Rebased and adjusted to upstream changes.

In D150264#4380575, @danlark wrote:

I don't have commit rights, can you help me submit the change?

Danila Kutenin
kutdanila@yandex.ru

LGTM after addressing the review comments.

LGTM!

I didn't do an in depth review, I mainly looked at the code out of curiosity.

LGTM modulo 2 nits.

Rebased and address review comments.

Thanks for the review!

In D151814#4390263, @Arthapz wrote:

Hello, i'll add the support of this patch in XMake today :)

In general the approach seems fine, but I want to do some testing with this patch.
Especially to see what happens when the regenerate_expected_results flag is toggled.

Thanks for the review!

Thanks for working on these improvements! LGTM!

LGTM!

Now with approval selected ;-)

LGTM!

In general I'm happy, but I would like to see whether custom_container_iterator can be removed.

Could you add some information in the commit message why this is a good idea?

In D151654#4385137, @ldionne wrote:

In D151654#4384747, @Mordante wrote:

Did we add documentation for this new feature?
LGTM!

It is documented in format.py like we do for other types of tests.

LGTM, thanks for fixing this.

I mainly glossed over it. I'm not to familiar with the details of these internals of ranges.

Did we add documentation for this new feature?
LGTM!

In D151740#4384671, @pzheng wrote:

In D151740#4384276, @philnik wrote:

Please don't land libc++ patches until the libc++ group has approved.

Oops, sorry for my ignorance, @philnik... I didn't know this rule existed for libc++ patch reviews. Thank you for educating me.

LGTM after undoing the formatting changes and the CI passes.
Can you make two followup commits

• format all the new files, I don't like to do that is the same commit, that makes tracking changes in git harder
• modernize the code _LIBCPP_INLINE_VISIBILITY -> _LIBCPP_HIDE_FROM_ABI, _VSTD -> std, typedef -> using

LGTM!

Thanks for catching me copy-pasting from the wrong place.

Addresses review comments.

Thanks for the reviews. I'll add release notes before committing.

Addresses review comments.

CI fixes

Adds an extra test.
Rebased on top of D151223.

Thanks for working on this! I didn't compare the code with the Standard. A few minor nits.

In D151344#4368926, @MaskRay wrote:

Thank you for making another try for the treewide change (which is admittedly very painful and not many people do such work).

Not looked at the patch closely but I wonder about the usage of the pragma and their scope.

In D150907#4376601, @ldionne wrote:

In D150907#4356510, @Mordante wrote:

LGTM! I wonder whether it would make sense to store the cache itself outside the docker image so is can be reused between runs or does the image keep running?

My understanding is that the container keeps running between CI jobs, so that shouldn't be needed.

• Rebased to fix CI
• Fixes and modernizes some tests

In D151244#4371101, @H-G-Hristov wrote:

I don't know if Xcode 14.3 (Apple Clang 14.0.3 base on LLVM15) is a supported compiler

Addresses review comments.

I'm happy, but since Louis had some remarks I leave the final approval to him.

LGTM, if possible I still would like benchmarks, but not a blocker.

@tahonermann when you have time can you have a look?

LGTM, thanks!

LGTM!

I like the change, but want to see a green CI before looking closer at it.

In D151274#4366483, @philnik wrote:

Benchmark numbers:

Would it make sense to add benchmarks? The number in D151274 look quite impressive.

Thanks, LGTM!

I've created D151344 @glandium @hans @thakis I really would appreciate when you can test the patch locally to avoid another revert round.

In D151224#4365477, @philnik wrote:

Are you sure this is expected? I can't really imagine a scenario where an attribute doesn't have a name. Clang itself also seems to assume that the pointer is non-null, since it just dereferences it instantly in quite a few places.

Rebased.