[iamthehimansh]

What

Part 1 of splitting #8658 into topic PRs, as @adriadam10 asked: the shared-file build fixes that let main compile further on macOS. #4627 (@mugurc) has since landed the Formatting.cpp and ExportMacros.hpp parts, so this PR carries only what is still missing on main. All changes are macOS-only; Linux and Windows compile the same code as before.

• FileStream.hpp: call feof_unlocked / ferror_unlocked unqualified. The macOS SDK defines them as macros, so ::feof_unlocked does not exist.
• StandardExceptions.cpp: const_cast the error-category vtable slot before the function-pointer reinterpret_cast, which Apple clang rejects as casting away qualifiers. Without it libc/src/exception/Exports.cpp fails to build since std::system_category landed.
• ThreadPriority.cpp: macOS has no rtkit or RLIMIT_RTTIME. high asks for the user-interactive QoS class, and rt for a Mach time-constraint policy (period 1 ms, computation 0.5 ms), documented in docs/dev/ENVIRONMENT.md.
• GuestContentSearch.cpp and GuestNpWebApi2.cpp: include <initializer_list>.

On macOS, main still fails 13 objects with this PR, each in a file another open macOS PR changes: GuestAllocations.cpp (branch), DirectMemory.cpp (#7631), Thread.cpp and Process.cpp (#7804), Exception.cpp and GuestRaiseException.cpp (#4630), and NativeStat.cpp, SharedMemory.cpp, Stdio.cpp, DynamicLoader.cpp, ModuleInfo.cpp and Rtld.cpp (#7201).

With BUILD_TESTING=ON, libSceAgcDriver also fails in Driver/Lifecycle.cpp (MAP_FIXED_NOREPLACE), Execution/src/GuestMemory.cpp (link.h), GuestBufferMemory.cpp (linux/udmabuf.h) and Texture.cpp (is_trivially_copyable on a std::shared_ptr). Those fixes were in #8658 and come in the AGC driver part of the split, since they need the driver memory and dma-buf changes there.

Tested

macOS 26 on an Apple M2 Pro, x86_64 under Rosetta 2, Release with Ninja, on main at a23f8fc9f. The library build's remaining failures are the 13 objects above. For Linux and Windows, I checked each changed file's view under unifdef -U__APPLE__, with -D__linux__ and with -D_WIN32: it equals main. python3 tools/check_conventions.py --base origin/main passes.

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 (Cana)

[MarcZX10]

Reviewed the current diff and hosted checks. The macOS-only branches address the reported SDK/compiler failures, the Linux/Windows paths remain unchanged, and the author reports a focused Apple Silicon build plus Rosetta x86_64 compile. Hosted conventions, eligibility, intake, labels, register-context, and write-watch checks pass; Linux and Windows product jobs were skipped by the repository path policy. I found no blocking issue.

[iamthehimansh]

@MarcZX10 Thank you for the review and for checking the hosted runs. I'll keep the branch current with main until it lands.

AI-assisted: yes (Cana)

[github-actions]

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 1896	docs/dev/ENVIRONMENT.md
PR 5021	core/libs/prx/libSceAgcDriver/Execution/src/ThreadPriority.cpp core/libs/prx/libc/include/FileStream.hpp
PR 6924	docs/dev/ENVIRONMENT.md
PR 7201	core/libs/prx/libSceAgcDriver/Execution/src/ThreadPriority.cpp core/libs/prx/libc/include/FileStream.hpp
PR 7447	core/libs/prx/libc/include/FileStream.hpp
PR 7449	docs/dev/ENVIRONMENT.md
PR 7457	docs/dev/ENVIRONMENT.md
PR 7514	docs/dev/ENVIRONMENT.md
PR 7604	docs/dev/ENVIRONMENT.md
PR 7642	docs/dev/ENVIRONMENT.md
PR 8004	docs/dev/ENVIRONMENT.md
PR 8077	docs/dev/ENVIRONMENT.md
PR 8161	docs/dev/ENVIRONMENT.md
PR 8246	docs/dev/ENVIRONMENT.md
PR 8484	docs/dev/ENVIRONMENT.md
PR 8542	docs/dev/ENVIRONMENT.md
PR 8896	docs/dev/ENVIRONMENT.md
PR 9176	docs/dev/ENVIRONMENT.md
PR 9248	docs/dev/ENVIRONMENT.md
PR 9287	docs/dev/ENVIRONMENT.md
PR 9289	docs/dev/ENVIRONMENT.md
PR 9345	docs/dev/ENVIRONMENT.md
PR 9356	docs/dev/ENVIRONMENT.md
PR 9385	docs/dev/ENVIRONMENT.md
PR 9387	docs/dev/ENVIRONMENT.md
PR 9669	docs/dev/ENVIRONMENT.md
PR 9676	docs/dev/ENVIRONMENT.md
PR 9852	docs/dev/ENVIRONMENT.md
PR 9898	docs/dev/ENVIRONMENT.md
PR 9989	docs/dev/ENVIRONMENT.md
PR 10193	docs/dev/ENVIRONMENT.md
PR 10194	docs/dev/ENVIRONMENT.md
PR 10405	core/libs/prx/libc/include/FileStream.hpp
PR 10442	docs/dev/ENVIRONMENT.md
PR 10547	docs/dev/ENVIRONMENT.md

