feat(libSceSystemService): open sceSystemServiceLaunchWebBrowser URLs in the host browser, opt-in with ANYPS5_OPEN_URLS by MarcZX10 · Pull Request #12895 · boykopovar/AnyPS5

MarcZX10 · 2026-10-29T13:42:13Z

What

sceSystemServiceLaunchWebBrowser (one URI argument; the null, 6015-byte and http:/https:/psno: checks already on main come from a measured library reading) returned EOPNOTSUPP for every valid address, so a title that asks for its browser (PPSA76677 falls back to a QR code) got nothing. It can now open the address in the host's default browser, opt-in through ANYPS5_OPEN_URLS=never|ask|always. Unset, empty and never keep today's result exactly.

Scope note: the requested target was libSceWebBrowserDialog, but no public source gives the layout of the sceWebBrowserDialogOpen parameter or the offset of its URL field (OpenOrbis-PS4-Toolchain @ 0a1aaf9dd4a92695538bdeb09fb056d06dd11725 include/orbis/WebBrowserDialog.h declares every function with no parameters; shadPS4 @ f68923fe807f657348e6719ce658bcbd0a2b1c72 webbrowserdialog.cpp takes none). The dialog is unchanged except for one debt line saying so. The store and profile dialogs carry ids, not URLs.

Design (all host-side, injected behind interfaces):

ValidateHostUrl: only http:// and https://, at most 2048 bytes, RFC 3986 characters only (so no control characters, spaces, non-ASCII or look-alike Unicode hosts), percent escapes well formed and not decoding to a control character, no credentials (user@, good.com@evil.com), non-empty host (DNS names or IPv6 literals), port 1 to 65535.
IUrlLauncher and IUrlConfirmation; OpenUrlOnHost applies the mode. An address that fails the validator reaches neither backend in any mode.
Backends: posix_spawnp of xdg-open (Linux) or /usr/bin/open (macOS) with the address as a single argv element and stdio on /dev/null, never a shell; ShellExecuteW on Windows after revalidating the address. Each backend refuses anything that is not a valid URL by itself.
Confirmation: SDL_ShowMessageBox (SDL2-static is now linked into libSceSystemService) showing the whole address, Cancel as the default for Return/Escape; no display means no. It does not use or stack on the overlay dialogs of feat(overlay): show PS5 system dialogs as real PC dialogs #12379 (no shared files; the conceptual overlap is "a host dialog"). A mutex serialises launches so two guest threads never stack boxes.
Results: launched returns 0 (the native success value is not measured, recorded as debt); declined, rejected, failed or disabled return the existing EOPNOTSUPP. An invalid ANYPS5_OPEN_URLS throws naming the variable at the first call past the argument checks.

Commits (22): validator + test; mode parsing + test; interfaces/policy + test; POSIX backend + test; Windows backend; POSIX launcher revalidation + test; SDL confirmation + test; wiring + test; env reference; usage; five debt statements (one per commit).

Overlap with open PRs: none edit the same lines. #12848 and #12752 also add rows/paragraphs to docs/dev/ENVIRONMENT.md and docs/user/USAGE.md and touch libSceSystemService (language and date format variables); #12285 touches libSceSystemService Export.cpp elsewhere. #12379 touches libSceMsgDialog and other CMake files, not these.

Tested

Linux x86_64, g++ 12, archive build with the host CPU gate patched (run_tests.sh).

Branch (built before the last two rebases, which only brought in unrelated upstream files; checks below re-run after): guest_system_service_open_url (new), guest_system_service, guest_system_service_language, guest_system_service_time_zone, guest_system_service_usb_storage, guest_system_service_app_title_id_{known,long,missing}, loadexec_arguments, guest_system_exit, guest_system_exit_parameters: 11/11 passed.
Baseline (upstream/main 5454cd2, second tree): the same ten existing tests, 10/10 passed, identical per test, so the default behaviour is unchanged (the existing test pins EOPNOTSUPP for http:, https:, psno: and the sourced error codes).
Negative controls (new test, mutation in the built tree, rebuild, ctest): 24 mutations, all aborted the test: credentials check removed, control-escape check removed, port upper bound, port zero, scheme prefix loosened, length limit raised, non-ASCII allowed, space allowed, empty host allowed, IPv6 characters unchecked, unknown mode returning never, variable name dropped from the message, policy without validation, ask skipping the question, always asking, never launching, launcher failure ignored, POSIX launcher without revalidation, POSIX launcher spawning sh, POSIX launcher reporting failure after spawn, confirmation wrap off, confirmation truncating, confirmation answering yes on SDL failure, wiring ignoring the mode, wiring always returning 0, wiring ignoring the environment, null-argument code changed. The failing line inside the test was not extracted (assertions use abort), so "intended case" is by construction of each mutation, not by line.
Checks: tools/check_conventions.py --base upstream/main, git diff --check, test_public_headers.py --audit-only, env_reference.py --check all clean on the final rebased head.
MinGW x86_64-w64-mingw32-g++ -fsyntax-only on the Windows backend.

