The saved ski route compared against an undefined skiKey, so opening a saved
ski tour threw. The held seed now lives in useSavedSkiSeed and is matched by
skiRouteKey(skiChoice). The ski route facts move to skiRouteFacts.ts so the
summary and leg rows only export components, the rail renders
SkiPlannerSummary, and openTour, useTourRoute and useAddonManager shed the
branches and lines the new ski code had added.
A ski leg is routed where the piste router planned it and free elsewhere;
the description said both. The collection tours entity takes its generated
form again after the merge.
A saved ski tour now opens with the line, heights, time, build and
planned legs it was saved with, read from the place this device keeps,
and nothing is asked of the piste router until a point, an option or the
discipline changes. Opened offline, while TREK Places is down or with
piste routing switched off, it no longer turns every leg straight, and a
rename saves the planned tour as it was. A kept line whose length is not
the tour's is another save's, and the tour is planned afresh.
- the save names the legs the router planned (ski_planned_legs, checked
against the tour's own legs); get_tour_plan reports each leg routed or
drawn from them instead of from the build alone, a tour that names none
counting as planned only with the router's time
- copies keep ski_options in every path: in the trip, taken over, kept
in a collection and made from one (collection_tours gets the column in
the same unshipped migration)
- the route summary shows the data build, the km of runs, the height
skied down on them and the lift rides, and the credit the server sends
in place of a fixed one; a tour opened as saved says so
- answers refused while piste routing is off, or planned on an older kept
copy, are not remembered, so they are asked again later
- terrain heights between known ones are filled in one walk along the line
- i18n: five summary keys in all 27 locales
The router thread now takes one job at a time from a queue kept on the
main thread, so a job's clock starts when the thread takes it up. A job
that times out ends only its own thread, and only while that thread is
still the current one; the jobs in line go to a fresh thread instead of
waiting out their own timers and killing the new thread in turn. A
thread that dies fails the job it held, and a thread that cannot start
fails each job on its own.
- the daily refresh fetches changed areas two at a time and stops once
the TREK Places breaker opens; the rest keep their older copies
- the winter index cap is on the unpacked JSON fetch hands over, so it
follows from the host's 8 MB gzip cap (64 MB) instead of matching it
- a build name is read to the rule a ski tour stores it by, so a build
TREK could not save is refused when the winter layer is read
- an unused import in the piste router contract is gone
A ski tour with a discipline now sends its legs to the piste router and
takes back runs, lifts and trails: each stretch is coloured by its grade
in the area's scheme, links and lift rides down are dotted, and a leg the
router cannot plan stays straight and unrouted with its reason and the one
change that would plan it. Snowshoe and winter hike legs off the piste
data follow summer paths instead, said so. The router's time and heights
are used once every leg is planned, and the options, the build and the
time are saved with the tour.
- skiRouteRepo and api: one request per tour, batched, remembered unless
the answer may change, offline legs failed as offline
- planner: router options per discipline, leg steps and failure rows with
fixes, a piste route summary, draft and reopen keep the options
- admin: switch for the piste router under Tours
- i18n: new ski keys in all 27 locales, ski info and time texts rewritten
A ski tour now saves the options its legs were planned with and the piste
data build they ran on, as one JSON column (a new migration, since the
ski migrations have shipped on this branch); the server keeps only the
options its discipline reads. Both stay off every other type. The refusal
of a ski time is lifted: the piste router works one out, and the planner
sends it only when every leg followed the piste data. get_tour_plan names
the legs of a tour the router planned as routed, and list_tours says what
the new fields hold.
A new Nest domain plans the legs of a ski and winter tour on the runs,
lifts, cross-country trails, ski touring and winter paths the Places host
builds from OpenStreetMap, one SGR1 graph per ski area. TREK fetches the
index once, downloads an area by its id when a planner first plans in it,
checks it against the index's SHA-256 and keeps it in data/ski; no
clicked point ever leaves the server. A leg whose area is slower than two
seconds answers loading, and an older copy stands in while the host is
down.
