensTare]

Makes sense.

My reasons for using -> were purely because it feels more natural (to me foo-> reads exactly as "access something declared within foo") compared to .. (which like you mentioned, makes me think of ranges). Having a second thought about this, something like foo!.bar(args..) seems readable and also align with cppfront's idea of having postfix unary operators (always given that postfix !/!. is not used anywhere else).

[MaxSagebaum]

Hurray. 🎉🎉🎉