fix: lower (nested) _braced-init-list_ (argument) #927
Conversation
| assert(board[2] == :std::array<u8, 3> = ('X', 'O', 'O')); | ||
|
|
||
| // Still parentheses | ||
| assert((:std::vector = (17, 29)).size() == 2); |
There was a problem hiding this comment.
For some reason, a semicolon doesn't work to delimit the unnamed declaration:
| assert((:std::vector = (17, 29)).size() == 2); | |
| assert(:std::vector = (17, 29);.size() == 2); |
Also, I think (0).f() should lower to CPP2_UFCS(f)({0}) (despite #542 (comment)),
just like f((0)) lowers to f({0}).
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
|
Commit 94bea67 did this for the empty list |
Since now we might want to write flags afterwards, e.g., `<Group, flag1, flag2>`, programmers who want the default group `<Default, flag1, flag2>` might expect to be able to equivalently write `<_, flag1, flag2>`, so let's make that work Until now, programmers would just omit `<Default>` and that continues the work as the basic default, as before
|
Thanks and sorry for the lag! This seems like it could be pri-3 (per #1287) if it were rebased to current |
hsutter
added
the
question - further information requested
JohelEGP
force-pushed
the
braced-init-list
branch
from
28ae3ef to
774b373
Compare
|
Rebased. |
| mystruct::mystruct(auto&& i_, auto&& j_, auto&& k_) | ||
| requires (std::is_convertible_v<CPP2_TYPEOF(i_), std::add_const_t<cpp2::i32>&> && std::is_convertible_v<CPP2_TYPEOF(j_), std::add_const_t<std::string>&> && std::is_convertible_v<CPP2_TYPEOF(k_), std::add_const_t<cpp2::u64>&>) | ||
| : base{ { 1 } } | ||
| : base{ 1 } |
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
There was a problem hiding this comment.
Nah, it must be a bug that I don't omit the braces when the context already implies some.
t: @struct type = {
this: std::vector<int> = (0, 1);
v: i32;
}
lowers to
: std::vector<int>{ { 01 } }(: std::vector<int>{ (01) } in the main branch: https://cpp2.godbolt.org/z/srM3vvscM).
There was a problem hiding this comment.
The problem is that emit_special_member_function centralizes the lowering of Cpp1 constructors and assignment operators.
#844 (comment) lists some of the bugs from that.
Maybe one of those PRs fixes this particular issue, so I'm leaving it as-is.
2 participants
Resolves #1343.
Resolves #1283.
Resolves #1037.
Resolves #568.
Resolves #542.
The
mainbranch already treats a return type contextually,only lowering a braced-init-list if it can work: https://cpp2.godbolt.org/z/sonKE4zan.
This PR recognizes two other places where an expression list can be an initializer to lower as a braced-init-list: