[libc++] [doc] Add issue tracking for spaceship operator<=> implementation

[mumbleskates]

Summary

Add issue tracking and assignment for the implementation of P1614R2: The Mothership has Landed.

[cjdb]

More status papers are good. I can't see anything obviously wrong with this and will merge it once it's gotten libc++ approval.

[Mordante]

Thanks for working on this. @rarutyun is also working in this area. I think D103478 should also be in this overview.
The documentation build is failing, can you fix this?

libcxx/docs/Status/SpaceshipIssues.csv

1 ↗

(On Diff #365647)

It would be great to have all related papers and issues in this table from the start. It should at the minimum contain P1614.

[mumbleskates]

Updates from feedback, more differentials and papers

[mumbleskates]

In D107877#2938647, @Mordante wrote:

Thanks for working on this. @rarutyun is also working in this area. I think D103478 should also be in this overview.
The documentation build is failing, can you fix this?

Thanks! I was able to traverse from here and find a few more in-progress diffs and specifications. I don't feel the need for this to be an unassailably complete accounting of the total state of work at present, I think it's more important that it exist at all first.

It looks like the documentation build failed because there was a promoted warning that that one CSV had no data rows.

libcxx/docs/Status/SpaceshipIssues.csv

1 ↗

(On Diff #365647)

Okay I've added that one.

I don't know of many other papers that cover the state of the library standards at the moment, though I did find in putting this together that the <=> operator in [stacktrace.entry.cmp] is not described in the standard; I added P1614R2 and the brand new P2404R0 and a few related LWGs from the C++20 list; let me know if this looks wrong.

[mumbleskates]

• Language standard is fine, stacktrace_entry is specified to model total ordering which is enough
• add missing quote

[Mordante]

Thanks for the update, LGTM! @Quuxplusone since you're also working on the spaceship operator can you have a look and give the libc++ approval?

[• Quuxplusone]

libcxx/docs/Status/Spaceship.rst

40–41

FWIW, I don't understand what this sentence means at all. I assume it was copied from somewhere else where its meaning is equally unclear.
My guess is that it just means "When there's a TODO in the code, we'll mark it with a TODO comment." Which is pretty obvious.

libcxx/docs/Status/SpaceshipPapers.csv

10

I don't understand what a Status of C++20 means (as opposed to |In Progress|).

libcxx/docs/index.rst

45–49

It'd be nice to a-z these: Format, then Ranges, then Spaceship.

[mumbleskates]

• sort status docs
• update LWGs status

libcxx/docs/Status/Spaceship.rst

40–41

Yeah this is all lifted from the Ranges one, which has one TODO item in it to add more tests when spaceship operator implementation increases. I can't think of anything that should go here now but it seems reasonable to leave a space for it.

libcxx/docs/Status/SpaceshipPapers.csv

10

That's the literal status of the LWGs; I can go through this and see what parts of this we've done, not done, or haven't affected our lack of doing things yet instead

[• Quuxplusone]

libcxx/docs/Status/SpaceshipPapers.csv

10

Ah. But on line 2 we've got one |In Progress| already, which I think makes more sense.
The "Status" here should be the status of the feature in re libc++, i.e. blank, |In Progress|, or |Complete|. If people want to know the status of the feature in re Standard adoption and/or progress-through-WG21, they can go look that up elsewhere. (Also, any paper that hasn't yet been at least nearly-adopted for some version of C++ probably shouldn't be in this list at all.)

[mumbleskates]

Add P2405 and fix quotes

[mumbleskates]

libcxx/docs/Status/SpaceshipPapers.csv

10

Yeah this is more up to date now.

I'm not 100% sure what the status is of Justin Bassett's recent papers, but I sense wisdom in them and I'm keeping on the list with no status.

[mumbleskates]

fix typo from copy/pasted paper title

[mumbleskates]

• Working on tuple, optional, & variant

[ldionne]

libcxx/docs/Status/SpaceshipPapers.csv

3–9

I'd like to track those using the regular Cxx20Papers.csv and Cxx20Issues.csv lists instead. SpaceshipProjects.csv should only contain stuff related to https://wg21.link/P1614, otherwise we duplicate information and things become too complicated.

Also, before you mark any LWG issue as "Nothing To Do", you need to explain why that's the case. Normally, this is done as a NFC commit after the fact.

[mumbleskates]

libcxx/docs/Status/SpaceshipPapers.csv

3–9

Ok, great. I'm working on cleaning this up further in D108502, we can fix it there.

I'm fine with removing entries here. I originally had this empty, then I went back after feedback to add papers related to the effort. If we don't want duplicates that's fine.

