MagicBeanDev/Longshot

Watch

Fork 19

fix(vk): on Most the Builder's progress bar leaves the prediction's part out; its check runs All and Most #164

Merged

HumanBean merged 1 commit from HumanBean/Longshot:fix/prepare-bar-most into main

2026-11-13 08:23:56 +00:00

HumanBean commented

2026-11-13 08:23:19 +00:00

Owner

Why

Since #158 the default is Most, and main fails one test: tests/test_prepare_cache_save.py compiles the Builder's preparation path with stubs and has no compile_all(). Testing Most also found that the Builder's progress bar jumps by the prediction's share on Most after a fresh preparation.

What changed

vk: started_up() in gpu/vk/vk_main.cpp (the Builder's --prepare-shaders path) takes the predicted pipelines' part out of the progress bar whenever the prediction does not run. Before, it did so only when the seed was already prepared, so on Most after a fresh preparation the step to the noted pipelines counted that part done at its guess.
tests: tests/test_prepare_cache_save.py gets compile_all() and runs every case with All and with Most: what runs (All predicts after a fresh seed, Most never) and the bar's values.

Flags and settings

None.

Measurements

The fix's measurement is its test: a check that fails on main and passes here.

main: 1d6b832a (2026-11-13)
this PR: fix/prepare-bar-most @ 539b7c14
Hardware: RTX 3090, Linux, 24 threads
Method: python -I -B tests/run_all.py

	main @ `1d6b832a`	this PR @ `539b7c14`
tests/run_all.py	471 passed, 1 failed (test_prepare_cache_save: `compile_all` undeclared)	472 passed, 0 failed, 65 skipped
bar on Most, fresh seed, at half the noted pipelines	0.714 (prediction's part counted done)	0.600

Testing

Tested: the full game-free suite on Linux, as above; the new cases of the check on All and on Most.

Not tested: the bar in the Builders' windows by eye; a Windows build (the change is four lines of portable C++ in the renderer, no host code or paths).

Needs other testers: None (the change is the bar's arithmetic, which the check covers).

Screenshots

Nothing on screen beyond the bar's value, which the check covers.

Risk and rollback

At worst the Builder's bar is off; the preparation itself does the same work. Reverting the commit is enough.

Issues

See #158

Checklist

For everyone

One topic.
The title is a Conventional Commit (type(scope): description).
Rebased on current main.
python -I -B tests/run_all.py run and passing (any failure that also fails on main is named above).
Tested in the game, or what wasn't is named under Testing.
Testing filled in: what was tested, what wasn't, and what needs other testers (with test requests).
Measured: main and this PR at named commits, on the same PC, scene and settings, at least 2 interleaved runs each, the median reported.
No game files, dumps, build output or generated files.
third_party/ untouched (a new folder with a PROVENANCE.md instead).
The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled.

If applicable

AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
A game-free test added for the fix, failing on main and passing here.
A check under tests/ that reads changed source updated in this PR, not weakened.