The reader decodes the writer's golden file byte for byte, including the
two edge flags it added (area_has_run, type_unknown). The search is A* in
a worker thread over a profile per discipline (alpine, snowboard,
cross-country by technique, touring, snowshoe, winter hike) with the
times and penalties of the cost model, up to five snap candidates per
point, lift stations as entry points, legs that go on from each other,
pieces of a split network stitched by their boundary nodes, other ways by
the penalty method, and a failed leg explained by relaxing one rule at a
time (difficulty, lift class, technique, riding down, uphill) without ever
handing out the relaxed line.
POST /api/trips/:tripId/ski/route, GET .../ski/area-near and
GET .../ski/network/:areaId sit behind Tours and the trip, with
plan_ski_route as their MCP twin. A daily job looks for a new build and
sweeps areas unused for 90 days. The admin switch ski_routing_enabled is
on wherever TREK Places is.
The binary heap and the flat-frame distance helpers are not about water;
the piste router needs the same two. They move to src/lib/geo, where any
domain may import them without reaching into paddle-routing. Nothing
changes in what they do.
The ways and routes layers now count failures on a breaker of their own, so
busy or slow layer answers no longer send place search to the fallback for a
minute, and a layer's 404 no longer clears the count of search failures.
When a later batch of /v1/ways fails, the batches that answered are kept and
only the ids still open go to the Overpass mirrors.
The wiki, the compose file and the chart now say that the way ids of a planned
tour and the horse routes box go to the TREK Places API.
The trail grades under a tour (BIK-05, MAP-03, MAP-04, RUN-02) and the
horse routes around a map view (HOR-01) asked the public Overpass mirrors.
The places host now serves both from its own planet build: GET /v1/ways
for the tags of ways by id and GET /v1/routes for route relations per box,
each in Overpass element form.
trek-places.client gains trekPlacesWays (batches of 500, one after
another) and trekPlacesRoutes, both through getJson with its off switch,
breaker, timeout, size cap and instance header, each element validated on
its own. TrekPlacesLayersService in the maps domain hands their answers to
the tours services, or why there is none (switched off, layer not built,
failed), and logs that once per layer until it answers again. Both tours
services ask the layer first and fall back to the existing Overpass path;
the responses TREK sends its clients are unchanged.
Since the estimate rule, an import without its own time shows the worked-out estimate on the day plan tile, and the sidebar metrics lead with it. Add the estimate field to the RS-01 fixture.
Move workedOutEndHint and totalText into tourDayTimeHint.ts so TourDayTimeParts only exports components, render the collection row's type icon without a component made during render, and pass the chase camera's lens and scene as one view object.
The flyover asked wheel and pinch to zoom around the centre with enable()
on handlers that were already on. MapLibre returns from enable() at once
in that case, so the option never took and a wheel zoom still went to the
cursor before the camera pulled back to the line. Each handler is now
switched off and on again with the centre anchor, and the same way back to
the cursor on pause, stop and unmount. The map stub refuses an enable() on
a handler that is on, like MapLibre does, so the tests see the real anchor.
The GPX file kept for the share sheet was only dropped when the tour's id
or name changed. A route saved elsewhere arrives live with a new revision
and the dialog kept handing out the old file. The kept file now carries the
revision it was read at; a new revision drops it and fetches again, and a
file a tap kept is not shared for another revision.
A fresh install gets Tours enabled from the addon seeder. An install that
already has the row is switched on by a migration that runs once, so an
admin who turns Tours off afterwards keeps it off.
The chase camera set a fixed zoom and pitch on every frame, so wheel,
pinch and the +/- keys were undone at once. It now keeps the zoom and
tilt of the user's latest gesture and only follows position and
bearing. While a zoom is still easing the flyover holds for those
frames, since any camera move would cancel the gesture. Wheel and pinch
zoom around the rider while flying, and drag panning is off until the
flyover stops. Starting over returns to the flyover's own zoom.
Sharing a tour fetched the GPX file first and only then called
navigator.share, so the tap that asked had often run out by then. The
retry path kept that file for one more tap only, and a tap while a sheet
was still open made the browser throw InvalidStateError, which read as a
failed share.
