[abderahmane-ai]

What

relinker --registry only described the imports of the executable, so a title whose own modules import system functions (PSN.prx, libc.prx and others) could not be audited completely. With --registry the relinker now also writes <output-stem>.<module>.registry.json for every converted bundled module.

An entry is written for each relocation, in DT_RELA and DT_JMPREL, against an undefined symbol that no other bundled module exports. That is the test GuestModuleBuilder already uses to decide that a dependency is satisfied by a sibling. Entries carry nid, library and targetOffset; the symbol has lost its # suffix by the time the module is read, so the NID is the bare 11 characters. The executable's registry is unchanged. The import audit from #1461 accepts these files as they are, and its docs are updated for them.

The first commit of this branch is #1461; the second is this change.

Tested

• core/relinker/relinker/tests/test_guest_registry.py (registered in CTest as guest_registry), for Linux and Windows output: two modules where one is provided by its sibling, both relocation tables, the same module importing the sibling's function once the sibling is removed, and no registry file without --registry.
• Broke the change three ways in throwaway copies (sibling exclusion removed, DT_JMPREL ignored, wrong target offset); the test caught all three.
• All 25 relinker Python tests pass with a relinker built from this branch on macOS 26.5 arm64, and the audit over the executable and module registries gave the expected counts.
• python3 tools/check_conventions.py --base origin/main passes. A trial merge with fix(relinker): resolve internal libc imports from guest libc #1454, the other open relinker PR, is clean.
• Not tested: GCC or MinGW builds (built with clang here), and a real title.

The code, tests and docs were written with AI assistance. I read every line and ran the checks above.

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 in TechnicalDebt
• 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: feat(tools): add an import audit for converted games #1461
• AI-assisted: yes

[adriadam10]

Rebased onto current main: #1461 landed, so its commit dropped out and one commit remains. Checked on Linux: build + full ctest green (464/464); guest_registry fails with the non-test changes reverted; the diff matches the description (relocations in DT_RELA and DT_JMPREL against symbols no other bundled module exports; executable registry unchanged); module registries are accepted by tools/import_audit.py as is. Merging once CI is green on the rebased 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
#1529	core/relinker/relinker/src/guest/GuestModuleBuilder.cpp L184-185
#1597	core/relinker/CMakeLists.txt L230