Skip to content

fix(ci): restore missing build.zig — CI fails at step one on every run - #615

Merged
gHashTag merged 2 commits into
mainfrom
fix/restore-build-zig
Aug 8, 2026
Merged

fix(ci): restore missing build.zig — CI fails at step one on every run#615
gHashTag merged 2 commits into
mainfrom
fix/restore-build-zig

Conversation

@gHashTag

@gHashTag gHashTag commented Aug 8, 2026

Copy link
Copy Markdown
Owner

Problem

CI / Build & Test has failed 12 out of the last 12 runs, always at the very first step:

zig build author-guard -Dci=true
error: no build.zig file found, in the current directory or any parent directories
##[error]Process completed with exit code 1

Because it dies immediately, every genuine problem in the build, tests, format check and schema drift gates is invisible — the workflow never reaches them. This masked five separate site-deploy defects I only found by inspecting runs one at a time (#611, #612, #614).

Cause

main has no build.zig, while it does have:

  • build.zig.zon (the Zig package manifest) at the root
  • ~2815 .zig sources, including everything ci.yml references (src/vm.zig, src/trinity.zig, src/tri/job_system.zig, .trinity/registry.json)

So this is a Zig project root missing only its build script. build.zig still exists on other branches — including every dependabot branch, which forked from an older main — which points at the cleanup/git-history-reduce-6gb-to-220mb history rewrite as where it was dropped.

Change

Restored build.zig from cleanup/git-history-reduce-6gb-to-220mb.

⚠️ Honest caveat — please look before merging

That branch is dated 2026-03-18; main is at 2026-08-05. The restored build script may not match five months of source changes, so CI may now fail differently — at a real error rather than at a missing file. That is still progress, but it is not automatically "fixed".

I deliberately did not verify locally: my toolchain is Zig 0.16 while CI pins 0.15.2, so a local result would be measured with the wrong instrument and could mislead. The authoritative check is this PR's own CI run.

Decision rule: if Build & Test gets past the author-guard step, this is a net improvement and worth merging (with follow-up for whatever it now reveals). If it fails in a way that suggests the wrong build script, the right fix is to commit the current build.zig from your working tree instead — you have one there, tracked on feat/tri-net-skill-radio-ops.

🤖 Generated with Claude Code

Dmitrii Vasilev and others added 2 commits August 8, 2026 15:41
…step one

The CI workflow is red on 12 of the last 12 runs, always at the same place:

    zig build author-guard -Dci=true
    error: no build.zig file found, in the current directory or any parent
    ##[error]Process completed with exit code 1

Cause: `build.zig` is absent from `main`, while `build.zig.zon` and ~2815 .zig
sources are present — so the repository is a Zig project root missing only its
build script. The file still exists on other branches (including every
dependabot branch, which forked from an older main), which points at the
`cleanup/git-history-reduce-6gb-to-220mb` history rewrite as where it was lost.

Restored from that branch. Note it dates from 2026-03-18 while main has moved on
to 2026-08-05, so it may need follow-up to match the current source tree — but
without it CI cannot even start, and every real failure stays invisible behind
this one.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
With build.zig restored the build script now compiles and runs — and reveals the
second half of the problem:

    no step named 'author-guard'
    access the help menu with 'zig build -h'

`ci.yml` invokes `zig build author-guard`, but no build.zig on any branch of this
repository defines that step (checked: cleanup/git-history-reduce-6gb-to-220mb,
feat/tri-net-skill-radio-ops, and the working tree — zero matches in all three).
The gate has therefore never been able to pass, and because it runs first it hid
the build, tests, format check and schema-drift gates behind it.

Marked continue-on-error so the pipeline reaches its real checks. Deliberately
NOT deleted: it encodes a policy someone intended, and retiring it should be the
owner's decision, not a silent side effect of fixing CI.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@gHashTag
gHashTag merged commit ce004b3 into main Aug 8, 2026
14 of 26 checks passed
@gHashTag
gHashTag deleted the fix/restore-build-zig branch August 8, 2026 09:05
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant