Rebuild OHM customizations as logical commits on top of upstream

[Rub21]

This PR replaces the old patch-accumulated history with 12 logical commits that contain all OHM customizations, rebuilt on top of the latest upstream master openstreetmap-website-5a02636.

This is the approach agreed in the ticket OpenHistoricalMap/issues#735 as a cleaner baseline for future development and upstream merges.

cc. @1ec5

[github-actions]

	1 Warning
⚠️	Number of updated lines of code is too large to be in one PR. Perhaps it should be separated into two or more?

Generated by 🚫 Danger

[coveralls]

Coverage Report for CI Build 35295283987

Coverage decreased (-0.4%) to 96.661%

Details

• Coverage decreased (-0.4%) from the base build.
• Patch coverage: 21 uncovered changes across 3 files (24 of 45 lines covered, 53.33%).
• 11 coverage regressions across 3 files.

Uncovered Changes

File	Changed	Covered	%
lib/date_range.rb	21	5	23.81%
app/helpers/browse_helper.rb	5	2	40.0%
app/controllers/searches/nominatim_queries_controller.rb	12	10	83.33%
Total (9 files)	45	24	53.33%

Coverage Regressions

11 previously-covered lines in 3 files lost coverage.

File	Lines Losing Coverage	Coverage
app/helpers/copyright_helper.rb	6	25.0%
app/controllers/accounts/pd_declarations_controller.rb	4	73.33%
app/models/community.rb	1	80.0%

---

Coverage Stats

Relevant Lines:	7636
Covered Lines:	7381
Line Coverage:	96.66%
Coverage Strength:	895.93 hits per line

---

💛 - Coveralls

[1ec5]

This is looking pretty good. I only inspected the code but haven’t tried running it. I haven’t gone through our issues to see if any features dropped out. It’ll be easier to detect them in testing anyways.

I included some nitpick comments, in case it’s easy enough to remove those from the diffs.¹ The cleaner the diff, the less likely we are to have merge conflicts in the future. Some of the whitespace changes are probably remnants from code that we added and removed at some point.

Footnotes

1. You can do an interactive rebase to rewrite this history. Change pick to edit, edit the file, then continue the rebase. ↩

[Rub21]

Tests are passing now, the map load timing was breaking some tests, so I bumped Capybara's default_max_wait_time so it waits for the page to finish loading.

[1ec5]

Sounds good, it was failing intermittently on other branches for the same reason.

[Rub21]

I rebased this branch on the latest upstream master and pulled in the recent changes from staging (map-styles 0.9.19 and the locale key rename), same 12 logical commits, just on a newer base.

[1ec5]

We’re almost there! I looked over everything except test/. I specifically looked for any sign of a bad merge, other than some problems we already have tail work ticketed out about.

[Rub21]

I went through config/locales/overrides/ again and I found dead entries worth removing before we merge. We ship 11,035 override entries across 110 files, and about 9% of them are dead, in three groups that are dead for different reasons.

Key	Files	Entries	Lines
site.communities.lede_text	47	47	308
activerecord.attributes.client_application.*	96	929	1,027
date.formats.contribution_calendar	19	19	19
Total		995	1,354

site.communities.lede_text — we removed the Communities page here, with its route and header link, so nothing reads this key in OHM. It is the only subkey under communities: in all 47 files.

activerecord.attributes.client_application.* — upstream renamed these to oauth2_application when it moved to Doorkeeper. There is no client_application in upstream's en.yml, no model here, and no code using it. We still translate it in 96 languages.

date.formats.contribution_calendar — in 19 files our value is the same as upstream, so the override does nothing.

Only config/locales/overrides/ changes. The upstream locale files stay as they are, so the diff against staging-new gets smaller.