- User Since
- Jun 26 2014, 12:44 PM (185 w, 6 d)
Is this really a "fix" or does it just hide the ODR issue?
@bcain Can we land it and then revert it once Clang decides what it's up to? That way we can get the 32bit builder green?
LGTM with the change mentioned by @lebedev.ri .
Fri, Jan 12
- Introduce SourceLocExprScope.h to help reduce code duplication.
- Merge with upstream.
- Address @rsmith's comments by removing the fallback implementation of the diagnostics.
- Improve diagnostic handling.
- Address inline comments about missing diagnostics for void pointers and function pointers.
- Address inline comments about only enabling when -fstrict-vtable-pointers is specified, and only on types with vtables.
Thu, Jan 11
- Address inline comments.
Is this behavior specified somewhere? Or are we simply adding an extension to Clang? If so I would really prefer to add my co_promise solution (but I need to write a paper in favor of it first).
LGTM other than inline comments.
Address inline comments.
@rsmith The trait is already documented in LanguageExtensions.rst, so I'll assumed you missed that and are not trying to suggest more documentation is needed.
This LGTM minus nits.
Mon, Jan 8
LGTM. Honestly, I'm surprised we don't already have a way to configure like this.
Sun, Jan 7
Dec 13 2017
Nov 25 2017
Nov 22 2017
Nov 21 2017
- Address spelling errors pointed out during review.
Nov 20 2017
This LGTM. I would love if another party interested in Windows could review it though.
Also, you're lacking tests for noexcept and the deduction guide.
I would much rather take this change in all dialects except for C++03.
- Improve quality of tests.
- Format code.
Nov 19 2017
Initially this LGTM. I'll take another pass in the morning.
- Fix argument initialization.
- Make constexpr.
Nov 18 2017
- Address inline comments.
Accepting for post-commit review. I don't want to be carrying this patch locally.
I've made an attempt to add __builtin_launder to clang in D40218.
- Use style suggested by @mgorny.
- Remove incorrect FIXME comment.
I think we may want a __launder function that we can use internally in all dialects.
Nov 14 2017
Nov 7 2017
Nov 6 2017
@K-ballo I always forget. Do you have commit access?
Are the changes that actually use this feature going to be upstreamed as well? Or are they specific to an internal build? It's never great to introduce seemingly dead or unused code.
I think I would prefer if this patch "left an explicit trace" as to the reasons for all the (void) casts. Specifically I think a macro like TEST_IGNORE_DISCARD <expr> might be nice.
LGTM minus inline comments.
Nov 3 2017
Oct 30 2017
We should backport this, right @mclow.lists?