Remove extra semi colon from folly/experimental/coro/Merge-inl.h

Summary:
`-Wextra-semi` or `-Wextra-semi-stmt`

If the code compiles, this is safe to land.

Reviewed By: Orvid

Differential Revision: D57975880

fbshipit-source-id: 2d60cb964fc917722e735c1afce1412f0f4c4458

Back out "Back out "Revert D61286579""

Summary:
Original commit changeset: 251ac011c75f

Original Phabricator Diff: D61629106

Reviewed By: dogdoggoos

Differential Revision: D61776376

fbshipit-source-id: aa8a17c55ccedd8e5c218ab19f985bace26fe2dd

Reduce duplication between `TaskPromise` specializations

Summary: Make the `TaskPromise<>` leaf classes as simple as they can be. This shortens the code by 30 lines, and makes it easier to follow.

Reviewed By: yfeldblum

Differential Revision: D61249849

fbshipit-source-id: c12ca1d96c2c99fd85d53df9a77146450d525ca6

Create an enum for superconsole interaction toggles

Summary:
Current implementation of console toggles relies on `char` and display is error-prone (.e.g we forgot to add `x` option to the Help). I initially tried to fix this in D60960924 but then realized that this could happen in the future. Better to make it type-safe so that we get it for free (via `strum::EnumIter`).

Also rename  `ConsoleInteraction` to `SuperConsoleInteraction`  to signify that these only for for superconsole.

Reviewed By: stepancheg

Differential Revision: D60965981

fbshipit-source-id: e0008b46dd752446b45877e6f0c2acf43f528c7b

folly concepts facebook#1 (facebook#2249)

Summary:
Pull Request resolved: facebook#2249

We have a need for some common concepts that are used in many places.

This is a proposal, please let me know what you think.

Usages example:

`uncvref_same_as`:

https://www.internalfb.com/code/fbsource/[a71ce59323a15fd7559d6132c1f8e2d0256ca7cb]/fbcode/multifeed/entrepot/lib/zdb_manager/ZdbUtil.h?lines=237

`uncvref_instantiated_from`:

https://www.internalfb.com/code/fbsource/[b366209d2c108133a0545edc7943b9461323623b]/fbcode/multifeed/mfaction/MFHashSerializer.h?lines=38

Reviewed By: yfeldblum

Differential Revision: D59396081

fbshipit-source-id: 4912c2f9fe683f4d3b37fce43466c9281d03ce13

Add `SparseByteSet::size()` method

Summary:
Useful, when `SparseByteSet` used as fast pre check for heavyweight
operation.

Reviewed By: marksantaniello

Differential Revision: D60285173

fbshipit-source-id: 8deac984b0f00096448af51fdfee283d0bb13ca2

simplify cache local/lifetime coordination in SingletonRelaxedCounter

Summary:
Rather than giving up on constraining the cache thread_local local and cache thread_local lifetime to be in the same DSO, now constrain it and assert on the constraint. This allows the cache thread_local local and the cache thread_local lifetime to be 1:1 with each other, rather than M:N with each other without the constraint.

Note that there is now no cross-DSO merging of these thread_local variables - they are now explicitly per-DSO.

Differential Revision: D59144274

fbshipit-source-id: 73b3a426059321bb9abfcb92fdbdd3547656adc8

folly: fix github ci tests on macos arm (facebook#2253)

Summary:
These were [failing on github](https://github.com/facebook/folly/actions/runs/9923539774/job/27413784309#step:48:6919)

I think the problem was the macos runner changed from intel to arm, expectations needed updating for arm.

Pull Request resolved: facebook#2253

Test Plan:
```
./build/fbcode_builder/getdeps.py --allow-system-packages build --src-dir=. folly

./build/fbcode_builder/getdeps.py --allow-system-packages --no-facebook-internal test --no-testpilot folly
```

