Actually push the change with the addressed feedback :P

Apply remaining review comments.

Apply review comments and fix CI.

Apply final few review comments.

Apply all review comments except for moving files around.

Apply review comments.

Apply review comments.

Apply review comments. Thanks Mark!

Go through https://libcxx.llvm.org/Contributing.html and fix what needs to be fixed.

Thanks for the cleanup.

Can't wait to start using this! (Note: this is not a full review, just some thoughts.)

Rebase.

Rebase.

Fix modules issues (hopefully).

Rebase.

Rebase (again).

Rebase.

Looks good to me. This seems like a pretty much obviously correct patch, so I wouldn't be opposed to just landing it.

Remove completed todo.

At runtime, not at compile time.
[...]
I'm open to something like what Louis suggested.

• Global constexpr friend -> friend constexpr.

Apply Louis' comments.

Apply Chris' review comments.

Apply remaining few review comments :)

Apply Louis' comments.

Apply Arthur's comments.

Rebase.

Apply Tim's comments and add a few more tests.

Just a note on the two comments above. I understand how they might come across as pedantic, or too picky, but I think it's important that we get this right. If we develop a good model here, it will mean we have a good model for the dozens of algorithms to come. That's why I think this patch should have a high bar (which is not to say that I wouldn't have made those two comments on any other patch).

Just a note on the two comments above. I understand how they might come across as pedantic, or too picky, but I think it's important that we get this right. If we develop a good model here, it will mean we have a good model for the dozens of algorithms to come. That's why I think this patch should have a high bar (which is not to say that I wouldn't have made those two comments on any other patch).

• Update module map.
• Update synopsis.
• Add correct enable_borrowed_range.

Rebase.

Apply review comments.

Generally this looks good to me. I have a few small comments, nothing too major. Looking forward to shipping this!

Rebse.

• Fix ARM and Windows: replace long with auto

Update static_assert message.

I may pick this patch up again later, but for now, I am not going to spend hours debugging why clang is crashing. This patch is not blocking anything.

Fix modules build (fully expecting it to crash).

• Update the release notes.

Apply review feedback.

Rebase.

• Re order when members are initialized in ctor.

• Alphabatize.
• Update module map.
• Make this a NFC for real.

EDIT: if we need to remove something, we'll need to precisely document it in our release notes (including why we needed to push the deletion ahead of schedule), and give users the chance to respond to this change.

Fix rebase fallout (in <iterator>).

Thanks, Louis! Does this mean we officially no longer support Clang 10 (for example) on main? Or does this only come into effect after llvm 13 is released?

Rebase on main.

A simpler way to say the words in the title: this removes <algorithm>'s dependency on both <memory> and <iterator>. (<memory> still depends on <iterator>.)

Rebase on main.

• Fix GCC: make current iterator public.

• I think I adressed all the review comments that apply from the other two views. (If not, sorry, let me know and I'll fix them here too.)
• I'm going to update common_iterator to use variant. Because it's not yet constexpr, I'm just going to comment out the static_assert in all the tests. This makes us support closer to the standard anyway. Note: I have not done this yet because it requries fixing a dependency cycle between iterator -> array -> variant -> iterator.

• Fix GCC for senintel type.
• Change comment.
• Re-format spacing.

• Remove indirectly_swappable.

• Fix GCC bots

• Update based on Chris' comments.
• Add tests for spaceship.
• Add iterator ctor tests and move maybe_const.