NOT RUN: Windows (ShellExecuteW backend and SDL box), macOS (open backend, main-queue dispatch of the box), any real launch of a browser or a real message box (tests use a recording script and fakes; the Linux no-display test only proves the box answers no without a display), a Wayland-only session, a console, a title (PPSA76677). No test rebuild after the final two rebases (my files and the tests are unchanged). The unfixed code was not run against the new tests beyond the baseline list above.

Checklist

oneandonlydean · 2026-10-29T13:48:26Z

@MarcZX10 thanks, reviewed at 7ae84d2. The validator, the shell-free launch on each OS (posix_spawnp argv / ShellExecuteW), the backend revalidation and the default-off behaviour read well, and the tests cover the rejections thoroughly. One point before merge, plus a small documentation one:

Repeated calls. Nothing limits how often a title can trigger this. Export.cpp:47-52 serializes calls with a mutex, but HostUrlPolicy.hpp:48-53 keeps no shared state, so in ask mode a title that retries sceSystemServiceLaunchWebBrowser after the user presses Cancel shows the box again straight away, and in always mode a call in a loop opens a new browser tab each time. With counting fakes over the exact validator and policy headers (no real browser or dialog), 1,000 ask calls answered "no" all reached the confirmation, and 1,000 always calls all reached the launcher. On POSIX each accepted launch also starts a child and a detached wait thread, so the mutex doesn't bound outstanding host work either.

Could you add a shared limit across URLs and guest threads? For example: after a decline, return the existing unavailable result without asking again for the rest of the session (or a cooldown), cap prompts and launches to a few per minute, and don't queue up further dialogs from callers waiting behind the active one. Deterministic tests with fake time and backends would cover it: repeated same and distinct URLs, repeated Cancel, concurrent callers, recovery after the interval, and never making no backend calls.
Docs. always opens any http/https host, including localhost and private-network addresses such as a router's page. Could docs/user/USAGE.md say that, so users know what always allows compared with ask?

There is also a small conflict with #12848 in Export.cpp and the environment/usage docs; whichever lands second rebases.

MarcZX10 · 2026-10-29T14:16:42Z

Re-checked — new head `bce7cb8`, both requests addressed

Both points were fair: nothing bounded repeated calls, and always was under-described.

What I ran

Head bce7cb8451d7efd3088ad8733c9d19325f8f6e2e, rebased onto upstream/main 1408347b2e (conflict-free; the history was rewritten by that rebase, so SHAs of the earlier commits changed). feat(libSceSystemService): take the console date and time format from ANYPS5_DATE_FORMAT and ANYPS5_TIME_FORMAT, with auto from the host #12848 and feat(libSceSystemService): take the console language from the host with ANYPS5_LANGUAGE=auto #12752 not touched; whichever lands second rebases.
Archive build on Linux x86_64 (g++ 12): run_tests.sh '^guest_system_service|^loadexec_arguments$|^guest_system_exit' with the new guest_system_service_open_url_limits_tests, guest_system_service_open_url_tests and the existing system-service, loadexec and exit targets.
check_conventions.py --base upstream/main, git diff --check, env_reference.py --check, test_public_headers.py --audit-only.

Changes (15 new commits, each with its test)

One shared OpenUrlGate for every address and guest thread, with an injected clock:
- a caller that arrives while a box or launch is active gets EOPNOTSUPP at once (try-lock, no queued dialog; the old queuing mutex is gone);
- at most 3 prompts or launches in a sliding 60 s;
- after a Cancel, no prompt or launch for 600 s. I chose a bounded cooldown over "rest of the session" so a user who cancelled once can still open a later link without restarting; suppressed calls do not use the per-minute attempts;
- never makes no backend call (1000 calls counted);
- POSIX: at most 2 openers (child plus wait thread) outstanding; a third is refused until one exits.
docs/user/USAGE.md: always opens any http/https host including localhost and private-network addresses such as a router page, unlike ask; the limits are described. New debt entry: the limits are host policy, not console measurements.

