[edubox]

What

Export std::numpunct<wchar_t>::id (XZNi3XtbWQ4) and implement std::numpunct<wchar_t>::_Getcat (SdXFaufpLIs), the two unversioned libc symbols that the libc.prx of Rayman 30 Aniversario (PPSA33016) and Fatal Frame II (PPSA32802) import and AnyPS5 does not export.

The facet is the same 0x30-byte C-locale object as the narrow numpunct<char>: grouping is a narrow string pointer at offset 0x10, falsename and truename are pointers to 16-bit strings at offsets 0x20 and 0x28, and decimal_point and thousands_sep are 16-bit values at 0x18 and 0x1a. The grouping virtual writes the narrow string with the narrow size and capacity rules (inline capacity 15, heap capacity length | 15), while falsename and truename write the guest 16-bit string with the same rules as collate<wchar_t>::do_transform (inline capacity 7, heap capacity length | 7, two bytes per unit). _Getcat returns 4, as the narrow entry does.

Console evidence: in the libc.prx shipped by PPSA33016 (unwrapped, dynsym), XZNi3XtbWQ4 is an 8-byte STT_OBJECT in .data, and SdXFaufpLIs is a 0x17b-byte function at 0xa2670 that allocates 0x30 bytes, stores the vtable, the empty grouping string, 16-bit '.' and ',', pointers to the 16-bit strings "false" and "true", and returns 4.

Tested

• guest_locale_tests passes. The new checks cover the wide id's 8-byte size, the facet's 16-bit ./,, the narrow empty grouping, and the 16-bit "false"/"true".
• Rayman 30 Aniversario PPSA33016: with this change the title resolves both NIDs and loads; it presented 11 frames in 120 s with no undefined symbol or crash. The title also imports the _Cnd_register_at_thread_exit/_Cnd_unregister_at_thread_exit pair, so that run used a local build that includes those registrations; this PR does not depend on that work.
• Fatal Frame II PPSA32802 imports the same two NIDs, but its run aborts later in a guest formatting path (__stack_chk_fail) that never reaches the wide facet, so it is not used as evidence here.

Checklist

• Based on current main; no other open PR implements the same functions
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt
• Unimplemented paths throw (NotImplemented_nid_no_patch); silent stubs are listed under Silent stubs in their library's technical debt file
• No notes, investigation .md files or images added to the repository (attach them to this PR)
• New third-party code is a submodule built from source
• Generic behaviour, not specific to one title
• Depends on: none
• Fixes: none
• Unblocks: none (PPSA33016 and PPSA32802 import these NIDs; neither has a Game bring-up discussion)
• GPUs tested: n/a
• AI-assisted: yes

[github-actions]

Progress report (ba62780 → 014001a)

📈 System libraries: 26.25% (+0.01%, +7 implemented functions)

✅ 7 implemented

Library	Name
libc	AssignWideCString
libc	WideNumpunctDecimalPoint
libc	WideNumpunctFalsename
libc	WideNumpunctGrouping
libc	WideNumpunctThousandsSep
libc	WideNumpunctTruename
libc	_ZNSt8numpunctIwE7_GetcatEPPKNSt6locale5facetEPKS1__nid_postfix

Changed libraries (implemented exports)

Library	Before	After	Change
libc	1489/1505 (98.94%)	1496/1512 (98.94%)	+0.0049%

[edubox]

Rebased onto current main (1b2af92baf); head is now 97f4f6c3b0. guest_locale_tests passes and check_conventions.py --base origin/main plus git diff --check are clean.

[github-actions]

Other open pull requests touch the same code. Merging one will break the other, or both implement the same thing. If this one builds on them, list them in Depends on.

Pull request	Conflicting files (lines on main)	Same exports	Same new files
PR 11172	core/libs/prx/libc/src/LocaleSupport.cpp L243-260

Some comparisons could not be completed because a pull request conflicts with the base branch. Review these shared files manually before concluding that there is no overlap.

