[rooklz]

What

Twelve single-byte profile getters that still threw now return profiles: sceCesRefersUcsProfileAscii, AsciiWithKatakana, Bs4730, GbT1988, GbT1988Tilde0x7e, JisX0201, JisX0201Katakana, JisX0201Roman, JisX0201RomanTilde0x7e, JisX0201Tilde0x7e, KsX1003 and KsX1003Tilde0x7e. They work with every existing single-byte scalar converter (sceCesSbcTo*, sceCes*ToSbc) and with the MBCS string and length functions through sceCesMbcsUcsContextInit, as the 27 profiles of #8923 do.

The tables are measured on a PS5. A homebrew payload that loads the console's libSceCesCs.sprx called sceCesSbcToUtf32 on every byte and sceCesUtf32ToSbc on every BMP scalar for these twelve profiles and the 30 already on main, and each of the twelve now hashes the same as the console's sweep (guest_ces_iso646 pins the hashes). On the console:

• The 7-bit sets (ASCII, BS 4730, GB 1988-80 and its tilde variant, JIS X 0201 Roman and its tilde variant, JIS X 0201 Katakana, KS X 1003 and its tilde variant) answer CES_ERROR_INVALID_ENCODE for bytes 0x80-0xFF, not UNASSIGNED_CODE.
• The changed positions: BS 4730 changes only 0x23 (£); 0x7E stays a tilde. GB 1988-80 puts ¥ at 0x24, JIS X 0201 Roman ¥ at 0x5C, KS X 1003 ₩ at 0x5C, and the base GB, JIS and KS sets put an overline at 0x7E, which the Tilde0x7e variants keep as ~. A reassigned ASCII character has no byte.
• JIS X 0201 Katakana is a 7-bit set: 0x21-0x5F are U+FF61-U+FF9F and 0x60-0x7E are unassigned; the 8-bit profiles (JisX0201, JisX0201Tilde0x7e, AsciiWithKatakana) carry the katakana at 0xA1-0xDF.

SingleByte.hpp gains a two-entry low-half override list per table, an Unassigned marker (0xFFFF) for bytes a set does not define, and a 7-bit flag; sbcToUnicode reports an unassigned byte and the converters answer INVALID_ENCODE for a high byte of a 7-bit set.

Getter signatures follow the Orbital type database. The other single-byte forms (sceCesSbcToUtf8, ToUcs2, the string functions) share the tables and were not measured separately. The same sweep found the surrogate answer 0x805C0015 for every single-byte profile and the C1 holes of Cp1254 and Cp1258 on main; those are other changes (#10183 has the Cp125x ones).

Tested

• Hosted Linux runner, head 209cbf758f (verify run on my fork; Ubuntu 24.04, GCC, Release, lavapipe): every guest_ces* test passes; its failures (agc_driver_sgpr_pattern_fill_keys) are main's on that runner. macOS (Apple clang 21, x86-64): every guest_ces* source passes against this Export.cpp, and a harness built from it matches the console dump for every input above.
• guest_ces_iso646 (new, in both link variants): every byte of every new profile decodes to the expected character or answers CES_ERROR_UNASSIGNED_CODE through sceCesSbcToUtf8, sceCesSbcToUcs2 and sceCesSbcToUtf32, and encodes back through sceCesUtf8ToSbc, sceCesUcs2ToSbc and sceCesUtf32ToSbc; the reassigned positions refuse their ASCII character (sceCesUtf32ToSbc(0x5C, JisX0201Roman) is unassigned, 0x00A5 gives 0x5C); the 63 katakana round-trip to U+FF61-U+FF9F; the Tilde0x7e variants keep ~ and refuse U+203E; the fourteen profiles are distinct pointers; and an MBCS context on JIS X 0201 converts 5C B6 41 7E to ¥ｶA‾ and back and reports 0xE0 as unassigned.
• guest_ces, guest_ces_unicode and guest_ces_mbc still pass, so the 27 earlier profiles and CP1252 are unchanged.
• Linux x86-64 container (Ubuntu 24.04, GCC 13.3, Release, this head built from scratch with ccache): guest_ces_iso646, guest_ces_iso646_module, guest_ces, guest_ces_module, guest_ces_mbc and guest_ces_unicode pass, each binary run directly with its exit status checked.
• Hosted runners (verify run on my fork, which fails a job on any ctest failure): the Windows full build (Server 2025, WinLibs GCC 15.2.0, Release) compiles everything and passes all sixteen guest_ces* tests, the two new ones included; the Linux full build (Ubuntu 24.04, GCC, Release, lavapipe) compiles everything and passes every guest_ces* test, its two failures being agc_driver_queued_index_read_order and agc_driver_sgpr_pattern_fill_keys, which main fails on that runner. An earlier run on 38445f98db, the same code without the Windows PATH entry for the new test, passed the Linux full build with every guest_ces* test green and agc_driver_sgpr_pattern_fill_keys as its only failure, which main fails on that runner; its Windows job could not load libSceCesCs.dll for the new test, which is why the entry is there.
• check_conventions.py passes. pr_overlap.py reports feat(ces): add eight sparse single-byte profiles #10056, fix(libSceCesCs): convert single-byte characters as measured on a PS5 #10183, fix(libSceCesCs): terminate and count MBCS string conversions as measured on a PS5 #10185 and fix(libSceCesCs): encode EUC-JP and Shift-JIS as measured on a PS5 #10186 on the sbcToUnicode lines and feat(libSceCesCs): add the code page 858 profile #10214 and feat(libSceCesCs): add the Windows-1253, 1255 and 1257 profiles #9912 on the SingleByte.hpp table list; none implements these twelve profiles (feat(libSceCesCs): add the Windows-1253, 1255 and 1257 profiles #9912 and feat(libSceCesCs): add the code page 858 profile #10214 add CP1253/1255/1257 and CP858, feat(ces): add eight sparse single-byte profiles #10056 eight sparse ISO 8859 and PC code pages). fix(libSceCesCs): convert single-byte characters as measured on a PS5 #10183 measured sceCesSbcToUtf8 and sceCesUtf8ToSbc on a PS5 and found undefined bytes answering 0x805C0020 in both directions, which is what the Unassigned entries here answer; it rewrites sbcToUnicode to the same bool shape, so whichever lands second resolves one hunk to the same code. feat(libSceCesCs): encode UCS-2, UTF-8, UTF-16 and UTF-32 characters into any MBCS profile #9175 adds a debt entry two lines above this one's.