Results

12/12 passed: guest_system_service_open_url, guest_system_service_open_url_limits (new) and the ten existing tests (system service, language, time zone, usb storage, three app title id, loadexec_arguments, two exit). The existing tests are the same set that passed on upstream/main earlier, so defaults are unchanged.
Counting fakes over the real validator and policy cover repeated same and distinct URLs (1000 calls each), repeated Cancel, 16 concurrent callers through the exported call while a box is open (one box, 16 refusals), recovery at exactly 60 s and 600 s, never, and the opener bound with a blocking script.

Negative control

20 mutations in the built tree, each rebuilt: blocking lock instead of try-lock (hang, 20 s timeout), cap off by one, window boundary, attempts not recorded, cap checked after recording, suppression removed, suppression boundary, decline not recorded, suppressed calls counted as attempts, never reaching the gate state, state kept across reset, validation removed from the gate, gate not shared in Export.cpp, injected clock ignored, opener cap raised, opener count never released, opener check disabled; and the three constants (cooldown 60 s, window 3600 s, max 4). Every one made at least one of the two new tests fail, except that the cooldown constant first passed because the test derived values from it, so I added literal assertions for 3, 60 s and 600 s and re-ran it: now caught.
Not mutation-tested: the posix_spawn_file_actions_init failure path of the opener counter.

Not verified

Windows, macOS, a real browser or message box, Wayland-only, a title, a console. The ShellExecuteW backend has no opener bound (it starts no child or thread here) and is only MinGW syntax-checked.
upstream/main moved again after my last build (unrelated files); I did not rebuild on it.

I will re-run the checks on any new head and say so here.

github-actions · 2026-10-29T14:31:41Z

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 4472	`docs/dev/ENVIRONMENT.md`
PR 5021	`docs/user/USAGE.md`
PR 5925	`docs/user/USAGE.md`
PR 6186	`docs/dev/ENVIRONMENT.md`
PR 6924	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
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` `docs/user/USAGE.md`
PR 7643	`docs/dev/ENVIRONMENT.md`
PR 8004	`docs/dev/ENVIRONMENT.md`
PR 8484	`docs/dev/ENVIRONMENT.md`
PR 8651	`core/libs/prx/libSceSystemService/Export.cpp` `docs/dev/debt/libSceSystemService.md`
PR 9012	`docs/dev/ENVIRONMENT.md`
PR 9289	`docs/dev/ENVIRONMENT.md`
PR 9345	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 9356	`docs/dev/ENVIRONMENT.md`
PR 9387	`docs/dev/ENVIRONMENT.md`
PR 9852	`docs/dev/ENVIRONMENT.md`
PR 10193	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 10442	`docs/dev/ENVIRONMENT.md`
PR 10678	`docs/dev/ENVIRONMENT.md`
PR 10757	`docs/dev/ENVIRONMENT.md`
PR 10894	`docs/dev/ENVIRONMENT.md`
PR 11435	`docs/dev/ENVIRONMENT.md`
PR 11436	`docs/dev/ENVIRONMENT.md`
PR 12324	`docs/dev/ENVIRONMENT.md`
PR 12544	`docs/dev/ENVIRONMENT.md`
PR 12746	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 12849	`docs/dev/ENVIRONMENT.md`
PR 13486	`core/libs/prx/libSceSystemService/Export.cpp` `docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 13793	`docs/dev/ENVIRONMENT.md`
PR 13809	`docs/dev/ENVIRONMENT.md`
PR 13863	`docs/dev/ENVIRONMENT.md`
PR 13920	`docs/dev/ENVIRONMENT.md`
PR 13928	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 13929	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 14051	`docs/dev/ENVIRONMENT.md`
PR 14212	`docs/dev/debt/libSceSystemService.md` `docs/dev/debt/libSceWebBrowserDialog.md`
PR 14344	`docs/dev/ENVIRONMENT.md`
PR 15032	`docs/user/USAGE.md`
PR 15059	`docs/dev/ENVIRONMENT.md`
PR 15136	`docs/dev/ENVIRONMENT.md`
PR 15169	`docs/dev/ENVIRONMENT.md` `docs/user/USAGE.md`
PR 15529	`docs/dev/ENVIRONMENT.md`
PR 15548	`docs/dev/ENVIRONMENT.md`
PR 15570	`docs/dev/ENVIRONMENT.md`
PR 15947	`docs/dev/debt/libSceSystemService.md`

JJenkx · 2026-10-29T16:23:19Z

Windows run of this head (bce7cb8451), merged onto main bb04a4f1ea (clean merge, local only). Windows 11 Enterprise 10.0.22621, Ryzen 9 5950X, WinLibs GCC 15.2.0 UCRT r7, Ninja, Release, BUILD_TESTING=ON, no Vulkan driver visible.