Pull request	Shared files not compared
PR 5021	docs/dev/debt/libc.md
PR 5303	docs/dev/debt/libc.md
PR 6565	docs/dev/debt/libc.md
PR 7081	core/libs/prx/libc/src/LocaleSupport.cpp docs/dev/debt/libc.md
PR 7386	docs/dev/debt/libc.md
PR 7447	docs/dev/debt/libc.md
PR 7715	docs/dev/debt/libc.md
PR 8150	docs/dev/debt/libc.md
PR 8965	docs/dev/debt/libc.md
PR 9340	docs/dev/debt/libc.md
PR 9371	docs/dev/debt/libc.md
PR 9447	docs/dev/debt/libc.md
PR 9526	core/libs/prx/libc/src/LocaleSupport.cpp core/libs/tests/GuestLocale.cpp docs/dev/debt/libc.md
PR 9929	docs/dev/debt/libc.md
PR 9964	docs/dev/debt/libc.md
PR 10113	docs/dev/debt/libc.md
PR 10226	docs/dev/debt/libc.md
PR 10245	docs/dev/debt/libc.md
PR 10250	docs/dev/debt/libc.md
PR 10253	docs/dev/debt/libc.md
PR 10265	docs/dev/debt/libc.md
PR 10274	docs/dev/debt/libc.md
PR 10315	docs/dev/debt/libc.md
PR 10325	docs/dev/debt/libc.md
PR 10328	docs/dev/debt/libc.md
PR 10345	docs/dev/debt/libc.md
PR 10349	docs/dev/debt/libc.md
PR 10405	docs/dev/debt/libc.md
PR 10506	docs/dev/debt/libc.md
PR 10510	docs/dev/debt/libc.md
PR 10512	docs/dev/debt/libc.md
PR 10565	docs/dev/debt/libc.md
PR 10987	docs/dev/debt/libc.md
PR 11260	docs/dev/debt/libc.md
PR 11262	docs/dev/debt/libc.md
PR 11266	docs/dev/debt/libc.md
PR 11332	docs/dev/debt/libc.md
PR 11831	core/libs/prx/libc/include/Api.hpp docs/dev/debt/libc.md
PR 11907	docs/dev/debt/libc.md
PR 12057	docs/dev/debt/libc.md
PR 12122	docs/dev/debt/libc.md
PR 12292	docs/dev/debt/libc.md
PR 12397	docs/dev/debt/libc.md
PR 12514	docs/dev/debt/libc.md
PR 12560	docs/dev/debt/libc.md
PR 12574	docs/dev/debt/libc.md
PR 12627	docs/dev/debt/libc.md
PR 13390	docs/dev/debt/libc.md
PR 13804	docs/dev/debt/libc.md
PR 14212	docs/dev/debt/libc.md

[edubox]

Rebased automatically onto current main (5bc84038c4); head is now c52c52f602. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[oneandonlydean]

@edubox thanks, the layout and vtable order match Dinkumware's numpunct, and the wide falsename/truename are right. One blocker and one small fix:

1. grouping must stay narrow for numpunct<wchar_t>: the standard has string grouping() const for both facets, and the pinned Dinkumware xlocnum keeps const char* _Grouping and string do_grouping() (only the two names use string_type). WideNumpunctGrouping currently builds it through AssignWideCString, so even the empty C-locale grouping gets the wide representation. Please use the narrow storage/result path (as AssignCString does), have the grouping check verify the narrow representation rather than only size() == 0, and correct the debt paragraph that calls grouping 16-bit.
2. LocaleSupport.cpp has a local extern for the id that Api.hpp already declares. Each export needs exactly one declaration (the public-header rule), so please drop the local one and include the header.

No other changes requested; _Getcat's measured signature stays as it is.

[edubox]

Addressed the review on 7d0f2a2f58:

• numpunct<wchar_t>::grouping now stores and returns a narrow string (const char* with AssignCString); the test checks the narrow inline capacity (15) and terminator instead of only size() == 0, and the debt entry no longer describes grouping as a 16-bit string.
• Removed the local _ZNSt8numpunctIwE2idE_nid_postfix extern; the declaration in Api.hpp is now the only one.

Verified: guest_locale passes, check_conventions.py --base origin/main passes, and git diff --check is clean.

AI-assisted: yes

[edubox]

Rebased automatically onto current main (486919e70a); head is now 1c77eb06ec. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (ca0e9557aa); head is now 37ee72bf5a. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (62b3ba2522); head is now e654d6ef6d. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (75f3702aed); head is now e4d0f50b53. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[maskym]

Windows result for head e4d0f50b53, merged into main 75f3702aed (the PR's Tested section names no Windows run).

Host: Windows 11 26300, Intel i7-12700K, WinLibs GCC 15.2.0 (x86_64-ucrt-posix-seh, r6), Release, BUILD_TESTING=ON, Ninja. Build clean.

