The SDK could not be published at all — bun publish runs prepare, and prepare needs bun2nix #115

Merged

enricobuehler merged 1 commits from worktree-sdk-publish-prepare-hook into main 2026-08-08 11:04:34 +00:00

Conversation 0 Commits 1 Files Changed 1

+1 -1

enricobuehler commented 2026-08-08 10:56:05 +00:00

Owner

Copy Link

Copy Source

sdk-v0.1.3 failed at the publish step: bun2nix: command not found, exit 127. Nothing reached the registry, so 0.1.3 is still free — @punktfunk/host is still at 0.1.2.

What happened

bun publish runs the prepare lifecycle script, and sdk's prepare is bun2nix -o bun.nix — regenerating the nix dependency file. bun2nix is a repo-maintenance tool, not something the oven/bun:1 publish container has; the workflow's own install is --ignore-scripts, so nothing put it on PATH either.

bun publish v1.3.14
$ bun run build
$ tsc -p tsconfig.build.json
$ bun2nix -o bun.nix
/usr/bin/bash: line 1: bun2nix: command not found
error: script "prepare" exited with code 127

This was latent, not new

prepare gained the bun2nix call on 2026-07-27 (1db8f763, "move the bun packages to bun2nix"), while the last SDK publish was 0.1.2, bumped 2026-07-20. So the hook has been broken for every SDK release since it landed — 0.1.3 is just the first one to try it.

@punktfunk/plugin-kit has no prepare at all, which is why kit 0.3.2 published fine inside that window and this stayed hidden.

Why not just copy web

web/package.json does the same job from postinstall. That is correct for web — it is never published — and would be worse here: a published package's postinstall runs in every consumer's install, so every plugin depending on @punktfunk/host would try to run bun2nix and fail.

prepare is the right hook for a published package (it does not run for consumers). It just must not assume a repo-maintenance tool exists wherever a publish happens.

The fix

The script skips when bun2nix is absent — and only then:

"prepare": "if command -v bun2nix >/dev/null 2>&1; then bun2nix -o bun.nix; fi"

A present-but-failing bun2nix still fails the script. Swallowing that (|| true) would let a publish go out with a silently stale bun.nix, which is the exact hand-maintained-hash problem 1db8f763 set out to end.

Verification

Both directions measured against the same sh -e that bun and the Gitea runner use for lifecycle steps:

case	exit	want
bun2nix absent (the publish container)	0	0
bun2nix present but failing	3	non-zero

bun publish --dry-run now completes and reports + @punktfunk/host@0.1.3, and the packed dist/ui.js:139 carries the category forwarding that is the point of the 0.1.3 release.

After merge

I will delete and re-push sdk-v0.1.3 (it published nothing, so the version is not burned), confirm the run goes green and that 0.1.3 actually appears on the registry, then tag plugin-kit-v0.3.3 — that order matters, since kit 0.3.3's peer range requires @punktfunk/host ^0.1.3 and publishing it first would leave a window where the kit is uninstallable.

