docs: credit NotRexed for Hookshot, and README.md, docs/BAKE.md, no changelog.txt #11

Merged

HumanBean merged 2 commits from HumanBean/Longshot:docs/credit-hookshot into main

2026-11-12 22:56:18 +00:00

HumanBean commented

2026-11-12 22:25:56 +00:00

Owner

Credits and history

README: a line under the intro: Longshot is a fork of Hookshot, which NotRexed made, and the history up to the fork is Hookshot's. The Credits section opens with the same statement and credits NotRexed for Hookshot (plus BlueShell, Skybound and the a64-kit tools). EvilOOT's link goes back to evilmfkr/Hookshot (evilmfkr/Longshot is a 404).
License (README): it says Longshot is a modified version of Hookshot under the same GPL-3.0-or-later, with two notices: Hookshot: Copyright (C) 2026 NotRexed and the Hookshot contributors and Longshot: Copyright (C) 2026 the Longshot contributors. LICENSE is the GPL's text, which carries no project copyright line and may not be changed, so it stays as it is.
CHANGELOG.md: a note at the top saying the entries up to 2.2.0-rc.1 are Hookshot's releases. The four entries whose text the rename changed (hookshot_log.txt, hookshot-k10, "what Hookshot connects to", hookshot_env.txt) are back as they were released. The note goes above the first ## [ heading, so release-please and release_promote.sh still insert new entries below it.
Other places the rename rewrote history: host/oot_heap_retry.c, nx2/oot_patches.py and tests/test_oot_heap_retry.py say TomShafford/Hookshot (a Hookshot fork) again, since TomShafford/Longshot doesn't exist. host/PROVENANCE.md and gpu/PROVENANCE.md say where code came from, so they say Hookshot again. The same goes for the measured runs of Hookshot builds in docs/cpu-profile.md and docs/saves.md, the Fritchii recipe note ("Not recorded by Hookshot"), the vk_vendor.h comment and two comments in tests/test_menu_size.py that quote old version strings.
docs/development.md: the third_party/ row now says a64-kit's changes are Hookshot's, "marked (longshot) since the rename".

Product names, binaries, paths, env vars and the (longshot) markers keep the Longshot name.

Cleanup

readme.md is renamed to README.md. Updated: CONTRIBUTING.md, docs (tested-configurations, saves, development), the three tests that read the option table, scripts/linux/appimage.sh (copies it into the AppImage), the two workflows' comments, and the opening comment of three host_shaders/*_spv.h. scripts/host_shaders.py keeps that comment when it regenerates the headers, and --check still reports 4 matches.
BAKE.md moves to docs/BAKE.md. Updated: its own table, gpu/shaderbatch.cpp, nx2/bake_mimo.py, tests/test_bake_ds.py and tests/test_bake_skeleton.py.
changelog.txt is deleted. Only docs mentioned it, and nothing built, shipped or read it. docs/versioning.md and the 2.0.14 entry in CHANGELOG.md now link Hookshot's own copy (NotRexed/Hookshot/blob/main/changelog.txt).

Overlap with the release/updater work: the only release-side files touched are docs/versioning.md (the changelog.txt paragraph) and comment lines in linux-appimage.yml and windows-builder.yml. The code that still points at NotRexed/Longshot (the update checks, the bug report and pipeline-share issue links, recomplauncher.json, the MSI's About link, -R NotRexed/Longshot in versioning.md) is left for that work.

Testing

python -I -B tests/run_all.py: 450 passed, 7 failed, 71 skipped, both before and after this change, with the same 7 failures. Five of them come from the sandbox's read-only /tmp and pass when run on their own outside it (test_program_cache_cleanup, test_program_cache_one_variant, test_control, test_longshot_ctl, test_vram_budget_row). The other two already fail on main: test_bake_portability (no zstandard module here) and test_builder_log_read_only (no libtk8.6). The checks that read README.md (test_readme_menu_rows, test_stick_dead_zone, test_camera_speed) pass, and so do test_host_shaders, test_oot_heap_retry and test_menu_size.

Docs, comments and file names only, with nothing on screen, so no screenshots and no run in the game.