The tree also carried the open libc PRs #11793, #11890, #11843 and #11846, all merged cleanly, so the run is a joint one; the test of this PR is guest_locale, which passed in 0.05 s (ctest --test-dir build -R "^guest_locale$"). Not run on a PS5.

AI-assisted: yes (Claude Code), the runs and the numbers above are from this machine.

[edubox]

Rebased automatically onto current main (0c4a8cb41a); head is now d54b9006de. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (14332cd8d0); head is now 25812b316c. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (46fe6f2437); head is now 382e8d5ca0. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (2364892547); head is now aa51c3bb69. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (3d1330179a); head is now 2ce4d516ed. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (de83e8f270); head is now 0fb5a3abd1. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Status on head 0fb5a3abd1: both review points are already in this head.

• WideNumpunctGrouping now uses AssignCString, so grouping keeps the narrow storage and result rules (inline capacity 15), and the test checks that narrow representation (size == 0, capacity == 15, buffer[0] == '\0'). The debt entry now says grouping is a narrow pointer at 0x10.
• The local extern for _ZNSt8numpunctIwE2idE_nid_postfix is gone; Api.hpp is included and carries the single declaration.

Verified on this machine: guest_locale passes, and python3 tools/check_conventions.py --base origin/main plus git diff --check are clean. The two failing write-watch jobs are not from this PR: core/libs/tests/write_watch/CMakeLists.txt on main registers write_watch_execute-fault and write_watch_read-fault twice, and #13220 removes the duplicate block. Branch left unchanged; no rebase or push.

AI-assisted: yes

[edubox]

Rebased automatically onto current main (45f4f82449); head is now 8e4a93402e. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Rebased automatically onto current main (1847a6f9da); head is now e60cd0b4cb. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Watchdog status on e60cd0b4cb: both review points are already in this head, so nothing is pushed.

• WideNumpunctGrouping uses AssignCString, so grouping keeps the narrow storage and result rules. The test now checks the narrow representation (capacity == 15, buffer[0] == '\0'), and the debt paragraph describes the narrow grouping.
• The local extern for the id is gone; Api.hpp is the single declaration.
• guest_locale passes; check_conventions.py --base origin/main and git diff --check are clean.
• The two write-watch failures are not from this patch: main registers write_watch_execute-fault and write_watch_read-fault twice in core/libs/tests/write_watch/CMakeLists.txt. test(libc): register the macOS write watch fault tests once #13220 fixes that registration and is owned by another author.

[edubox]

Rebased automatically onto current main (b8f0a6b202); head is now 7bacb3d9aa. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Watchdog status on 7bacb3d9aa: both review points are already in this head, so nothing is pushed.

• WideNumpunctGrouping uses AssignCString, so grouping keeps the narrow storage and result rules. The test checks the narrow representation (capacity == 15, buffer[0] == '\0'), and the debt paragraph describes the narrow grouping.
• The local extern for the id is gone; Api.hpp is the single declaration.
• The #11504 overlap raised by @verusferro is resolved: feat(libc): throw std::filesystem::filesystem_error from _Xfilesystem_error as the console's libc does #11504 merged on 2026-10-29, this head is rebased on b8f0a6b202, and both calls use GuestLocale::AllocateStringStorage.
• guest_locale passes; check_conventions.py --base origin/main and git diff --check are clean.
• The two write-watch failures are not from this patch: main registers write_watch_execute-fault and write_watch_read-fault twice in core/libs/tests/write_watch/CMakeLists.txt. test(libc): register the macOS write watch fault tests once #13220 fixes that registration and is owned by another author.

[edubox]

Watchdog re-check on 7bacb3d9aa: no review requests a change and the head has not changed since the last status, so nothing is pushed and the branch is not rebased (16 commits behind, below the 25-commit threshold).

• Both points from @oneandonlydean are already in the head: WideNumpunctGrouping uses AssignCString and the test checks the narrow representation, and the local extern for the id is gone (Api.hpp is the only declaration). The feat(libc): throw std::filesystem::filesystem_error from _Xfilesystem_error as the console's libc does #11504 overlap is resolved: feat(libc): throw std::filesystem::filesystem_error from _Xfilesystem_error as the console's libc does #11504 merged on 2026-10-29 and this head calls GuestLocale::AllocateStringStorage.
• guest_locale passes; check_conventions.py --base origin/main and git diff --check are clean.
• The two write-watch failures are not from this patch: main still registers write_watch_execute-fault and write_watch_read-fault twice in core/libs/tests/write_watch/CMakeLists.txt. test(libc): register the macOS write watch fault tests once #13220 (e9f09f54fa, open) fixes that registration; this branch is left untouched.

