- User Since
- Jun 26 2014, 12:44 PM (251 w, 1 d)
Wed, Apr 17
It may start being reported as a leak by tooling, in the same way static T* t = new T would, which is the other idiomatic way to do this.
Tue, Apr 16
I don't think this is correct.
I think this is fine.
Mon, Apr 15
We should apply this to both versions of __do_compare_assert and to the call operator taking non-const arguments as well.
Sorry I didn't get to this sooner.
Sun, Apr 14
This doesn't work for a number of reasons. First, the destructor is only trivial when all the union members are. second, you can't put constexpr on functions that never produce a valid constant expression
Sat, Apr 13
LGTM. This fix is correct given the current state of things.
Thu, Apr 11
Hmm. I thought I had fixed this using perfect forwarding, but I guess not. My bad.
It took me a couple of passes to spot it.
Wed, Apr 10
The fix LGTM. I'll try to figure out what's going on with the tests
So _symbol_difference is used to check for added or removed symbols. changed_symbols is the bit of logic that detects if the properties of the symbol changed.
I take back my comments about libc++ style. I didn't understand what was going on here. my apologies.
Theres more to the ARM config. The macro should reflect were implementing the ARM C++ ABI and not the plain itanium one.
Tue, Apr 9
There's a circular dependency between libc++ and libc++abi, which causes normal CMake dependencies not to work for PSTL. Specifically, libc++abi will complain about not having access to pstl/internal/XXX when building.
How does this include pattern work with modules? It's important the libc++ headers maintain their modularity.
I don't really understand how a lot of the changes in this patch implement the changes in the paper. Can you explain the rationale for each change?
I don't believe this patch implements the perfect forwarding call wrapper requirements laid out in the paper. Take a look at the not_fn implementation as an example of this (perhaps even generalize it's implementation to be used here as well).
This patch isn't correct. The is_flexable_invocable trait is an ODR violation. And I don't think the bug this is solution to PR41360.
Every line of this change should be covered by a test
Mon, Apr 8
Sun, Apr 7
Fri, Apr 5
I landed these changes as a series of smaller cleanups.
Thu, Apr 4
Wed, Apr 3
- Revert use of builtins.
- Fix configuration error.
I've come around to Howards position that this is a beneficial extension, and so I'm abandoning this change.
When LIT says "N warnings in tests" It's not talking about warnings compiling code; it's refering to warnings while configuring the test suite.
Tue, Apr 2
In future, you don't need review before adding or improving tests.
Mon, Apr 1
Address review comments.
- Convert the tests to ASSERT_SAME_TYPE
- Fix layout incongruity for non-apple platforms.
- Various cleanups.