EricWF added a comment.
If we choose to special case this I think we *need* to do it with a new
specialization of `default_delete`, not using the primary template.
The primary templates defines `default_delete::pointer` as `T*[N]`, which
`T*` will not convert to. That means this patch still rej
STL_MSFT resigned from this revision.
STL_MSFT removed a reviewer: STL_MSFT.
STL_MSFT added a comment.
My only concern is that the tests in test/std shouldn't exercise this bug in
the standard, as MSVC's implementation doesn't special-case it. If you'd like
to special-case it in your product cod
EricWF added a comment.
My personal preference would be to diagnose `std::default_delete` and
`std::unique_ptr`, but I'll defer to your judgment here.
An alternative implementation would be to provide a "default_delete"
specialization. (Note that adding this specialization does not change the
mclow.lists updated this revision to Diff 60651.
mclow.lists added a comment.
@EricWF pointed out that this didn't actually suppress the warning that we were
trying to suppress.
All I can say is that it did several weeks ago when I wrote it. Apparently
clang has gotten pickier.
Use tag dispatch
mclow.lists created this revision.
mclow.lists added reviewers: EricWF, STL_MSFT.
mclow.lists added a subscriber: cfe-commits.
There's a bug in the standard, where the default deleter will always call
`delete x`; even if `x` is an array type. This shows up for `shared_ptr`.
Do the right thing,