AI-assisted: yes

[edubox]

Watchdog re-check on 7bacb3d9aa against origin/main 474030e584: both points from @oneandonlydean are already in the head, so there is no push and no rebase (16 commits behind, below the 25-commit threshold).

• WideNumpunctGrouping uses AssignCString, the test checks the narrow representation (inline capacity 15, empty C string), and the debt paragraph now describes grouping as narrow.
• The local extern for the id is gone; Api.hpp is the only declaration.
• The feat(libc): throw std::filesystem::filesystem_error from _Xfilesystem_error as the console's libc does #11504 overlap is resolved: feat(libc): throw std::filesystem::filesystem_error from _Xfilesystem_error as the console's libc does #11504 merged on 2026-10-29 and this head calls GuestLocale::AllocateStringStorage.
• guest_locale rebuilt and passes; python3 tools/check_conventions.py --base origin/main and git diff --check are clean.
• The two write-watch failures are not from this patch: main registers write_watch_execute-fault and write_watch_read-fault twice in core/libs/tests/write_watch/CMakeLists.txt. test(libc): register the macOS write watch fault tests once #13220 (e9f09f54fa, open) fixes that registration; this branch is left untouched.

AI-assisted: yes

[edubox]

Rebased automatically onto current main (ca073ec3ad); head is now 8e89d8959c. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Watchdog verification on head 8e89d8959c (no code changes, no rebase; 12 commits behind main 16452da934).

• ctest --test-dir build -R '^guest_locale$': passes.
• python3 tools/check_conventions.py --base origin/main and git diff --check: clean.
• Both points from @oneandonlydean are already in this head: WideNumpunctGrouping uses AssignCString, the test verifies the narrow empty grouping (capacity 15 and a narrow buffer) and the debt paragraph now says grouping is narrow; the local extern for the wide id is gone, and the only declaration is the one in Api.hpp, which LocaleSupport.cpp includes.
• The write-watch failures are upstream duplicate add_test names in core/libs/tests/write_watch/CMakeLists.txt on main; test(libc): register the macOS write watch fault tests once #13220 fixes them.

Nothing is pushed.

[edubox]

Rebased onto current main 0a148be963 (head 90d4acce8e). The macOS write-watch failures were the duplicate execute-fault/read-fault registrations from main, already fixed on main by 19375d89c4; the rebased tree inherits the single registration and cmake -S core/libs/tests/write_watch configures cleanly here.

@oneandonlydean the two review points stay as in the previous head: grouping is stored and returned through the narrow AssignCString path (test checks the narrow inline representation), and LocaleSupport.cpp includes Api.hpp with no local extern for the id.

• guest_locale passes (1/1).
• tools/check_conventions.py --base origin/main: all checks passed.

AI-assisted: yes (opencode).

[uttam12331]

Hi @edubox, #14373 (the numpunct<char> members) touches the same lines of libc/include/Api.hpp and libc/src/LocaleSupport.cpp as this PR. Since this one is older, I stacked #14373 on your two commits and declared Depends on: #11773, so this lands first and nothing changes on your side. I'll rebase #14373 onto main once this merges. Let me know if you'd prefer a different order.

[edubox]

Rebased automatically onto current main (ba627802e2); head is now 014001a2d0. Focused tests pass; check_conventions.py --base origin/main and git diff --check are clean.

[edubox]

Watchdog verification on the current head 014001a2d0 (already rebased onto main ba627802e2; no further changes or push).

• Both points from @oneandonlydean are in this head: WideNumpunctGrouping uses the narrow AssignCString path (the test checks the narrow empty grouping: inline capacity 15 and a narrow buffer), and LocaleSupport.cpp has no local extern for the id; Api.hpp is the single declaration.
• guest_locale rebuilt and passed (1/1).
• python3 tools/check_conventions.py --base origin/main and git diff --check: clean.
• The earlier #11504 overlap is resolved (it merged and the head calls GuestLocale::AllocateStringStorage).

AI-assisted: yes