[Delciob]

What

Complete stubs in libSceHttp, libSceSsl and libSceConvertKeycode.

Tested

Windows 10 x64, net and input stubs.

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: no

[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 6797	core/libs/prx/libSceConvertKeycode/Export.cpp
PR 9630	core/libs/prx/libSceHttp/Export.cpp
PR 9834	core/libs/prx/libSceHttp/Export.cpp
PR 10167	core/libs/prx/libSceSsl/Export.cpp
PR 10437	core/libs/prx/libSceConvertKeycode/Export.cpp
PR 10535	core/libs/prx/libSceHttp/Export.cpp
PR 10900	core/libs/prx/libSceHttp/Export.cpp
PR 11134	core/libs/prx/libSceHttp/Export.cpp core/libs/prx/libSceSsl/Export.cpp
PR 11145	core/libs/prx/libSceSsl/Export.cpp

[MarcZX10]

I reviewed head 6cdc37108d03603d0dbd950e9e2f8a5109e7ea84 against current main and found regressions/overlap:

• sceSslGetPem copies crt->raw bytes directly and reports the DER length. The existing GuestSslLoad.cpp test requires PEM output (including the PEM size query, PEM text and short-buffer error), and the current main implementation at sceSslGetPem already performs that encoding. The PR version would fail those assertions and should not replace the newer implementation.
• sceConvertKeycodeGetCharacterFromKeyboardData is changed from a no-argument stub to a constant error return, but current main already has a five-argument implementation that reads the requested keyboard data. Keep the current implementation after syncing.
• sceHttpSetCookieEnabled(..., 1) now returns success without enabling cookies; docs/dev/debt/libSceHttp.md says enabling throws because the offline backend has no cookie engine. Please retain the throw until that behavior exists.

Current main merge-tree reports conflicts in libSceConvertKeycode/Export.cpp and libSceSsl/Export.cpp. python3 tools/check_conventions.py --base upstream/main and git diff --check pass, but the PR description gives no focused test command/results. Please rebase onto current main, preserve its existing implementations, keep unsupported cookie enabling explicit, and rerun the focused keycode/SSL tests.

[MarcZX10]

This changes sceHttpSetCookieEnabled(id, 1) from the required unimplemented-path throw to success, but the option is not stored or applied anywhere in this diff. A caller is told cookies are enabled when the implementation still has no cookie support; this directly violates CONTRIBUTING.md’s every-function contract and its limited silent-stub exception. Keep the unsupported case throwing (and retain a test for it) until cookie behavior is implemented. The overlap bot also could not compare libSceHttp/Export.cpp against #9834; both diffs modify this exact export, and #9826 does too. Please coordinate merge order with those authors and declare Depends on where needed.

[MarcZX10]

Two blocking issues in the changed behavior:

• sceSslGetPem copies crt->raw verbatim. In this backend, certificate inputs are parsed into Mbed TLS X.509 objects and raw is the DER certificate buffer (see the bundled mbedtls_x509_crt_parse_der path); this does not produce PEM text. Also, when a non-null output buffer is too short, the function currently reports success after only updating *pemLen. Please serialize the DER certificate to PEM and match the established required-size/short-buffer behavior (for example, compare sceSslGetSerialNumber), with focused tests for sizing, short buffers, and returned PEM content.
• sceHttpSetCookieEnabled(id, 1) now returns success but this PR adds no cookie storage or handling. AGENTS.md says never return made-up success. Keep the unsupported enable path throwing until cookie behavior exists.

The cookie setter change is also duplicated in open #9826; the overlap report and checklist need an agreed order/dependency rather than Depends on: none. The branch is based on 11f976b2, 44 commits behind current main a6646f84. The report only identifies a Windows build, and current product jobs are skipped; focused behavior tests are needed on the final rebased head. I did not run local tests for these changes.

[adriadam10]

Closing, with reasons so you can reopen a fixed version:

• sceHttpSetCookieEnabled(id, 1) now returns success but nothing is stored or applied; it must keep throwing until cookies are implemented (CONTRIBUTING: exact or throw).
• sceSslGetPem copies crt->raw (DER) and reports it as PEM, and returns success when the buffer is too short. main already has a PEM implementation that GuestSslLoad.cpp tests, so this would regress it.
• sceConvertKeycodeGetCharacterFromKeyboardData replaces the throw with a made-up kErrInvalidAddress and has no signature or reference behind it.
• No tests, the branch conflicts with main, and it overlaps feat(libSceHttp): add synchronous HTTP request transport #9834, fix(http): preserve context and epoll state on allocation failure #9630 and feat(libSceConvertKeycode): implement sceConvertKeycodeGetCharacterFr… #6797 on the same exports. Three change requests from @MarcZX10 on 2026-10-22 have had no answer.

Welcome to reopen with one function per PR, a reference for each behaviour, and a test that fails without the change.