Nothing To Do LWG entries are all individually justifiable because they affect parts of the standard that are not yet implemented, already implemented with the specified changes, or the changes do not affect the implementation.

[ldionne]

libcxx/docs/Status/SpaceshipPapers.csv

3–9

Nothing To Do LWG entries are all individually justifiable because they affect parts of the standard that are not yet implemented, already implemented with the specified changes, or the changes do not affect the implementation.

If the relevant part of the standard is not implemented in libc++ yet, then the LWG issue isn't "Nothing to do". We say "Nothing to do" when our implementation already implemented the LWG issue resolution.

I want those "Nothing To Do" entries to be done as a separate review so we can have a discussion on those.

libcxx/docs/Status/Spaceship.rst

+.. spaceship-status:
+
+==============================================
+libc++ Spaceship Operator Status (operator<=>)
+==============================================
+
+.. include:: ../Helpers/Styles.rst
+
+.. contents::
+   :local:
+
+
+Overview
+================================
+
+This document contains the status of the C++20 spaceship operator support
+in libc++. It is used to track both the status of the sub-projects of the effort
+and who is assigned to these sub-projects. This is imperative to effective
+implementation so that work is not duplicated and implementors are not blocked
+by each other.
+
+If you are interested in contributing to this effort, please send a message
+to the #libcxx channel in the LLVM discord. Please *do not* start working on any
+of the assigned items below.
+
+
+Sub-Projects in the Implementation Effort
+=========================================
+
+.. csv-table::
+   :file: SpaceshipProjects.csv
+   :header-rows: 1
+   :widths: auto
+
+
+Misc. Items and TODOs
+====================================
+
+(Note: files with required updates will contain the TODO at the beginning of the
+list item so they can be easily found via global search.)
+
+
+Paper and Issue Status
+====================================
+
+.. csv-table::
+   :file: SpaceshipPapers.csv
+   :header-rows: 1
+   :widths: auto

libcxx/docs/Status/SpaceshipPapers.csv

+"Number","Name","Status","First released version"
+`P1614R2 <https://wg21.link/P1614>`_,The Mothership has Landed,|In Progress|,
+`P2404R0 <https://wg21.link/P2404>`_,"Relaxing equality_comparable_with's, totally_ordered_with's, and three_way_comparable_with's common reference requirements to support move-only types",,
+`P2405R0 <https://wg21.link/P2405>`_,nullopt_t and nullptr_t should both have operator<=> and operator==,,
+`LWG3330 <https://wg21.link/LWG3330>`_,Include <compare> from most library headers,|In Progress|,
+`LWG3347 <https://wg21.link/LWG3347>`_,"std::pair<T, U> now requires T and U to be less-than-comparable",|Nothing To Do|,
+`LWG3350 <https://wg21.link/LWG3350>`_,Simplify return type of lexicographical_compare_three_way,|Nothing To Do|,
+`LWG3360 <https://wg21.link/LWG3360>`_,three_way_comparable_with is inconsistent with similar concepts,|Nothing To Do|,
+`LWG3380 <https://wg21.link/LWG3380>`_,common_type and comparison categories,|Nothing To Do|,
+`LWG3395 <https://wg21.link/LWG3395>`_,Definition for three-way comparison needs to be updated,|Nothing To Do|,

