main: test_vcore_fs_open and test_vcore_video_acquire no longer compile since #2042 met #2301 and #2322 #2379

Closed

opened 2026-11-11 16:44:12 +00:00 by mooki000 · 0 comments

mooki000 commented

2026-11-11 16:44:12 +00:00

(Migrated from github.com)

Hookshot version

main b744b1ec (every main since d522ec0e, 2026-11-11 16:36 UTC), from source

What kind of problem

Something else

What happened

Two game-free tests no longer compile on main, so tests/run_all.py (the "Game-free source regressions" check) fails there:

tests/test_vcore_fs_open.py:        fixture.cpp:77:24: error: use of undeclared identifier 'env_on'
tests/test_vcore_video_acquire.py:  fixture.cpp:159:28: error: use of undeclared identifier 'env_on'

(g++ says 'env_on' was not declared in this scope at the same lines.)

The cause is three merges meeting, and two of them are mine. #2042 made env_on() the rule for on/off switches and changed, among others, the HOS_NO_MOVIES read in OpenFile (host/nn_fs.c) and the HOS_VIDEO_NO_CATCHUP read in dec_acquire (host/nn_video.c). #2301 and #2322, merged two and three minutes later, each added a test that cuts one of those functions out of its source file and compiles it in a small fixture. Both were written on a base before #2042, so their fixtures have no env_on. Each pull request passed alone.

Only the two tests' fixtures are affected. The source files themselves include gpu/vk/compat/hookshot_env.h.

Where in the game

Not in the game: tests/run_all.py

How often

Every time

Graphics card and driver version

Does not depend on it

CPU, RAM and operating system

i9-13900K, 96 GB, Arch Linux (g++ 16.2.1, clang++ 23.1.3)

Changes from the defaults

None

Other software hooked into the game

No, nothing hooked in

hookshot_log.txt

Not a game run. To see it: python3 tests/test_vcore_video_acquire.py --cxx clang++ on main.

Claim

I'm working on this: the fix is one line in each test's compile command (the header with -include, as #2042 did for the other tests that cut code out of the sources), and its pull request follows this issue directly. By docs/workflow.md: S2 (main's checks are red for everyone) x reach (every contributor's run) x confidence 1.0 (cause found, fix passes with both compilers here).

### Hookshot version main b744b1ec (every main since d522ec0e, 2026-11-11 16:36 UTC), from source ### What kind of problem Something else ### What happened Two game-free tests no longer compile on main, so `tests/run_all.py` (the "Game-free source regressions" check) fails there: ``` tests/test_vcore_fs_open.py: fixture.cpp:77:24: error: use of undeclared identifier 'env_on' tests/test_vcore_video_acquire.py: fixture.cpp:159:28: error: use of undeclared identifier 'env_on' ``` (g++ says `'env_on' was not declared in this scope` at the same lines.) The cause is three merges meeting, and two of them are mine. #2042 made `env_on()` the rule for on/off switches and changed, among others, the `HOS_NO_MOVIES` read in OpenFile (`host/nn_fs.c`) and the `HOS_VIDEO_NO_CATCHUP` read in `dec_acquire` (`host/nn_video.c`). #2301 and #2322, merged two and three minutes later, each added a test that cuts one of those functions out of its source file and compiles it in a small fixture. Both were written on a base before #2042, so their fixtures have no `env_on`. Each pull request passed alone. Only the two tests' fixtures are affected. The source files themselves include `gpu/vk/compat/hookshot_env.h`. ### Where in the game Not in the game: `tests/run_all.py` ### How often Every time ### Graphics card and driver version Does not depend on it ### CPU, RAM and operating system i9-13900K, 96 GB, Arch Linux (g++ 16.2.1, clang++ 23.1.3) ### Changes from the defaults None ### Other software hooked into the game No, nothing hooked in ### hookshot_log.txt Not a game run. To see it: `python3 tests/test_vcore_video_acquire.py --cxx clang++` on main. ### Claim I'm working on this: the fix is one line in each test's compile command (the header with `-include`, as #2042 did for the other tests that cut code out of the sources), and its pull request follows this issue directly. By `docs/workflow.md`: S2 (main's checks are red for everyone) x reach (every contributor's run) x confidence 1.0 (cause found, fix passes with both compilers here).

Commenting is not possible because the repository is archived.