This PR: guest_system_service_open_url, guest_system_service_open_url_limits pass 3 of 3 (ctest --repeat until-fail:3).
Negative control: not possible. With this PR's source files (core/libs/prx/libSceSystemService/Export.cpp, core/libs/prx/libSceSystemService/HostUrl.hpp, core/libs/prx/libSceSystemService/HostUrlBackends.hpp, core/libs/prx/libSceSystemService/HostUrlConfirmation.cpp, core/libs/prx/libSceSystemService/HostUrlLauncherPosix.cpp, core/libs/prx/libSceSystemService/HostUrlLauncherWindows.cpp and 1 more) put back to main's version, and the ones it adds removed, the build fails against the rest of the PR, so there is no source-only control.

AI-assisted: yes (Claude Code).

libSceSystemService/CMakeLists.txt: keep main's system_service_host_locale interface library and add it to this branch's link line next to SDL2-static. docs/dev/ENVIRONMENT.md: keep main's ANYPS5_LANGUAGE row and this branch's ANYPS5_OPEN_URLS row.

…and credentials A file URL, an address with credentials and one with a space start nothing.

HostMessageBoxConfirmation shows the whole address (wrapped every 96 characters for display only) in a warning box with Cancel as the default for Return and Escape. SDL_ShowMessageBox needs no initialised video subsystem and no particular thread on Windows and X11, and the macOS implementation dispatches to the main queue. When SDL cannot show a box, for example with no display, the answer is no. libSceSystemService now links SDL2-static for this.

…answer The wrapped text keeps the whole address and no line over 96 characters. On Linux, with DISPLAY and WAYLAND_DISPLAY removed, the message box cannot be shown and the confirmation answers no without opening anything.

…r on the host After the sourced argument checks (null, 6015 bytes, http:, https: or psno:), the call now consults ANYPS5_OPEN_URLS. Unset, empty or never keep today's EOPNOTSUPP result. With ask or always, a URL that passes the host validator is confirmed and launched through the injectable backends and the call returns 0; a declined, rejected or failed launch still returns EOPNOTSUPP. The launch is serialised so two guest threads never stack confirmation boxes. An invalid ANYPS5_OPEN_URLS value throws an error naming the variable at the first call that gets past the argument checks.

With injected backends: unset, empty and never keep EOPNOTSUPP and touch neither backend; ask confirms before launching and a declined or failed launch returns EOPNOTSUPP; always launches without asking; invalid addresses and the sourced 0x80a10003, 0x80a10007 and 0x80a10008 results reach neither backend in any mode; a bogus variable throws naming it after the argument checks.

Adds the variable to the environment reference: never, ask or always, with unset and empty keeping the current failing result.

Lists the three modes, why always is discouraged, what a missing display does, the accepted address forms and the refusals, and how each host starts the browser.

The silent-stub entry for sceSystemServiceLaunchWebBrowser now states the failure applies when ANYPS5_OPEN_URLS is unset, empty or never.

…bBrowser The result is the default one; ANYPS5_OPEN_URLS=ask or always changes it for addresses the host opens.

…e browser launch The native result for a launched browser is unknown, and the host accepts fewer addresses than the console's own checks.

Only the Linux opener is exercised; the Windows and macOS backends, the macOS main-queue dispatch and Wayland without Xwayland are reasoned from the SDL and platform documentation.

sceWebBrowserDialogOpen does not open anything on the host because no public source gives the parameter layout or the URL offset.

OpenUrlGate runs OpenUrlOnHost behind a try-lock: a caller that arrives while a confirmation box or a launch is in progress gets Busy at once instead of queuing another dialog behind it. never still makes no backend call and an invalid address is rejected before the lock is touched. The gate takes an injectable clock for the limits that follow. Nothing calls it yet.

While a confirmation is blocked, other callers for the same and distinct addresses, in ask and always, get Busy without a second dialog or launch; a never call and an invalid address are answered without touching the lock; 1000 never calls make no backend call.

…nute Every confirmation shown in ask and every launch in always counts as an attempt in a sliding 60 second window kept by the shared gate. A fourth attempt inside the window returns Limited without a dialog or a launch, whatever the address, and the window empties as attempts age out of it by the injected clock.

…unts Three attempts pass in ask and in always, then 1000 same and distinct addresses are limited with no further dialog or launch; the window opens again exactly 60 seconds after its oldest attempt. Rejected addresses do not count, failed launches do.