libcxx/docs/Status/SpaceshipProjects.csv

+Section,Description,Dependencies,Assignee,Complete
+[cmp.concept],"| `three_way_comparable <https://reviews.llvm.org/D103478>`_
+| `three_way_comparable_with <https://reviews.llvm.org/D103478>`_",None,Ruslan Arutyunyan,In progress
+[cmp.result],| `compare_three_way_result <https://reviews.llvm.org/D103581>`_,None,Arthur O'Dwyer,In progress
+[expos.only.func],"| `synth-three-way <https://reviews.llvm.org/D107721>`_
+| `synth-three-way-result <https://reviews.llvm.org/D107721>`_",[cmp.concept],Kent Ross,In progress
+[comparisons.three.way],| `compare_three_way <https://reviews.llvm.org/D80899>`_,[cmp.concept],Christopher Di Bella,In progress
+[cmp.alg],"| `strong_order <https://reviews.llvm.org/D107036>`_
+| `weak_order <https://reviews.llvm.org/D107036>`_
+| `partial_order <https://reviews.llvm.org/D107036>`_",None,Arthur O'Dwyer,In progress
+[alg.three.way],| `lexicographical_compare_three_way <https://reviews.llvm.org/D80902>`_,[comparisons.three.way],Christopher Di Bella,In progress
+[coroutine.handle.compare],| coroutine_handle,[comparisons.three.way],Unassigned,Not started
+[pairs.spec],| `pair <https://reviews.llvm.org/D107721>`_,[expos.only.func],Kent Ross,In progress
+[syserr.errcat.nonvirtuals],| error_category,[comparisons.three.way],Unassigned,Not started
+[syserr.compare],"| error_code
+| error_condition",None,Unassigned,Not started
+[tuple.rel],| `tuple <https://reviews.llvm.org/D108250>`_,[expos.only.func],Kent Ross,In progress
+"[optional.relops]
+[optional.nullops]
+[optional.comp.with.t]","| optional
+| nullopt",None,Kent Ross,In progress
+"[variant.relops]
+[variant.monostate.relops]","| monostate
+| variant",None,Kent Ross,In progress
+[unique.ptr.special],| unique_ptr,[comparisons.three.way],Unassigned,Not started
+[util.smartptr.shared.cmp],| shared_ptr,[comparisons.three.way],Unassigned,Not started
+[type.index.members],| type_index,None,Unassigned,Not started
+[stacktrace.entry.cmp],| stacktrace_entry,None,Unassigned,Not started
+[stacktrace.basic.cmp],| basic_stacktrace,[alg.three.way],Unassigned,Not started
+[string.cmp],| `basic_string <https://reviews.llvm.org/D80895>`_,None,Christopher Di Bella,In progress
+[string.view.comparison],| `basic_string_view <https://reviews.llvm.org/D80891>`_,None,Christopher Di Bella,In progress
+[array.syn],| array,[expos.only.func],Unassigned,Not started
+[deque.syn],| deque,[expos.only.func],Unassigned,Not started
+[forward.list.syn],| forward_list,[expos.only.func],Unassigned,Not started
+[list.syn],| list,[expos.only.func],Unassigned,Not started
+[vector.syn],| vector,[expos.only.func],Unassigned,Not started
+[associative.map.syn],"| map
+| multimap",[expos.only.func],Unassigned,Not started
+[associative.set.syn],"| multiset
+| set",[expos.only.func],Unassigned,Not started
+[queue.ops],| queue,None,Unassigned,Not started
+[stack.ops],| stack,None,Unassigned,Not started
+[reverse.iter.cmp],| reverse_iterator,None,Unassigned,Not started
+[move.iter.cmp],| move_iterator,None,Unassigned,Not started
+[counted.iter.cmp],| counted_iterator,None,Unassigned,Not started
+[range.iota.iterator],| ranges::iota_view::iterator,[concepts.cmp],Unassigned,Not started
+[range.transform.iterator],| ranges::transform_view::iterator,[concepts.cmp],Unassigned,Not started
+[range.elements.iterator],| ranges::elements_view::iterator,[concepts.cmp],Unassigned,Not started
+"[time.duration.comparisons]
+[time.point.comparisons]
+[time.cal.day.nonmembers]
+[time.cal.month.nonmembers]
+[time.cal.year.nonmembers]
+[time.cal.md.nonmembers]
+[time.cal.mdlast]
+[time.cal.ym.nonmembers]
+[time.cal.ymd.nonmembers]
+[time.cal.ymdlast.nonmembers]
+[time.zone.nonmembers]
+[time.zone.leap.nonmembers]
+[time.zone.link.nonmembers]","| chrono::duration
+| chrono::time_point
+| chrono::day
+| chrono::month
+| chrono::year
+| chrono::month_day
+| chrono::month_day_last
+| chrono::year_month
+| chrono::year_month_day
+| chrono::year_month_day_last
+| chrono::time_zone
+| chrono::leap_second
+| chrono::time_zone_link",None,Unassigned,Not started
+[fs.path.nonmember],| filesystem::path,None,Unassigned,Not started
+[fs.dir.entry.obs],| filesystem::directory_entry,None,Unassigned,Not started
+[re.submatch.op],| sub_match,None,Unassigned,Not started
+[thread.thread.id],| thread::id,None,Unassigned,Not started

libcxx/docs/index.rst

.. toctree::
    ReleaseNotes
    UsingLibcxx
    BuildingLibcxx
    TestingLibcxx
    Contributing
    Status/Cxx14
    Status/Cxx17
    Status/Cxx20
    Status/Cxx2b
-   Status/Ranges
    Status/Format
+   Status/Ranges
+   Status/Spaceship
 
 
 .. toctree::
     :hidden:
 
     AddingNewCIJobs
     FeatureTestMacroTable