The rename to Longshot also rewrote NotRexed's credit ("NotRexed: Longshot, BlueShell, ...") and parts of the project's history, as if Longshot were the original. Longshot is a fork of NotRexed's [Hookshot](https://github.com/NotRexed/Hookshot), and the readme, credits and history should say so. ## Credits and history - **README:** a line under the intro: Longshot is a fork of Hookshot, which NotRexed made, and the history up to the fork is Hookshot's. The Credits section opens with the same statement and credits **NotRexed** for Hookshot (plus BlueShell, Skybound and the a64-kit tools). EvilOOT's link goes back to `evilmfkr/Hookshot` (`evilmfkr/Longshot` is a 404). - **License (README):** it says Longshot is a modified version of Hookshot under the same GPL-3.0-or-later, with two notices: `Hookshot: Copyright (C) 2026 NotRexed and the Hookshot contributors` and `Longshot: Copyright (C) 2026 the Longshot contributors`. `LICENSE` is the GPL's text, which carries no project copyright line and may not be changed, so it stays as it is. - **CHANGELOG.md:** a note at the top saying the entries up to 2.2.0-rc.1 are Hookshot's releases. The four entries whose text the rename changed (`hookshot_log.txt`, `hookshot-k10`, "what Hookshot connects to", `hookshot_env.txt`) are back as they were released. The note goes above the first `## [` heading, so release-please and `release_promote.sh` still insert new entries below it. - **Other places the rename rewrote history:** `host/oot_heap_retry.c`, `nx2/oot_patches.py` and `tests/test_oot_heap_retry.py` say `TomShafford/Hookshot (a Hookshot fork)` again, since `TomShafford/Longshot` doesn't exist. `host/PROVENANCE.md` and `gpu/PROVENANCE.md` say where code came from, so they say Hookshot again. The same goes for the measured runs of Hookshot builds in `docs/cpu-profile.md` and `docs/saves.md`, the Fritchii recipe note ("Not recorded by Hookshot"), the `vk_vendor.h` comment and two comments in `tests/test_menu_size.py` that quote old version strings. - **`docs/development.md`:** the `third_party/` row now says a64-kit's changes are Hookshot's, "marked `(longshot)` since the rename". Product names, binaries, paths, env vars and the `(longshot)` markers keep the Longshot name. ## Cleanup - `readme.md` is renamed to `README.md`. Updated: CONTRIBUTING.md, docs (tested-configurations, saves, development), the three tests that read the option table, `scripts/linux/appimage.sh` (copies it into the AppImage), the two workflows' comments, and the opening comment of three `host_shaders/*_spv.h`. `scripts/host_shaders.py` keeps that comment when it regenerates the headers, and `--check` still reports 4 matches. - `BAKE.md` moves to `docs/BAKE.md`. Updated: its own table, `gpu/shaderbatch.cpp`, `nx2/bake_mimo.py`, `tests/test_bake_ds.py` and `tests/test_bake_skeleton.py`. - `changelog.txt` is deleted. Only docs mentioned it, and nothing built, shipped or read it. `docs/versioning.md` and the 2.0.14 entry in CHANGELOG.md now link Hookshot's own copy (`NotRexed/Hookshot/blob/main/changelog.txt`). **Overlap with the release/updater work:** the only release-side files touched are `docs/versioning.md` (the changelog.txt paragraph) and comment lines in `linux-appimage.yml` and `windows-builder.yml`. The code that still points at `NotRexed/Longshot` (the update checks, the bug report and pipeline-share issue links, `recomplauncher.json`, the MSI's About link, `-R NotRexed/Longshot` in versioning.md) is left for that work. ## Testing `python -I -B tests/run_all.py`: 450 passed, 7 failed, 71 skipped, both before and after this change, with the same 7 failures. Five of them come from the sandbox's read-only `/tmp` and pass when run on their own outside it (`test_program_cache_cleanup`, `test_program_cache_one_variant`, `test_control`, `test_longshot_ctl`, `test_vram_budget_row`). The other two already fail on main: `test_bake_portability` (no `zstandard` module here) and `test_builder_log_read_only` (no `libtk8.6`). The checks that read README.md (`test_readme_menu_rows`, `test_stick_dead_zone`, `test_camera_speed`) pass, and so do `test_host_shaders`, `test_oot_heap_retry` and `test_menu_size`. Docs, comments and file names only, with nothing on screen, so no screenshots and no run in the game.

HumanBean added 1 commit

2026-11-12 22:25:56 +00:00

docs: credit NotRexed for Hookshot, and README.md, docs/BAKE.md, no changelog.txt 72345cf66b

The rename to Longshot also rewrote NotRexed's credit ("NotRexed: Longshot") and
parts of the project's history. Longshot is a fork of Hookshot, so say so.

- README: Longshot is a fork of NotRexed's Hookshot (top and Credits); NotRexed
  is credited for Hookshot; EvilOOT's link back to evilmfkr/Hookshot; the License
  section names Hookshot's copyright next to the Longshot contributors'.
- History back to Hookshot: CHANGELOG.md's four renamed entries and a note that
  the entries up to 2.2.0-rc.1 are Hookshot's; TomShafford/Hookshot (a Hookshot
  fork) in oot_heap_retry; the PROVENANCE notes; measured runs of Hookshot builds
  in docs/cpu-profile.md and docs/saves.md; the Fritchii recipe's note.
- readme.md is README.md, BAKE.md is docs/BAKE.md, with every reference.
- changelog.txt is gone: Hookshot's 0.1.0 changelog is linked in its repository.

HumanBean added 1 commit

2026-11-12 22:45:08 +00:00

fix: links to Longshot's own issues, and app IDs of its own 4e495812e8

The bug report, the pipeline share and the Windows installer still sent players to
github.com/NotRexed/Longshot, which does not exist, and the app IDs put Longshot
under NotRexed's name.

- Bug report: the issues and the Game problem form on git.magicbean.dev (Forgejo
  names a template by its path: ?template=.github/ISSUE_TEMPLATE/1-game-problem.yml).
- Save my pipeline list: Longshot's own "Submit your pipelines" issue (#14), in the
  menu, its translations, the Windows builder, nx2/pipeline_share.py and the docs. The
  lists sent before the fork keep their link to Hookshot's #2034.
- The MSI's about link: the repository on git.magicbean.dev.
- App IDs: dev.magicbean.Longshot (the window class and the .desktop file's
  StartupWMClass) and MagicBean.Longshot.Builder (Windows' AppUserModelID).

The updater and release links are left to the release rework.

HumanBean merged commit d3bd280503 into main

2026-11-12 22:56:18 +00:00

HumanBean deleted branch docs/credit-hookshot

2026-11-12 22:56:18 +00:00

HumanBean referenced this pull request from a commit

2026-11-12 22:56:19 +00:00

docs: credit NotRexed for Hookshot, and README.md, docs/BAKE.md, no changelog.txt (#11)

~~HumanBean referenced this pull request 2026-11-12 23:25:40 +00:00~~

feat(release)!: releases on Forgejo, signed by one key; tools.lock; Forgejo Actions only #13