A declined confirmation suppresses every further prompt and launch, for any address, until 600 seconds later by the injected clock; those calls return Suppressed at once and do not count against the per-minute cap. The pause is bounded, not for the whole session, so a user who cancelled once can still open a link they ask for later without restarting the game.

After a decline every same or distinct address is suppressed in ask and always with no dialog or launch, until exactly 600 seconds later; repeated Cancel shows one box per interval; suppressed calls do not use up the per-minute attempts.

The fake clock, launcher and confirmation move out of the limits test so the exported-call tests can reuse them. No assertion changes.

…the shared gate All guest threads and all addresses now share one OpenUrlGate: callers that arrive during a dialog or launch, repeats after a Cancel and more than three attempts a minute return the existing EOPNOTSUPP without touching a backend. The mutex that queued callers is gone. The test hook takes the clock too and resets the gate; the existing exported-call test advances a fake clock between calls so its sequence is unaffected by the limits.

With counting fakes and a fake clock: repeated Cancel on the same and distinct addresses shows one box until the cooldown ends; more than three launches a minute are refused and resume after the window; 1000 never calls reach no backend; sixteen concurrent callers during an open confirmation get EOPNOTSUPP and add no second box.

Each accepted launch left a child and a detached wait thread for as long as the opener lived. The launcher now counts running openers, shared with the wait threads so it survives the launcher, and refuses a third until one exits. The gate's three a minute still apply on top.

A blocking opener script holds two launches open: a third and twenty more are refused and start nothing; once the first two exit a launch succeeds again.

…esses ask shows the address before anything opens; always opens whatever http or https host the game names, including localhost and a router's page.

Busy callers are refused, three attempts a minute, and ten minutes of silence after Cancel; refused requests fail like the default.

The busy refusal, the attempt cap, the cooldown after Cancel and the opener bound are not measured on the console.

The usage text promises three attempts, 60 seconds and 10 minutes; the limits test derived its sequences from the constants, so changing a constant moved test and code together. The values are now asserted literally.

TechLevelZero · 2026-11-04T15:38:07Z

✅ AMD GPU test: passed

PR head 737ac71 merged onto main b11c7bf · tested 2026-11-04 15:38 UTC · 20 min


GPU	AMD Radeon RX 6300 (RADV NAVI24), driver Mesa 25.2.8-0ubuntu0.24.04.4, Vulkan 1.4.318
Host	Ubuntu 26.04.1 LTS, kernel 7.0.0-42-generic, container ubuntu:24.04
Logs	build and test logs for this run
Build	✅ passed (Release, `-DANYPS5_ENABLE_SPIRV_TOOLS=ON`)
Tests	✅ 2,165 passed · 9 failed (0 new, 6 also fail on main, 3 unstable on main) · 10 skipped (`ANYPS5_REQUIRE_VULKAN=1`)
Test setup	CPU-only tests without a GPU; GPU tests (and any CPU-only test that failed or skipped without one) once on the GPU with `VK_LAYER_KHRONOS_validation` and `ANYPS5_NO_SHADER_CACHE=1`; main ran the same way
Test selection	2,184 of 3,894 tests run: the ones this change can reach, chosen by `tools/select_tests.py` from main's build graph
Vulkan validation	✅ no new validation IDs vs main (967 tests run on the GPU)
Slow tests (>30 s, GPU tests timed with validation layers)	⚠️ `agc_driver_pipeline_journal` 120.3 s, ⚠️ `agc_driver_graphics` 107.0 s, ⚠️ `agc_driver_graphics_maps_fallback` 101.0 s, ⚠️ `agc_driver_bda_device` 58.9 s, ⚠️ `agc_driver_image_sample_unnormalized` 53.3 s

Unstable on main (3, these also failed in recent main runs on this machine; not counted against this PR)

agc_driver_array_layer_target_cpu: failed in 3 of the last 10 main runs
agc_driver_pipeline_journal: failed in 8 of the last 10 main runs
agc_driver_two_sample_alpha_to_coverage_sample_mask_export: failed in 3 of the last 10 main runs

Failing on main too (6, not counted against this PR)

agc_driver_alpha_to_coverage_sample_mask_export
agc_driver_guest_memory_lifetime_direct
agc_driver_image_table
agc_driver_native_depth_bda_clear
agc_driver_native_multisample_depth
agc_driver_submit_progress

