[AlxcBTW]

What

Adds a guest user to libSceUserService, the second user a game needs to see before it lets a second player join locally. Share play's together mode (#29) will log it in when the mode starts and out when it ends.

• UserServiceSetGuestUser_nid_postfix(userIndex, name) and UserServiceClearGuestUser_nid_postfix(userIndex) in UserService.hpp. This is the seam: nothing in this PR calls it, the share code does later.
• The guest is the slot's existing user id (0x10000000 + index), so scePadOpen for that user already returns the handle of the pad on the same port. Pad 2 of a share is port 1, therefore user index 1. No libScePad change is needed.
• A user is logged in while a controller on the slot or a guest claim holds it. Login and logout events, the login list and the registered list change only when that combined state changes, so a physical controller plugging in or out on the slot cannot log the guest out, and the other way round.
• sceUserServiceGetUserName returns the display name the caller gave, cut to 16 bytes at a UTF-8 character boundary, control characters and invalid UTF-8 dropped, trimmed. With no name, or one that is empty after that, it returns the slot's default name (Player2).
• sceUserServiceGetNpAccountId returns 0 for a guest. The host's PSN account must not be handed to another person.
• The spec's bridge messages (share open, share mode, share pad) carry no guest handle, so the name is a parameter the share code fills when the bridge learns it. Until a name field exists, the default applies. A guest_name string on share mode would be the smallest addition; that is a network spec change, not part of this PR.

Based on main, not on #29: no API of the share branch is used. Once #29 lands, its together handling calls these two functions, and the "Share play guest pad" debt entry in libScePad (Closes when: the user service has a second user) can be closed.

Tested

Not tested on a title. New host test guest_user_service_guest_user (GuestUserServiceGuestUser.cpp): claim before and after sceUserServiceInitialize, event order and collapsing of a claim that ends before it is read, login and registered lists, user number, age level, account id 0, name sanitising (truncation at a character boundary, control characters, invalid UTF-8, empty, whitespace), claim and hotplug on the same slot in both orders, out-of-range slots ignored. The Windows and Linux build and test jobs of this PR run it with the existing user service tests. tools/check_conventions.py passes locally; the C++ tree was not built locally, only syntax-checked.

Needs a real game: whether a title's local-multiplayer join flow accepts a user with no PSN account, and what it does on the logout event while it is in a match. Recorded in docs/dev/debt/libSceUserService.md.

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: the together mode of feat(bridge): share play frame ring, guest pad and host override #29
• GPUs tested: n/a
• AI-assisted: yes

---

^{Need help on this PR? Tag @codesmith-bot with what you need.}

[AlxcBTW]

@parl0v, #30 is merged with current main and green, ready for your re-review. I have not merged it.

What changed in the merge: main merged in with a merge commit (no rebase, no force push; head a18a3b7). There were no conflicts. I also merged the fix from PR #40 into this branch, because every full build.yml run has been red since #9: bridge_e2e_game skips without a net service and check_skips.py rejects that skip. That is unrelated to the guest user.

Build: dispatch run https://github.com/anyplaynet/AnyPS5/actions/runs/40604915342 on this head: Windows, Linux and the relinker jobs (macOS arm64 and x86_64, asan) are all green.

Note for the order of merging: #29 has a TODO at the together mode switch that calls UserServiceSetGuestUser_nid_postfix(1, name) and clears it on leaving, closing and a lost connection. Whichever of #29 and #30 merges second wires it. The guest's display name is not in the bridge messages yet; the proposed follow-up is a guest_name field on share mode.