[MIG] hr_*: analyses for simple submodules (v2 with reviewer fixes)

[dnplkndll]

What this is

Mirror of OCA#5634 with the 6 reviewer asks (hbrunn, remi-filament) addressed. Opened on the ledoent fork (not OCA upstream) for review before any decision about whether/when to push the fixes back to OCA#5634.

The corresponding commit on OCA#5634 (`f9e80d8`) predates both reviewers' feedback and hasn't been updated since 2026-05-20.

What changed vs the OCA#5634 commit

#	Reviewer ask	File(s)	Fix
1	"'Nothing to do' goes in col 2, not col 3 — also all of below" (hbrunn)	`docsource/modules180-190.rst`	Switched 13 hr_* rows from `Done | Nothing to do` to `Nothing to do | (preserve col-3)`
2	"don't delete valuable info from col 3" (hbrunn)	same	Restored `No DB layout changes.` on `hr_recruitment_survey` and `hr_work_entry_holidays`
3	"you don't touch the noupdate_changes.xml file … in this case, it just needs to be loaded in the post migration script" (hbrunn, on hr_presence)	`hr_presence/19.0.1.0/`	Reverted hand-edit of `noupdate_changes.xml`; added `post-migration.py` with `openupgrade.load_data(env, "hr_presence", "19.0.1.0/noupdate_changes.xml")`
4	"need to be deleted using delete_records_safely_by_xml_id" (hbrunn, on hr_fleet)	`hr_fleet/19.0.1.0/`	Added `pre-migration.py` with `delete_records_safely_by_xml_id` for the 5 orphan xmlids (4 ir.rule + 1 ir.ui.view); work-file annotation updated
5	"see above" (hbrunn, on hr_recruitment_survey)	`hr_recruitment_survey/19.0.1.0/`	Same as #3 — reverted hand-edit, added `post-migration.py` with `load_data`
6	"needs to load noupdate_changes.xml" (hbrunn, on hr_timesheet)	`hr_timesheet/19.0.1.0/`	Added `post-migration.py` with `load_data`; work-file annotation updated

Also reverted whitespace-only edits to `event/19.0.1.9/noupdate_changes.xml` and `hr_recruitment/19.0.1.1/noupdate_changes.xml` (per "don't touch noupdate_changes.xml").

Pre-submit verifier output

Ran `.claude/skills/openupgrade-contribute/scripts/verify-pr-staged.py` against the staged set:

• 0 MISS
• 0 WARN (after a verifier check-7 precision fix — section-level DONE markers no longer false-match against the line above)
• 9 MAYBE-MISS — sibling hr_* modules with noupdate_changes.xml not in this PR's scope. All are out-of-scope per the original [19.0][MIG] hr_*: analyses for simple submodules OCA/OpenUpgrade#5634 split (handled by other PRs or already `Done`). Soft exit by design.

Lab validation

Two of the touched modules are installed in our lab's prod seed:

• `hr_timesheet` — its new `post-migration.py` would fire during `make migrate`
• `hr_skills` — not changed by this PR (docsource entry only)

`hr_presence`, `hr_recruitment_survey`, and `hr_fleet` are not installed in our seed, so the lab can't validate their scripts directly. End-to-end validation deferred until we either (a) install those modules in the seed or (b) decide to push this back to OCA#5634 and let OCA's `Test OpenUpgrade migration` workflow exercise the scripts on its 18.0.psql seed.

Status

Draft, fork-only. Per the 2026-05-15 fork-first pivot, no decision yet on whether/when to push this back to OCA#5634. Three options:

1. Force-push [19.0][MIG] hr_*: analyses for simple submodules OCA/OpenUpgrade#5634 head with these fixes — single ceremony, addresses both reviewers' open feedback.
2. Open a brand-new OCA PR pointing at this branch — leaves [19.0][MIG] hr_*: analyses for simple submodules OCA/OpenUpgrade#5634 dormant.
3. Leave as-is on the fork — defer upstream contribution.

[dnplkndll]

Pushed to OCA#5634 (head branch is the same: ledoent:19.0-mig-hr-family-trivial). Closing this fork mirror to avoid double-tracking; the branch (19.0-mig-hr-family-trivial-v2) stays in case we need to iterate locally before another force-push.

Audit-trail for this round of review is committed at test-results/openupgrade-review/fork-pr38/81b8f846/ in the lab.