Validated but not compared (64, not in the main baseline's validation pass)

public_headers
public_headers_compile_0
public_headers_compile_1
public_headers_compile_10
public_headers_compile_11
public_headers_compile_12
public_headers_compile_13
public_headers_compile_14
public_headers_compile_15
public_headers_compile_16
public_headers_compile_17
public_headers_compile_18
public_headers_compile_2
public_headers_compile_20
public_headers_compile_21
public_headers_compile_22
public_headers_compile_23
public_headers_compile_24
public_headers_compile_25
public_headers_compile_26
public_headers_compile_27
public_headers_compile_28
public_headers_compile_29
public_headers_compile_3
public_headers_compile_30
public_headers_compile_31
public_headers_compile_32
public_headers_compile_33
public_headers_compile_34
public_headers_compile_35
public_headers_compile_36
public_headers_compile_37
public_headers_compile_38
public_headers_compile_39
public_headers_compile_4
public_headers_compile_40
public_headers_compile_41
public_headers_compile_42
public_headers_compile_43
public_headers_compile_44
public_headers_compile_45
public_headers_compile_46
public_headers_compile_47
public_headers_compile_48
public_headers_compile_49
public_headers_compile_5
public_headers_compile_50
public_headers_compile_51
public_headers_compile_52
public_headers_compile_53
public_headers_compile_54
public_headers_compile_55
public_headers_compile_56
public_headers_compile_57
public_headers_compile_58
public_headers_compile_59
public_headers_compile_6
public_headers_compile_60
public_headers_compile_61
public_headers_compile_62
public_headers_compile_63
public_headers_compile_7
public_headers_compile_8
public_headers_compile_9

Validation IDs already present on main (8 IDs in 677 tests)

VUID-VkDeviceCreateInfo-pNext-pNext (677 tests): agc_compute_pipeline_replay, agc_driver_address_draw_space_reuse, agc_driver_address_draw_space_reuse_held, agc_driver_address_draw_space_reuse_off, agc_driver_address_draw_space_reuse_transitions, agc_driver_address_draw_space_reuse_verify, agc_driver_address_passes, agc_driver_address_passes_draw_transitions, and 669 more
VUID-vkCmdCopyImage-srcImage-01548 (11 tests): agc_driver_depth_clear_hle, agc_driver_depth_plane_cpu_write_after_memo, agc_driver_depth_plane_cpu_write_after_page_unwatch, agc_driver_depth_plane_cpu_write_after_partial_unwatch, agc_driver_depth_plane_partial_store_image_copy, agc_driver_depth_plane_partial_store_unimported, agc_driver_depth_plane_partially_unmapped_plane, agc_driver_depth_plane_unmapped_plane, and 3 more
VUID-vkCmdCopyImage-srcImage-01551 (11 tests): agc_driver_depth_clear_hle, agc_driver_depth_plane_cpu_write_after_memo, agc_driver_depth_plane_cpu_write_after_page_unwatch, agc_driver_depth_plane_cpu_write_after_partial_unwatch, agc_driver_depth_plane_partial_store_image_copy, agc_driver_depth_plane_partial_store_unimported, agc_driver_depth_plane_partially_unmapped_plane, agc_driver_depth_plane_unmapped_plane, and 3 more
VUID-VkPipelineColorBlendStateCreateInfo-pAttachments-00605 (2 tests): agc_driver_prepared_exports, agc_driver_srgb8_color_target
VUID-VkSamplerCreateInfo-addressModeU-01079 (1 test): agc_driver_image_gather_volume
VUID-VkSamplerCreateInfo-addressModeU-parameter (1 test): agc_driver_image_gather_volume
VUID-VkSamplerCreateInfo-addressModeV-parameter (1 test): agc_driver_image_gather_volume
VUID-VkSamplerCreateInfo-addressModeW-parameter (1 test): agc_driver_image_gather_volume

Skipped (10)

agc_driver_copied_writer_label_order: skipped, the device reports a minStorageBufferOffsetAlignment of 4, so the dispatch binds its record at +0x1004 in place and nothing is written back; the test needs an alignment above 4
agc_driver_copied_writer_label_order_drain: skipped, the device reports a minStorageBufferOffsetAlignment of 4, so the dispatch binds its record at +0x1004 in place and nothing is written back; the test needs an alignment above 4
agc_driver_float16_unused_nan_payloads: skipped, the device has no native f16 conversions with RTE rounding and kept f16 denormals
agc_driver_gcn_display_present: skipped, no display: Vulkan support is either not configured in SDL or not available in current SDL video driver (offscreen) or platform
agc_driver_linear_target_direct_write_completion: skipped, the device does not keep host imports write-watched, so no completion stamp over the target is tracked
agc_driver_ngg_vertex_wave_reduction: skipped, s3 describes a 32-lane wave only on 32-lane subgroups, and the device has 64 lanes
agc_driver_scattered_runs: skipped, host imports are unavailable: /dev/udmabuf is not accessible
agc_driver_sparse_dma_buf_import: skipped, host imports are unavailable: /dev/udmabuf is not accessible
agc_driver_vertex_elected_atomic_max: skipped, the device runs vertex programs in subgroups, so they keep their cross-lane reductions
video_out_flip_eop_main: skipped, no display or Vulkan device: DisplayWindow: Vulkan support is either not configured in SDL or not available in current SDL video driver (offscreen) or platform