`sdk-v0.1.3` failed at the publish step: **`bun2nix: command not found`, exit 127**. Nothing reached the registry, so **0.1.3 is still free** — `@punktfunk/host` is still at 0.1.2. ## What happened `bun publish` runs the `prepare` lifecycle script, and sdk's `prepare` is `bun2nix -o bun.nix` — regenerating the nix dependency file. bun2nix is a repo-maintenance tool, not something the `oven/bun:1` publish container has; the workflow's own install is `--ignore-scripts`, so nothing put it on PATH either. ``` bun publish v1.3.14 $ bun run build $ tsc -p tsconfig.build.json $ bun2nix -o bun.nix /usr/bin/bash: line 1: bun2nix: command not found error: script "prepare" exited with code 127 ``` ## This was latent, not new `prepare` gained the bun2nix call on **2026-07-27** (`1db8f763`, *"move the bun packages to bun2nix"*), while the last SDK publish was 0.1.2, bumped **2026-07-20**. So the hook has been broken for **every** SDK release since it landed — 0.1.3 is just the first one to try it. `@punktfunk/plugin-kit` has no `prepare` at all, which is why kit 0.3.2 published fine inside that window and this stayed hidden. ## Why not just copy web `web/package.json` does the same job from **`postinstall`**. That is correct for web — it is never published — and would be *worse* here: a published package's `postinstall` runs in every **consumer's** install, so every plugin depending on `@punktfunk/host` would try to run bun2nix and fail. `prepare` is the right hook for a published package (it does not run for consumers). It just must not assume a repo-maintenance tool exists wherever a publish happens. ## The fix The script skips when bun2nix is absent — and **only** then: ``` "prepare": "if command -v bun2nix >/dev/null 2>&1; then bun2nix -o bun.nix; fi" ``` A present-but-failing bun2nix still fails the script. Swallowing that (`|| true`) would let a publish go out with a silently stale `bun.nix`, which is the exact hand-maintained-hash problem `1db8f763` set out to end. ## Verification Both directions measured against the same `sh -e` that bun and the Gitea runner use for lifecycle steps: | case | exit | want | |---|---|---| | bun2nix absent (the publish container) | 0 | 0 | | bun2nix present but failing | 3 | non-zero | `bun publish --dry-run` now completes and reports `+ @punktfunk/host@0.1.3`, and the packed `dist/ui.js:139` carries the `category` forwarding that is the point of the 0.1.3 release. ## After merge I will delete and re-push `sdk-v0.1.3` (it published nothing, so the version is not burned), confirm the run goes green and that 0.1.3 actually appears on the registry, then tag `plugin-kit-v0.3.3` — that order matters, since kit 0.3.3's peer range requires `@punktfunk/host ^0.1.3` and publishing it first would leave a window where the kit is uninstallable.

enricobuehler added 1 commit 2026-08-08 10:56:08 +00:00

fix(ci/sdk): the SDK could not be published at all — bun publish runs prepare ... 1df39d9617

`sdk-v0.1.3` failed at the publish step with `bun2nix: command not found`, exit 127.
Nothing was published, so 0.1.3 is still free.

`bun publish` runs the `prepare` lifecycle script, and sdk's `prepare` is
`bun2nix -o bun.nix` — regenerating the nix dependency file. That tool is a
devDependency of the repo, not something the `oven/bun:1` publish container has, and
the workflow's own install is `--ignore-scripts`, so nothing put it on PATH either.

This was latent, not new. `prepare` gained the bun2nix call on 2026-07-27 (1db8f763,
"move the bun packages to bun2nix"), while the last SDK publish was 0.1.2, bumped
2026-07-20. So the hook has been broken for every SDK release since it landed, and
0.1.3 is simply the first one to try. `@punktfunk/plugin-kit` has no `prepare` and was
never affected, which is why kit 0.3.2 published fine in that window and hid this.

The fix is NOT to copy `web/package.json`, which does the same job from `postinstall`.
That is right for web — it is never published — and would be worse here: a published
package's `postinstall` runs in every CONSUMER's install, so every plugin depending on
`@punktfunk/host` would try to run bun2nix and fail. `prepare` is the correct hook for
a published package (it does not run for consumers); it just must not assume a
repo-maintenance tool exists wherever a publish happens.

So the script skips when bun2nix is absent — and ONLY then. A present-but-failing
bun2nix still fails the script, because swallowing that would publish with a silently
stale bun.nix, which is the exact hand-maintained-hash problem 1db8f763 set out to end.
Both directions measured against the same `sh -e` bun and the Gitea runner use:
absent → exit 0, present-and-failing → exit 3.

`bun publish --dry-run` now completes and reports `+ @punktfunk/host@0.1.3`.

enricobuehler merged commit 27ceab2f6c into main 2026-08-08 11:04:34 +00:00

enricobuehler deleted branch worktree-sdk-publish-prepare-hook 2026-08-08 11:04:41 +00:00

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

No items

No labels

Milestone

No items

No Milestone

Projects

Clear projects

No projects

Assignees

Clear assignees

FernandoMX enricobuehler (Enrico Bühler) kszlim nannku neoney sassycrown

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: unom/punktfunk#115