web: follow the tile57 display_priority property rename#53
Merged
Conversation
tile57 renamed the MVT paint-order properties: draw_prio -> display_priority, cat -> display_category, and added display_plane (S-101 DisplayPlane, emitted only when OverRadar). The client reads these keys directly, so it has to move with them. Three live reads plus one inspector label: - chart-style.mjs fill-sort-key on the areas layer - chartplotter.mjs pick-report comparator - s52-style.mjs display-category membership filter - dev-tools.mjs feature-inspector value formatting The Go side needs no change: chartplotter uses the bake and metadata APIs, not the draw-callback surface, and tile57.Feature (S-57 metadata) has none of the renamed fields. `go test ./...` and `make build` are clean. Re-vendored style-engine.wasm — it is built from tile57's style engine and had the old key names compiled in, so a stale binary would emit draw_prio while the JS read display_priority. Verified end-to-end by baking testdata through `chartplotter bake` and decoding a z12 tile: display_priority and display_category present, zero draw_prio or cat. NOT included: the tile57 submodule pin. The engine commit carrying this rename is not on main yet, so the pin stays where it is rather than recording a reference nobody else can resolve. Two things to know before this merges: CI floats the submodule (`git submodule update --remote tile57`, ci.yml:32) and ignores the pin, so it builds against tile57 main. Until the rename lands there, CI on this branch is GREEN BUT WRONG — Go tests never touch the property keys, and the style expressions coalesce a missing display_priority to 0, so the chart renders at a single flat priority instead of failing. Baked archives must be rebaked; old tiles carry the old keys and fail the same silent way. `make clear-cache` then bake-ienc / bake-noaa — note both targets skip when the output file already exists, so a stale archive persists unless deleted. display_plane is carried in the tiles but not yet honoured in layer ordering. It outranks display_priority per S-52 §10.3.4.2, and chartplotter encodes paint order as MapLibre layer sequence, so respecting it is a behaviour change rather than part of this rename.
tile57's compose ABI now finds/reuses/rebuilds the ownership partition beside the archives itself (compose_open/compose_tree dropped partition_path; compose_save_partition is gone). Drop all the host-side sidecar plumbing: NewComposer no longer takes a partitionPath, Composer.SavePartition and livePartitionPath are deleted, bake.go stops writing partition.tpart, and openLiveComposer just opens — no load, no save. Fixes the StalePartition dance that forced a blocking owned-face rebuild at boot on an engine bump. Bumps the tile57 submodule to origin/main (2c36da6, PR #31 label-cache).
Bake pre-warm and the live provider opened a compositor over an explicit slice of per-cell archives, round-tripping tile57_chart_open across cgo once per archive (~35 ms each) — on a national library that turned boot registration into a minute of chart_open churn. Route both through NewComposerTree (tile57 OpenComposeTree): one engine call walks, mmaps and composes the whole tree on the batch path. The ownership partition is still the engine's — found beside the archives, reused when it matches, rebuilt when not. Registration measured 60 s -> 7 s. Share the Composer construction between the two openers via newComposerFrom.
The prior commit routes bake pre-warm and the live provider through tile57.OpenComposeTree, but the pinned engine (2c36da6) predates it, so the CI cross-build failed: undefined: tile57.OpenComposeTree. Float the submodule to main (b0e1b30, PR #32 merged), where the batch compose-tree open lands.
…gine's TestImportPacks stat'd <setDir>/partition.tpart, a host-written contract from before the ownership partition moved into the engine (OpenComposeTree). The engine now writes the sidecar only during a bake, beside the archives, and the compose-open path merely refreshes an existing one; the host neither writes nor reads it, so there is nothing at setDir to assert on.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
tile57 renamed the MVT paint-order properties: draw_prio -> display_priority,
cat -> display_category, and added display_plane (S-101 DisplayPlane, emitted
only when OverRadar). The client reads these keys directly, so it has to move
with them.
Three live reads plus one inspector label:
The Go side needs no change: chartplotter uses the bake and metadata APIs, not
the draw-callback surface, and tile57.Feature (S-57 metadata) has none of the
renamed fields.
go test ./...andmake buildare clean.Re-vendored style-engine.wasm — it is built from tile57's style engine and had
the old key names compiled in, so a stale binary would emit draw_prio while the
JS read display_priority. Verified end-to-end by baking testdata through
chartplotter bakeand decoding a z12 tile: display_priority anddisplay_category present, zero draw_prio or cat.
NOT included: the tile57 submodule pin. The engine commit carrying this rename
is not on main yet, so the pin stays where it is rather than recording a
reference nobody else can resolve.
Two things to know before this merges:
CI floats the submodule (
git submodule update --remote tile57, ci.yml:32) andignores the pin, so it builds against tile57 main. Until the rename lands there,
CI on this branch is GREEN BUT WRONG — Go tests never touch the property keys,
and the style expressions coalesce a missing display_priority to 0, so the chart
renders at a single flat priority instead of failing.
Baked archives must be rebaked; old tiles carry the old keys and fail the same
silent way.
make clear-cachethen bake-ienc / bake-noaa — note both targetsskip when the output file already exists, so a stale archive persists unless
deleted.
display_plane is carried in the tiles but not yet honoured in layer ordering.
It outranks display_priority per S-52 §10.3.4.2, and chartplotter encodes paint
order as MapLibre layer sequence, so respecting it is a behaviour change rather
than part of this rename.