test: avoid unnamed parameter in function type template argument #1291
Conversation
| pure2-last-use.cpp2:279:272: error: no member named 'move_only_function' in namespace 'std' | ||
| requires (std::is_convertible_v<CPP2_TYPEOF(f_), std::add_const_t<std::add_pointer_t<int()>>&> && std::is_convertible_v<CPP2_TYPEOF(g_), std::add_const_t<std::add_pointer_t<int(cpp2::impl::in<int> in_)>>&> && std::is_convertible_v<CPP2_TYPEOF(mf_), std::add_const_t<std::move_only_function<int()>>&> && std::is_convertible_v<CPP2_TYPEOF(mg_), std::add_const_t<std::move_only_function<int(cpp2::impl::in<int> in_)>>&>) | ||
| requires (std::is_convertible_v<CPP2_TYPEOF(f_), std::add_const_t<std::add_pointer_t<int()>>&> && std::is_convertible_v<CPP2_TYPEOF(g_), std::add_const_t<std::add_pointer_t<int(cpp2::impl::in<int> inx)>>&> && std::is_convertible_v<CPP2_TYPEOF(mf_), std::add_const_t<std::move_only_function<int()>>&> && std::is_convertible_v<CPP2_TYPEOF(mg_), std::add_const_t<std::move_only_function<int(cpp2::impl::in<int> inx)>>&>) |
There was a problem hiding this comment.
This seems to be one of the lowering bugs mentioned at 212eb15#commitcomment-147040717.
| 2137 | return false; | ||
| | ^ | ||
| 2137 | template<typename T, specialization_of_template<std::optional> X> | ||
| | ^~~~~~~ |
There was a problem hiding this comment.
Some GCC 10 version doesn't output these ~ squiggles.
Maybe @hsutter's local version or a CI runner's.
In the main branch's commit c5feb42,
they are again removed: c5feb42#diff-aa97a613d9dea020a78ec4046393f329f881475e3d67f8796e6e8afec5849968R4.
|
Thanks! There are merge conflicts so I won't try to merge this for 0.8. As described in #1322, that means I need a new one-time CLA that will cover all your future contributions. I've emailed you the new CLA, and one it's completed I can look at this PR. |
| @@ -1,12 +1,21 @@ | |||
| pure2-last-use.cpp2:273:36: error: expected variable name or 'this' in lambda capture list | |||
There was a problem hiding this comment.
Thanks! Regressions weren't re-run so I did that and committed the results.
Note these Clang-12 results in particular changed, can you check the diffs and see if they're benign?
There was a problem hiding this comment.
Yes, the change looks right.
Resolves 212eb15#commitcomment-147400224
by avoiding the unnamed parameters,
rather than changing the lowering to Cpp1.