MapColonies / shigola

47%

master: 44%

LAST BUILD ON BRANCH refactor/explicit-tile-grid

branch: refactor/explicit-tile-grid

CHANGE BRANCH

x

Reset

Sync Branches

• refactor/explicit-tile-grid
• build/cgo-free-end-to-end
• chore/ci-security-updates
• chore/complete-rebrand
• chore/dockerignore
• chore/drop-vendor
• ci/gate-vet-and-gofmt
• ci/publish-release-images
• ci/release-please
• development
• docs/contributing-for-this-repo
• docs/contributing-trunk
• feat/grafana-dashboard
• feat/log-trace-correlation
• feat/mapcolonies-log-format
• feat/otel-tracing
• feat/otlp-log-export
• feat/postgis-query-span
• feat/remove-capabilities
• feat/remove-viewer
• feat/serve-layer-collections
• feat/trace-exemplars
• fix/ci-docker-build-job
• fix/config-type-hints
• fix/grpc-go-2026-6348
• fix/land-ci-dependency-updates
• fix/log-call-site-spacing
• fix/release-image-stamp-and-registry
• fix/tile-crs-envelope
• master
• refactor/delete-geometry-math-trees
• refactor/protobuf-apiv2
• refactor/remove-unsupported-providers
• refactor/retire-standard-provider
• refs/tags/backup/pre-rebrand-world-crs84quad-tiles-pr-fixes
• test/cite-mvt-postgis
• test/coverage-baseline
• test/postgis-athens-fixture
• test/tile-content-scheme-edges
• update/update-helms

Build # 4900f5741

Build Type

Pull #10

github

Committed by NivGreenstein

Commit Message

feat(cache): warn when a seed or purge run picks its own scheme

A run without --tile-matrix-set still writes a whole pyramid under one,
and which one it is decides every key it touches. Until now it said so
only at INFO, in a line that reads the same whether the operator chose
the scheme or the command did.

The wrong scheme is not visibly wrong. A seed fills a cache nothing will
read; a purge leaves the tiles the operator meant to remove in place and
removes another scheme's instead. Both exit 0.

So warn, and name the scheme and where the choice came from — "a default
was applied" would leave the operator to work out which. Both implicit
paths warn, not just the WebMercatorQuad one: taking the first entry of
a map's tile_matrix_sets is the more surprising of the two, because
nothing sorts that list and swapping two entries changes what the same
command seeds.

Emitted after the support check, so a run about to fail says so once
rather than announcing a scheme it will not use.

Refs MAPCO-11486

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Pull Request Pull Request #10: refactor(atlas): make the tiling scheme explicit on the seed and encode seams (MAPCO-11486)

Coverage Stats

54 of 69 new or added lines in 5 files covered. (78.26%)

2 existing lines in 1 file now uncovered.

9342 of 19703 relevant lines covered (47.41%)

210.09 hits per line