Reproduce

git fetch origin 737ac7199d6fa4048d32016835442d5f421cd434 b11c7bf84f5d2f636ff60166cc3371bd071e24d5
git checkout b11c7bf84f5d2f636ff60166cc3371bd071e24d5 && git merge --no-ff 737ac7199d6fa4048d32016835442d5f421cd434
git submodule update --init --recursive
cmake -S . -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=ON -DANYPS5_ENABLE_SPIRV_TOOLS=ON
cmake --build build && cmake --build build --target libs
VK_DRIVER_FILES=/nonexistent/icd.json ctest --test-dir build -E '^(agc|shader|recompiler|videoout|video_out|guest_video_out|vulkan)|gpu' --timeout 120 --parallel 12 --output-on-failure
ANYPS5_REQUIRE_VULKAN=1 ANYPS5_NO_SHADER_CACHE=1 VK_INSTANCE_LAYERS=VK_LAYER_KHRONOS_validation \
  ctest --test-dir build -R '^(agc|shader|recompiler|videoout|video_out|guest_video_out|vulkan)|gpu' --timeout 120 --parallel 16 --output-on-failure
# failures are then re-run alone: --timeout 360 --parallel 2; new validation IDs: the test alone, twice, --parallel 1

_{Automated build and test on a contributor's AMD machine; it does not replace upstream CI. A new comment is posted for each new push. Full logs.}

MarcZX10 force-pushed the feat/host-open-url branch from 7ae84d2 to bce7cb8 Compare October 29, 2026 14:16

oneandonlydean self-assigned this Oct 29, 2026

github-actions Bot added build docs prx needs rebase Conflicts with main, rebase needed labels Oct 30, 2026

kyon-eth mentioned this pull request Oct 30, 2026

feat(libSceSystemService): declare 32 callable exports #13192

Merged

12 tasks done

This was referenced Oct 30, 2026

feat(libSceSystemService): answer the codec activation calls as firmware 13.42 does #13322

Merged

feat(libSceSystemService): implement sceSystemServiceGetPSButtonEvent #13333

Merged

kyon-eth mentioned this pull request Oct 30, 2026

feat(libSceWebBrowserDialog): declare limited export #13666

Merged

12 tasks done

This was referenced Oct 31, 2026

feat(libSceIme): show a notice when PSN-only system dialogs open #13701

Merged

feat(libSceAgcDriver): ask whether to keep a stalled title running #13707

Open

MarcZX10 force-pushed the feat/host-open-url branch from bce7cb8 to d0a4124 Compare October 31, 2026 12:11

github-actions Bot removed the needs rebase Conflicts with main, rebase needed label Oct 31, 2026

LeandroLP23 mentioned this pull request Oct 31, 2026

docs(debt): correct stale, contradictory and misplaced entries #14085

Merged

12 tasks done

github-actions Bot added the needs rebase Conflicts with main, rebase needed label Oct 31, 2026

rooklz mentioned this pull request Oct 31, 2026

refactor(libSceSystemService): throw the documented running-title rejections #14188

Merged

12 tasks done

MarcZX10 force-pushed the feat/host-open-url branch from d0a4124 to 76ce444 Compare October 31, 2026 23:18

BunnyFl mentioned this pull request Oct 31, 2026

feat(libSceSystemService): answer two ShellCoreUtil calls as firmware 13.42 does #14203

Open

12 tasks done

github-actions Bot removed the needs rebase Conflicts with main, rebase needed label Oct 31, 2026

ken00H mentioned this pull request Nov 1, 2026

feat(libSceSystemService): store the notice screen skip flag #14844

Closed

12 tasks done

CnRJay mentioned this pull request Nov 1, 2026

feat(libSceSystemService): take the console date and time format from ANYPS5_DATE_FORMAT and ANYPS5_TIME_FORMAT, with auto from the host #12848

Open

12 tasks done

github-actions Bot added needs rebase Conflicts with main, rebase needed and removed needs rebase Conflicts with main, rebase needed labels Nov 2, 2026

