fix(ci): align the Idris2 version assert with the pinned image (0.8.0) - #31
Merged
Conversation
With the assert made diagnosable, the actual mismatch is visible:
idris2 --version -> Idris 2, version 0.8.0-6ca00e72e
::error::expected Idris 2 0.7.0
The digest-pinned idris2-pack image ships 0.8.0, so the 0.7.0 assertion could
never have matched and this job has never passed on its own terms.
Nothing in the repository actually requires 0.7.0: abi.ipkg declares no version
constraint (it depends only on `base`), and the `idris2 0.7.0` line in
.tool-versions is commented out, so it pins nothing. The only 0.7.0 references
were the workflow asserts and their comments — stale documentation that drifted
from the digest.
The immutable digest is the authority, so the version notes are brought into
step with it rather than the other way round. The typecheck now actually runs.
Also fixes the identical brittle assert in release.yml, which piped
`idris2 --version` into `grep -Fx` under pipefail exactly as ci.yml did.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
hyperpolymath
added a commit
that referenced
this pull request
Jul 27, 2026
…ns (#32) `CI / required` is the last red check on `main`. After #31 fixed the version assert, the job got two steps further and now fails at **Validate repository contracts** with exit code 2 and one line of output: ``` jq is required ``` ## Root cause `tests/check-examples.sh:15`: ```bash command -v jq >/dev/null || { echo "jq is required"; exit 2; } ``` The pinned `idris2-pack` image does not ship `jq`. So the **Trope IR conformance test — the one asserting every example lowers to valid Trope IR — has never actually executed in CI.** It bailed before testing anything. That makes this more than a build fix: it turns a step that always died early into one that genuinely exercises the language's core guarantee. ## Fix Same class of problem as the missing `xz` in #28, so the step now handles both: collect what is absent, install in a single `apt-get` pass, print both versions so the log records what ran. A no-op on images that already have them. ## Verified locally against `origin/main` (d29f89e) ``` bash -n over tests/ scripts/ ............... all parse tests/check-examples.sh .................... exit 0 → "examples: all lower to valid Trope IR" tests/aspect_tests.sh ...................... exit 0 → PASS=3 FAIL=0 WARN=0 scripts/check-root-shape.sh ................ exit 0 → 45 entries, 46 permitted tests/workflows/validate_workflows_test.sh . exit 0 → Workflow validation PASSED ``` Note `check-examples.sh` skips the verdict round-trip when no sibling `tropecheck` binary is built — that part still needs `trope-checker`, and is reported as a `note`, not a pass. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.



Now that #30 made the assert diagnosable, the real mismatch is visible:
The digest-pinned
idris2-packimage ships 0.8.0, so the0.7.0assertioncould never have matched —
CI / requiredhas never passed on its own terms.This is the last red check on
main.Which side is wrong?
Nothing in the repository actually requires 0.7.0:
abi.ipkgdeclares no version constraint — it depends only onbase.idris2 0.7.0line in.tool-versionsis commented out, so it pins nothing.So "0.7.0" was stale documentation that drifted from the digest. The immutable
digest is the authority, and the version notes are brought into step with it
rather than the reverse. The typecheck now actually runs against the toolchain
that is genuinely installed.
Also fixes the identical brittle assert in
release.yml, which pipedidris2 --versionintogrep -Fxunderpipefailexactly asci.ymldid — itwould have failed the same way, silently, on the next release.
If 0.7.0 was intended
Then the fix is the other direction: re-pin the container to a 0.7.0 image digest.
Say so and I will swap it — but the current pair is contradictory either way.
🤖 Generated with Claude Code