[iamthehimansh]

@oneandonlydean @adriadam10 Could one of you review this when you have time? It's split PR 1 of #8658, as @adriadam10 asked: 6 files of macOS-only build fixes. Linux and Windows compile the same code as main, the hosted checks are green, @MarcZX10 has approved it, and it's current with main. It's in the same area as #4627, #8279 and #10052, which you merged.

AI-assisted: yes (Cana)

[iamthehimansh]

@oneandonlydean @adriadam10 a gentle ping for review when you have time.

It changes 6 files. Every code change is macOS-only: with __APPLE__ undefined, the Linux and Windows preprocessor views of each file are identical to main. All checks pass at 4050dce20, and I keep the branch merged with main. MarcZX10 approved it, and krishalivaibhav built and ran it on an M2 MacBook Air. Happy to split it further or change anything you'd like.

AI-assisted: yes (Cana)

[krishalivaibhav]

@iamthehimansh a note for planning your next split PRs: on current main, a relinked macOS game can't start yet.

I built main (57f506f) plus the open macOS PRs on an M2 (x86_64 Release) and ran Balatro (PPSA21402). It stops in dyld before any guest code runs: symbol not found '_Aps5StartGuest_nid_no_patch'. The relinker's macOS output already binds that symbol (MacOsImage.cpp:668), but nothing on main defines it, and I found no open PR that does.

On a local scratch branch (not for a PR), I took these pieces from your 66a0ac4968, in the order the run needed them:

1. libc: GuestStart.cpp, GuestTls.cpp, include/MainThread.hpp, the two source lines and the CoreFoundation link in libc/CMakeLists.txt, and the 0x400 load-command reserve in the relinker's MacOsImage.cpp. With these, Balatro gets past _Aps5StartGuest.
2. The next stop is _Aps5RunOnMainThread_nid_no_patch, which libSceVideoOut.prx needs: the VideoOut main-thread hunks from the same commit.
3. A current VideoOut links libSceAgcDriver, which doesn't build on macOS on main yet (SpirvOptimizer.cpp:42 pthread_getattr_np, plus Lifecycle.cpp, GuestMemory.cpp, GuestBufferMemory.cpp and Texture.cpp). I stopped there.
4. libSceAudioIn.prx also needs build(libSceAudioIn): keep --exclude-libs off Apple's linker #10669 to link.

Are 1 and 2 already in one of your planned splits? If you open them, I'll run Balatro on M2 with them and report back.

AI-assisted: yes (Claude Code)

[iamthehimansh]

@krishalivaibhav thanks, that's a useful map of what a game needs to start on main.

Yes, 1 and 2 are both planned, as one "start relinked games on macOS" PR built from 66a0ac4968: the libc guest entry, TLS and main-thread pieces, the CoreFoundation link, the relinker's 0x400 load-command reserve, and the VideoOut main-thread hunks. The four libSceAgcDriver files from your step 3 go in a separate AGC-driver PR. The SpirvOptimizer.cpp pthread_getattr_np failure is new since my branch, so I'll add a macOS fix for it there.

The repo's intake check allows one open PR for an author with no merged PRs, so these open one at a time after this one merges. I'll tag you on each one, and a Balatro run on your M2 would be very welcome.

AI-assisted: yes (Cana)

[adriadam10]

Re-checked on current main (6f58a04): the merge is clean. Main's change to GuestContentSearch.cpp (the extern block moved to Api.hpp) doesn't touch the added <initializer_list> include, and the macOS branches of the six files are unchanged. Linux and Windows views of each file are the same as main. Merged.

[iamthehimansh]

@krishalivaibhav the game-start part from your Balatro note is now up as #11284: guest entry, guest TLS, main-thread windows and the relinker's load-command reserve. The SpirvOptimizer.cpp fix stays for the graphics-driver PR.

Balatro still needs libkernel to build on macOS before libc can link, which is your #10552 plus #11246, #11249, #11282 and #10898. If you get a chance to run it on your M2 once those are in, that would be the real test.

AI-assisted: yes (Cana)