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
• GPUs tested: n/a
• AI-assisted: yes

[github-actions]

Progress report (b7105f7 → 0a1ee9f)

📈 System libraries: 49.45% (+0.1%, +12 implemented functions)

✅ 12 implemented

Library	Name
libSceCesCs	sceCesRefersUcsProfileAscii
libSceCesCs	sceCesRefersUcsProfileAsciiWithKatakana
libSceCesCs	sceCesRefersUcsProfileBs4730
libSceCesCs	sceCesRefersUcsProfileGbT1988
libSceCesCs	sceCesRefersUcsProfileGbT1988Tilde0x7e
libSceCesCs	sceCesRefersUcsProfileJisX0201
libSceCesCs	sceCesRefersUcsProfileJisX0201Katakana
libSceCesCs	sceCesRefersUcsProfileJisX0201Roman
libSceCesCs	sceCesRefersUcsProfileJisX0201RomanTilde0x7e
libSceCesCs	sceCesRefersUcsProfileJisX0201Tilde0x7e
libSceCesCs	sceCesRefersUcsProfileKsX1003
libSceCesCs	sceCesRefersUcsProfileKsX1003Tilde0x7e

Changed libraries (implemented exports)

Library	Before	After	Change
libSceCesCs	228/287 (79.44%)	240/287 (83.62%)	+4.18 pp

[MarcZX10]

Reviewed head cb3277e. The profile tables and converter paths consistently handle the low-byte overrides and unassigned high-byte sentinels, and the added tests exercise scalar and MBCS round trips across the profiles. I independently confirmed the JIS X 0201 Roman/katakana mappings against Unicode's JIS0201 mapping and ran C++20 syntax checks on Export.cpp and GuestCesIso646.cpp; both pass. The author reports the full focused Linux and Windows test runs in the PR description. The current hosted run passes conventions, eligibility, intake, labels and report; platform builds are skipped for these paths. I could not run the product tests in this environment because the SDL2 submodule and XRandR development dependency are unavailable. I found no blocking issue at this head.

[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 10056	core/libs/prx/libSceCesCs/Export.cpp L92 core/libs/prx/libSceCesCs/PendingExports.cpp L151-230 core/libs/prx/libSceCesCs/include/Api.hpp L240-255 docs/dev/debt/libSceCesCs.md L26-27
PR 11614	core/libs/prx/libSceCesCs/Export.cpp L95-97 core/libs/prx/libSceCesCs/SingleByte.hpp L36, L39, L320 docs/dev/debt/libSceCesCs.md L26-27

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 9912	core/libs/prx/libSceCesCs/Export.cpp core/libs/prx/libSceCesCs/PendingExports.cpp core/libs/prx/libSceCesCs/SingleByte.hpp docs/dev/debt/libSceCesCs.md

[rooklz]

Reworked against a PS5 sweep of every single-byte ISO 646 profile: high bytes are now rejected with 0x805C0014, BS 4730 only remaps 0x23, and JIS X0201 Katakana is GL 0x21-0x5F. The test checks the console sweep hashes for all 12 profiles.

[oneandonlydean]

@rooklz At 5cb7146, remove the test-local APS5_VABI declarations and include the canonical libSceCesCs public header. The exact-head production audit fails on duplicate declarations.

Please run ctest -R public_headers, including the applicable public_headers_compile_* tests, after the header changes; rerun the affected guest tests.

[rooklz]

@oneandonlydean done at da91d0c0c3. The test-local APS5_VABI declarations are removed and the test includes prx/libSceCesCs/include/Api.hpp, which already declares every function it uses. In a Linux x86-64 container (Ubuntu 24.04, GCC 13.3, Release), ctest -R '^guest_ces|^public_headers' passes: public_headers, every public_headers_compile_* batch, and all guest_ces* tests.

[rooklz]

@MarcZX10 could you re-review when you have a moment? Your approval at cb3277e predates the later branch changes. The branch is now exactly two commits: the feature you reviewed (restacked as 350daeee41) and the canonical-header fix dean requested across the stack (0a1ee9fd20, test-only: GuestCesIso646.cpp now includes prx/libSceCesCs/include/Api.hpp and drops the local APS5_VABI declarations, +1/-26). Since then only rebases and the SPIRV-Headers/SPIRV-Tools gitlink rollback dean flagged on #10428; the gitlinks match main's current tip 77f68f314b. Hosted eligibility/conventions/labels/report are green on 0a1ee9fd20.