Bump external/wirejs-native to merged master post-#4#13
Conversation
Wire-Network/wirejs-native#4 merged on 2026-04-21 with three additional commits of review feedback applied after this repo's abieos #12 merged on 2026-04-20. The submodule pointer here was pinned at df218917 — the initial push of wirejs-native's feature branch, before any review cleanup. Now that wirejs-native master is at b096df42 (merge of #4) and the feature branch has been deleted, retarget the pointer so fresh clones pick up the reviewed-and-merged state instead of a dangling pre-review commit. Picked up between df218917 and b096df42: - orphaned legacy kv_table* types removed - int64 enum width docs - get_table_rows unification (drop get_kv_table_rows, add find/index_name)
dtaghavi
left a comment
dtaghavi
left a comment
There was a problem hiding this comment.
Review
Verified the submodule bump and every claim in the description against the remote state.
Correctness — all checks out
b096df42is the currentHEADofwirejs-native/master(confirmed via branches API).feature/wire-sysio-table-idhas been deleted upstream — onlymasterandpre-releaseremain on wirejs-native — sodf218917is indeed a dangling commit reachable by SHA only. Retargeting is the right fix.- wirejs-native#4 merged
2026-04-21T12:42:28Z; abieos#12 merged2026-04-20T19:50:58Z. The ~17h gap matches the description. - The three review-feedback commits are exactly what's picked up between the two SHAs:
1cb1d79bAddress PR #4 review feedback (kv_table* removal)21c97dfeDocument int64 enum value width in Abi interface34750244Adopt unified get_table_rows; drop get_kv_table_rows
- Diff is a one-line gitlink change — no source drift.
API-impact sanity check
wirejs-native#4's review commits remove get_kv_table_rows and the orphaned kv_table* types, which could be breaking for consumers. I ran a code search across this repo for both symbols — zero hits — so abieos itself isn't affected and doesn't need compensating changes. Worth a quick confirm from anyone who knows downstream consumers of abieos that might reach through to these types transitively.
CI
13/14 checks green; macos-intel still pending at review time. Submodule-only change across a matrix that's already green on clang/gcc/macos-arm makes a macos-intel-only failure very unlikely, but worth waiting on before merge.
Process note (non-blocking)
The root cause here is that abieos#12 was merged against wirejs-native's feature-branch tip before that branch's review was complete. This PR patches the artifact cleanly, but it might be worth a team convention: land submodule bumps only after the upstream PR has actually merged, so the pointer is never at a pre-review SHA that can later become dangling. Would have avoided both the dangling-commit window and the need for this follow-up.
Nits
None. Commit message and PR body are unusually thorough — the named commit SHAs, the "why retarget" framing, and the explicit "submodule pointer only, no source changes" line all make this trivial to verify.
LGTM pending macos-intel.
2 participants
Summary
Retarget
external/wirejs-nativefromdf218917(pre-review feature-branch tip) tob096df42(wirejs-native master head after Wire-Network/wirejs-native#4 merged on 2026-04-21).Background
abieos #12 merged on 2026-04-20 with the submodule pointed at the then-tip of wirejs-native's
feature/wire-sysio-table-idbranch. wirejs-native #4 merged the following day after three additional review-feedback commits:1cb1d79b— remove orphaned legacykv_table*types21c97dfe— documentint64enum value width34750244—get_table_rowsunification (dropget_kv_table_rows, addfind/index_name)After the feature branch on wirejs-native was deleted,
df218917became a dangling commit — still reachable by SHA, but fresh clones pulling it miss all three review rounds. Retargeting tob096df42fixes that.Submodule pointer only; no source changes.