## Why Since #158 the default is Most, and main fails one test: `tests/test_prepare_cache_save.py` compiles the Builder's preparation path with stubs and has no `compile_all()`. Testing Most also found that the Builder's progress bar jumps by the prediction's share on Most after a fresh preparation. ## What changed - vk: `started_up()` in `gpu/vk/vk_main.cpp` (the Builder's `--prepare-shaders` path) takes the predicted pipelines' part out of the progress bar whenever the prediction does not run. Before, it did so only when the seed was already prepared, so on Most after a fresh preparation the step to the noted pipelines counted that part done at its guess. - tests: `tests/test_prepare_cache_save.py` gets `compile_all()` and runs every case with All and with Most: what runs (All predicts after a fresh seed, Most never) and the bar's values. ## Flags and settings None. ## Measurements The fix's measurement is its test: a check that fails on `main` and passes here. main: 1d6b832a (2026-11-13) this PR: fix/prepare-bar-most @ 539b7c14 Hardware: RTX 3090, Linux, 24 threads Method: `python -I -B tests/run_all.py` | | main @ 1d6b832a | this PR @ 539b7c14 | |---|---|---| | tests/run_all.py | 471 passed, 1 failed (test_prepare_cache_save: `compile_all` undeclared) | 472 passed, 0 failed, 65 skipped | | bar on Most, fresh seed, at half the noted pipelines | 0.714 (prediction's part counted done) | 0.600 | ## Testing **Tested:** the full game-free suite on Linux, as above; the new cases of the check on All and on Most. **Not tested:** the bar in the Builders' windows by eye; a Windows build (the change is four lines of portable C++ in the renderer, no host code or paths). **Needs other testers:** None (the change is the bar's arithmetic, which the check covers). ## Screenshots Nothing on screen beyond the bar's value, which the check covers. ## Risk and rollback At worst the Builder's bar is off; the preparation itself does the same work. Reverting the commit is enough. ## Issues See #158 ## Checklist ### For everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`). - [x] Rebased on current `main`. - [x] `python -I -B tests/run_all.py` run and passing (any failure that also fails on `main` is named above). - [x] Tested in the game, or what wasn't is named under Testing. - [x] Testing filled in: what was tested, what wasn't, and what needs other testers (with test requests). - [x] Measured: main and this PR at named commits, on the same PC, scene and settings, at least 2 interleaved runs each, the median reported. - [x] No game files, dumps, build output or generated files. - [x] `third_party/` untouched (a new folder with a `PROVENANCE.md` instead). - [x] The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled. ### If applicable - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [x] A game-free test added for the fix, failing on `main` and passing here. - [x] A check under `tests/` that reads changed source updated in this PR, not weakened.

HumanBean added 1 commit

2026-11-13 08:23:19 +00:00

fix(vk): on Most the Builder's progress bar leaves the prediction's part out; its check runs All and Most

CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Has been skipped

Details

CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Has been skipped

Details

CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Has been skipped

Details

Third-party parts / The same as upstream (pull_request) Has been skipped

Details

CI / Conventional Commit title (pull_request) Successful in 5s

Details

CI / Workflow lint (actionlint) (pull_request) Successful in 6s

Details

Third-party parts / As their notes say (pull_request) Successful in 13s

Details

PR description / Description follows the template (pull_request_target) Successful in 9s

Details

539b7c14b3

The Builder's preparation (vk_main.cpp, --prepare-shaders) took the predicted pipelines' part out of the bar only when the
seed was already prepared. On Most after a preparation the prediction does not run, and the step to the noted pipelines
counted its part done at its guess: the bar jumped by the prediction's share. Now it is skipped whenever the prediction
does not run.

tests/test_prepare_cache_save.py compiles that path with stubs and lacked compile_all (feat(vk): Compile shaders): it now
runs each case with All and with Most, checking what runs (All predicts after a fresh seed, Most never) and the bar's
progress (Most, fresh seed: 0.600 at half the noted pipelines, not the 0.714 with the prediction's part counted done).

Co-Authored-By: Claude Fable 5.2 <[email protected]>

HumanBean changed title from ~~WIP: fix(vk): on Most the Builder's progress bar leaves the prediction's part out; its check runs All and Most~~ to fix(vk): on Most the Builder's progress bar leaves the prediction's part out; its check runs All and Most

2026-11-13 08:23:35 +00:00

HumanBean merged commit 3ece96e942 into main

2026-11-13 08:23:56 +00:00

HumanBean deleted branch fix/prepare-bar-most

2026-11-13 08:23:56 +00:00

HumanBean referenced this pull request from a commit

2026-11-13 08:23:57 +00:00

fix(vk): on Most the Builder's progress bar leaves the prediction's part out; its check runs All and Most (#164)

HumanBean referenced this pull request

2026-11-13 08:44:37 +00:00

feat(prepare): Compile shaders All / Most / None, Most skipping the predicted pipelines #158