XairRrR mentioned this pull request Nov 3, 2026

docs(tracing): say APS5_TRACE_TEXTURES counts nonzero bytes at 64-byte steps #15417

Open

12 tasks done

MarcZX10 added 27 commits November 4, 2026 11:36

test(system-service): check the POSIX launcher refuses other schemes …

040915d

…and credentials A file URL, an address with credentials and one with a space start nothing.

docs(env): document ANYPS5_OPEN_URLS

61d93e5

Adds the variable to the environment reference: never, ask or always, with unset and empty keeping the current failing result.

docs(usage): describe ANYPS5_OPEN_URLS and the addresses it opens

f51a530

Lists the three modes, why always is discouraged, what a missing display does, the accepted address forms and the refusals, and how each host starts the browser.

docs(debt): say the browser request only fails by default

d9cfcde

The silent-stub entry for sceSystemServiceLaunchWebBrowser now states the failure applies when ANYPS5_OPEN_URLS is unset, empty or never.

docs(debt): qualify the EOPNOTSUPP result of sceSystemServiceLaunchWe…

ecac5b6

…bBrowser The result is the default one; ANYPS5_OPEN_URLS=ask or always changes it for addresses the host opens.

docs(debt): record the unmeasured success value and host limits of th…

d3bcb3c

…e browser launch The native result for a launched browser is unknown, and the host accepts fewer addresses than the console's own checks.

docs(debt): record which host launch backends are unverified

ab35d08

Only the Linux opener is exercised; the Windows and macOS backends, the macOS main-queue dispatch and Wayland without Xwayland are reasoned from the SDL and platform documentation.

docs(debt): record that the web browser dialog's URL field is unsourced

b9cb9d8

sceWebBrowserDialogOpen does not open anything on the host because no public source gives the parameter layout or the URL offset.

test(system-service): move the counting fakes into a shared test header

1b3df26

The fake clock, launcher and confirmation move out of the limits test so the exported-call tests can reuse them. No assertion changes.

test(system-service): cover the bound on running POSIX openers

4e969b9

A blocking opener script holds two launches open: a third and twenty more are refused and start nothing; once the first two exit a launch succeeds again.

docs(usage): say that always opens localhost and private-network addr…

fa4c0fe

…esses ask shows the address before anything opens; always opens whatever http or https host the game names, including localhost and a router's page.

docs(usage): describe the browser request limits

0dfdccd

Busy callers are refused, three attempts a minute, and ten minutes of silence after Cancel; refused requests fail like the default.

docs(debt): record that the browser request limits are host policy

924a338

The busy refusal, the attempt cap, the cooldown after Cancel and the opener bound are not measured on the console.

test(system-service): pin the documented limit values

737ac71

The usage text promises three attempts, 60 seconds and 10 minutes; the limits test derived its sequences from the constants, so changing a constant moved test and code together. The values are now asserted literally.

MarcZX10 force-pushed the feat/host-open-url branch from b8bd072 to 737ac71 Compare November 4, 2026 11:39

github-actions Bot removed the needs rebase Conflicts with main, rebase needed label Nov 4, 2026

Navigation Menu

Repository navigation

feat(libSceSystemService): open sceSystemServiceLaunchWebBrowser URLs in the host browser, opt-in with ANYPS5_OPEN_URLS - #12895

feat(libSceSystemService): open sceSystemServiceLaunchWebBrowser URLs in the host browser, opt-in with ANYPS5_OPEN_URLS#12895
MarcZX10 wants to merge 37 commits into
boykopovar:mainfrom
MarcZX10:feat/host-open-url

MarcZX10 commented Oct 29, 2026

Uh oh!

oneandonlydean commented Oct 29, 2026

Uh oh!

MarcZX10 commented Oct 29, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Oct 29, 2026 •

edited

Loading

Uh oh!

JJenkx commented Oct 29, 2026

Uh oh!

TechLevelZero commented Nov 4, 2026

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

MarcZX10 commented Oct 29, 2026

What

Tested

Checklist

Uh oh!

oneandonlydean commented Oct 29, 2026

Uh oh!

MarcZX10 commented Oct 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Re-checked — new head bce7cb8, both requests addressed

What I ran

Changes (15 new commits, each with its test)

Results

Negative control

Not verified

Uh oh!

github-actions Bot commented Oct 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JJenkx commented Oct 29, 2026

Uh oh!

TechLevelZero commented Nov 4, 2026

✅ AMD GPU test: passed

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

MarcZX10 commented Oct 29, 2026 •

edited

Loading

Re-checked — new head `bce7cb8`, both requests addressed

github-actions Bot commented Oct 29, 2026 •

edited

Loading