julia-ci.yml reads the ci-timings branch but nothing writes it — LPT planning has never activated

[sotashimozono]

julia-ci.yml reads the ci-timings orphan branch but contains no job that writes it, so the
timings-informed LPT planning it advertises can never activate for a caller.

The header says:

timings-informed LPT planning kicks in automatically when a ci-timings orphan branch with
timings.tsv exists (else deterministic round-robin).

and the plan job implements the read (lines 60–67). Nothing in the file produces timings.tsv.
The test job does not upload per-shard timings and there is no record-timings job.

Measured, 2026-07-30

caller	shards	ci-timings branch
QAtlasHub/AbstractQAtlas.jl	4	absent
QAtlasHub/DataVault.jl	1	absent
QAtlasHub/ParamIO.jl	1	absent
lab-sotashimozono/ParaLinearAlgebra.jl (inline CI, has record-timings)	8	present, updated 2026-07-30T03:33:30Z

shards: 1 needs no timings, so DataVault and ParamIO are unaffected. AbstractQAtlas is not: it
fans out four ways and its planner has been round-robin on every run since adoption, with no path to
ever becoming anything else. The failure is silent — round-robin is a legitimate documented fallback,
so the log line "no ci-timings branch yet — round-robin fallback" looks like a first-run state
forever rather than a permanent one.

Fix

Port the producer half from ParaLinearAlgebra.jl's inline CI.yml (jobs test upload +
record-timings, lines 110–160):

• test uploads test/.ci-out/timings-*.tsv as an artifact per shard;
• a record-timings job downloads them, merges into the existing timings.tsv (latest value wins),
  and force-pushes the ci-timings orphan branch. Needs contents: write.

Gate it on inputs.shards > 1 so unsharded callers pay nothing, and on github.ref == 'refs/heads/main'
so PR runs do not race each other for the branch.

Why this blocks a larger migration

lab-sotashimozono/.github has no julia-ci, and 45 repos across the three owners still carry an
inline CI.yml (6016 lines total) that this reusable is meant to absorb — see
lab-sotashimozono/.github#15. ParaLinearAlgebra's inline version does record timings, so adopting
the reusable as it stands would be a regression for it and for every other sharded repo. Fixing the
producer here first, then porting, keeps the known defect from being copied into a second hub.

Not a gap

The compat job (minimum LTS + latest) has no counterpart here, but it does not need one:
AbstractQAtlas.jl already expresses it as a second call with julia-version: "1.10", which is the
better factoring. Only the timings producer is missing.

[sotashimozono]

Fixed on main by #20 (merged 2026-07-30T07:10Z). Not yet delivered to any caller, and one correction to the issue above.

Correction: three callers, not four

The issue listed QAtlasHub/.github among the callers. It is not one — my scan matched the uses:
example inside julia-ci.yml's own header comment. Verified: only DataVault.jl, ParamIO.jl and
AbstractQAtlas.jl call it. This does not change the finding (the table's per-caller rows were
correct), but it does change what merging accomplishes — see below.

Delivery needs the v1 tag moved

Every caller references julia-ci.yml@v1, and v1 is a moving tag at v1.0.3 == the previous
main. So main now has the fix and no caller sees it. Nothing here calls @main, so the merge
exercised no code path at all.

Delivering it is the established release step: tag v1.0.4 at the new main and move v1 to it.
I have deliberately not done that — it is a fleet-wide release across ~20 repos in three orgs, and
it is the kind of step worth being asked for rather than assumed. Risk is low if someone wants it:
the change is purely additive, the producer is gated on shards > 1 and push:main, and
record-timings is continue-on-error, so the worst case is an extra warning line or a skipped
bookkeeping job.

It cannot be verified before that

record-timings only runs on push:main, so no PR anywhere can exercise it. The first real proof
will be AbstractQAtlas.jl (shards: 4) growing a ci-timings branch on its next merge to main
after v1 moves. Worth stating plainly rather than calling the fix confirmed — right now it is
lint-clean and reasoned, not observed.