The share button now fetches the file ahead where sharing is offered
(not while the planner has unsaved changes), and every tap hands a fresh
File built from the kept bytes to the sheet before anything is awaited.
A file a tap had to fetch is kept for the next tap. A tap while a sheet
is open is let be, and InvalidStateError and a closed sheet stay quiet;
only a real failure still says the download is left.
The get_tour_plan description named the estimate of a tour "such a field"
right after the values of a recording, so it could be read as one of them.
It now follows the sentence on derived fields and says outright that it is
one of them, worked out from distance and heights or the pace of a run.
The arrow between a tour's start and end points from the start to the end
in a right-to-left language as well. A tour just added to a day, which
carries no tour_place_id until the next refresh, now says its overrun like
the row above it. A phone row of an import with a total the travellers
entered shows that one time, the estimate joining its words as on the
desktop. The German overrun line reads naturally without an item name.
A tour on a day had no place in the day's timeline: its time was not tied
to the slot it was planned in, so several hikes on one day could not be
lined up. A tour now sits there the way a place does, on the times of its
day entry, one model for every tour type. With a start and a known time
(the planned total, else the estimate of an import) its end follows as the
start plus that time; it is shown fainter, marked with an approximate sign
where it comes from an estimate, and says in its tooltip how it was worked
out. An end typed by hand wins. A day that holds one stage of a multi-day
tour gets no end from the whole route.
When the end runs past the start of the day's next timed item, the row
says so in one quiet line, as a late connection between two city walks
does; nothing is moved. The existing time fields of the day entry, on the
desktop form and the phone sheet, say what the end works out to while none
is typed. City walks keep their arrivals per station. The logic lives in
one module both day plans run on.
A hike imported from GPX, KML/KMZ or wanderer had no time at all, so
several hikes on one day could not be planned against each other. It now
gets the walking time the planner times a hike by (DIN 33466), worked out
from its own distance, ascent and descent; a run gets its distance at the
pace entered. Without heights in the track the time is unknown rather
than the time on the flat, and types timed by a router, by a speed of
their own or by nothing get no estimate.
The estimate is derived on every read and never stored, so tours imported
before get it too and no stored or entered time is touched. The rule moved
to @trek/shared, where the planner, the client and the server read it.
The tour list, the day rows on desktop and phone and the tour detail say
it as an estimate, apart from a planned total; list_tours and
get_tour_plan carry it as time_estimate and name it as derived.
The times fill their row when there are two, the trip days are boxes in
the look of the measures with the date under the day, the footer's
collection button carries its label like the others and the copy button
says Copy, and the name field keeps its distance to a save button as large
as the close button.
A tour kept in a collection was a card in a box of its own above the list,
and its button lost its look to the collections page's button reset. It is
now a row of the list like a saved place: its grade or type tile, its name,
type and the trip it came from, its measures, and Add to a trip, with
removing dimmed beside it. The list's search narrows the tours too. The
phone shows the same rows and takes a tour into a trip from a sheet, on the
same hooks as the desktop.
The walk from a Google Maps link, the take-over from another trip and the
other new tour dialogs sat at the top of the screen; they now open centred
like the planner's own dialogs.
Snowshoe is no longer exported, useRouteCalculation takes the stage stops
type through dayRoutePlan, and tourPlannerFacts and useTourRoute hand the
saved duration, the ski check and the planned key to small helpers.
The collection tour fixture no longer writes the dropped winter_activity
column, the drafts are saved as version 11, and a hike no longer offers
the pistes layer, which belongs to a ski track only.
The winter label and four-discipline migrations are already on
feat/tours-board-rest, so a database that ran them would refuse to boot
or fail the old CHECK. 045500, 045520 and 046100 are back as they
shipped, and a new migration drops both winter_activity columns and
adds ski_discipline again with the six disciplines. The entity and
table types follow the column to its new place.