Before, broken
```
+ /opt/homebrew/opt/cmake/bin/ctest \
+      --output-on-failure \
+      -j \
+      3 \
+      --rerun-failed
Test project /private/var/folders/xz/lz5thm6s3vb1vdkk77vmkgbr0000gn/T/fbcode_builder_getdeps-ZUsersZrunnerZworkZfollyZfollyZbuildZfbcode_builder/build/folly
    Start 1646: arena_test.Arena.SizeSanity
    Start 2039: constexpr_math_test.ConstexprMathTest.constexpr_exp_floating
1/2 Test facebook#1646: arena_test.Arena.SizeSanity ....................................***Failed    0.01 sec
Note: Google Test filter = Arena.SizeSanity
[==========] Running 1 test from 1 test suite.
[----------] Global test environment set-up.
[----------] 1 test from Arena
[ RUN      ] Arena.SizeSanity
/Users/runner/work/folly/folly/folly/memory/test/ArenaTest.cpp:81: Failure
Expected: (arena.totalSize()) <= (maximum_size), actual: 1544 vs 1536

/Users/runner/work/folly/folly/folly/memory/test/ArenaTest.cpp:91: Failure
Expected: (arena.totalSize()) <= (maximum_size), actual: 1544 vs 1536

[  FAILED  ] Arena.SizeSanity (0 ms)
[----------] 1 test from Arena (0 ms total)

[----------] Global test environment tear-down
[==========] 1 test from 1 test suite ran. (0 ms total)
[  PASSED  ] 0 tests.
[  FAILED  ] 1 test, listed below:
[  FAILED  ] Arena.SizeSanity

 1 FAILED TEST

2/2 Test facebook#2039: constexpr_math_test.ConstexprMathTest.constexpr_exp_floating ...***Failed    0.01 sec
Note: Google Test filter = ConstexprMathTest.constexpr_exp_floating
[==========] Running 1 test from 1 test suite.
[----------] Global test environment set-up.
[----------] 1 test from ConstexprMathTest
[ RUN      ] ConstexprMathTest.constexpr_exp_floating
/Users/runner/work/folly/folly/folly/test/ConstexprMathTest.cpp:990: Failure
Expected equality of these values:
  std::exp(471.L)
    Which is: 3.5702693074778485e+204
  a
    Which is: 3.5702693074778518e+204

[  FAILED  ] ConstexprMathTest.constexpr_exp_floating (0 ms)
[----------] 1 test from ConstexprMathTest (0 ms total)

[----------] Global test environment tear-down
[==========] 1 test from 1 test suite ran. (0 ms total)
[  PASSED  ] 0 tests.
[  FAILED  ] 1 test, listed below:
[  FAILED  ] ConstexprMathTest.constexpr_exp_floating

 1 FAILED TEST

0% tests passed, 2 tests failed out of 2

Total Test time (real) =   0.13 sec

The following tests FAILED:
	1646 - arena_test.Arena.SizeSanity (Failed)
	2039 - constexpr_math_test.ConstexprMathTest.constexpr_exp_floating (Failed)
```

After, works:
```
...
3067/3067 Test facebook#2213: fbstring_test.FBString.testFixedBugsD4355440 .........................................................***Skipped   0.03 sec

100% tests passed, 0 tests failed out of 3066

Total Test time (real) =  36.02 sec

The following tests did not run:
	1934 - small_locks_test.SmallLocks.PicoSpinLockThreadSanitizer (Skipped)
	1946 - small_locks_test.SmallLocksk.MicroSpinLockThreadSanitizer (Skipped)
	1993 - atomic_unordered_map_test.AtomicUnorderedInsertMap.MegaMap (Disabled)
	2213 - fbstring_test.FBString.testFixedBugsD4355440 (Skipped)
```

Reviewed By: ndmitchell

Differential Revision: D59733599

Pulled By: ahornby

fbshipit-source-id: e7f00184e867a5bf296836ea8c4a4f2430d9635a

Make sure folly's memcpy/memset not replaced by glibc's implementation

Summary:
`__folly_memcpy` and `__folly_memset` are supposed to be used in prod. However, they may be replaced by libc's implementation when `link_group_map` is used.

The issue comes from two places: (1) The `link_group_map` may cause "folly/__memcpy-use__/__objects__/memcpy.S.o" to be put after "libc.so" in the xxx.thinlto.index.argsfile. (2) Even if "folly/__memcpy-use__/__objects__/memcpy.S.o" comes before "libc.so", `__folly_memcpy` may still be replaced by libc's implementation in some cases. For example: if there's no reference to `memcpy` before "folly/__memcpy-use__/__objects__/memcpy.S.o", the symbol `memcpy` inside memcpy.S.o is considered to be a lazy symbol, which will be replaced by the definition inside "libc.so".

To make sure `__folly_memcpy` and `__folly_memset` are linked, this diff sets link_whole for cpp_library memcpy-use and memset-use. Because memcpy-use and memset-use only contain the necessary symbols, link_whole doesn't cause size bloat.

Reviewed By: WenleiHe

Differential Revision: D58921357

fbshipit-source-id: f5c7161dda24edd1ac808042715931d9b8dfaf10

Change lock order expecation between StaticMeta.lock_ and ThreadEntry…

…Set lock.

Summary:
The dispose() of a thread local copy of an object is called when a
thread does a reset() on it. It can invoke arbitrary code including calling
destroy() or reset() on other ThreadLocals. Such a call will need to acquire
the StaticMeta's lock_. To enforce that lock order, we need to fix paths that
today acquire a ThreadEntrySet lock after acquiring StaticMeta's lock_.

Reviewed By: yfeldblum

Differential Revision: D59076052

fbshipit-source-id: 0109d66b3a5f74d7460b8a2d27103f2e919677e9