German, Hungarian, Indonesian and English no longer name the whole type
with the ski touring word, and the info dot speaks of ski tours and
tracks. search_tour_pois and its wiki row list rides and ski and winter
tours. Snowshoeing gets a snowshoe glyph instead of the walk's
footprints, a ski import without a discipline can be planned over (the
planner asks for it), and an overlay shown by default tries its tiles
afresh each time it comes on.
Ski becomes "Ski & winter", a type the planner plans with six disciplines
(alpine, snowboard, cross-country, ski touring, snowshoeing, winter hiking),
nothing preselected; saving needs a discipline, and a technique for
cross-country. Every leg is drawn straight between the points through
planner/skiRouting.ts, the one place the piste router will plug in, with its
heights from the elevation source and no ski time worked out.
The hike's winter label goes entirely: its migration, column, field, marks,
GPX rule, MCP field and strings. A hike is summer only; a ski tour with the
snowshoe discipline exports as snowshoeing. The unreleased ski migration takes
the six disciplines and the collection migration loses the winter column.
The pistes are offered for a ski tour, shown until switched off, and with no
tour open, never for a hike. Export, copy, collections, stages (untimed),
actual tracks, the atlas, list and day marks (one glyph per discipline), 3D,
planning over a ski import and list_tours/get_tour_plan know the planned ski
tour.
The server's src and tests together are past 570,000 lines and the test
typecheck needs about 4.3 GB, over Node's default heap; it now gets 8 GB like
the test jobs. Forty new client files are formatted, and the react-hooks/refs
baseline drops by the one warning a split hook removed.
Which places are tours, their planned times and where their stages run
travel together to the desktop and the phone day plan. useTripPlannerPage
now puts them together as dayPlanTours, and the page passes that bundle
instead of naming the three props at both places.
The day route plan and the day plan sidebar both built the map of a day's
stops projected onto their tours, each stage on its own stretch, by hand.
projectedById in dayTourProjection builds it for both, and the sidebar's
drive runs reuse the route plan's point type instead of spelling it out
twice.
reportGlViewport says what a GL map shows, its bounds and its zoom on the
raster tile scale, and has its own test for that scale. MapViewGL only
wires it to moveend and idle, and its test checks that wiring. The GL map
tests put the zoom back to its default before each test, as they do for
every other value a test overrides, instead of each test restoring it.
The hike's limit and preferences, the bike profile and the run profile were
four states with four refs kept in step by hand in every handler that set
them. useRoutingChoices holds them as one value with one ref, so a history
step, an undo, a new draft and an opened tour each lay them in with one
call, and useTourPlanner is back under the size limit.
The confirmed field is checked through z.object(...).pick() as the dawarich
and doc-sync tests read their tool shapes, and the stage edits pass typed
waypoints through the request contract as the walk tests do. The previous
forms sent the test typecheck out of memory.
Raw request bodies are objects, the collection snapshot names the run, ski,
winter and track time fields, the walk write says it has no hiking limit,
the MCP field shapes are read as zod shapes, and the stage edits go through
the tour request contract the API parses them with.
The map focus helpers split out of useTripPlannerPage returned object
literals ahead of the hook's own return, which the B4A wiring test reads as
the view model. They move to tourMapFocus.ts beside the hook.
Clears the client lint ratchet: helpers move to sibling .ts files for react-refresh, complex render paths get small helpers, floating promises are voided, the stage rail lets a stale request go during render, and the collections page reads the active list's tours and rights in ActiveCollectionTours. StageNumber is no longer exported.
The type order puts run before kayak, the import button reads Import GPX or KML, and the mocks gain what the components now read: routeRun, packing items and links, the map's zoom, the translation provider and the compare store's map flag.
Expectations follow createAssignment's options, the stage_ends column, the editor's trip access and the broadcasts of the take-over. Refusals in the run and ski e2e tests go one by one so supertest does not reset a connection under another request. A parity test keeps the source hint enums in step with the contract.
The activity section used Footprints without importing it. Opening one stop's menu now closes any other. lostStageEnds no longer counts a stop the change kept without its mark as lost twice over. The wanderer watch switch keeps the same tree once the connection is known, so it is not remounted and keeps its focus.
Maurice