3c087a43cdd94a4540e64f4688d8c5b18c3c1385
15 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
| 3c087a43cd |
fix(assets): a busy shard is not a broken one, and a column that reached no existing install (Phase 9a)
Two defects the Phase 9 acceptance walk found on a real rig, one of them ours and
one of them released (docs/link/v8.md §17.14).
## "The shard is not answering for client files" about a shard that was fine
The status call exhausts its 425 backoff whenever something else holds the
shard's single asset slot -- an import the operator started, or the item-art warm
pass refilling itself after a client patch. Phase 8's panel rendered that with
the same banner as a shard that is down or has the plane switched off, and left
it standing, because the page only re-reads after an action. On the rig it was up
for a quarter of an hour while the warm pass refilled 313 pictures and every
direct call to the same route answered normally.
BUSY now says what it is, and one automatic re-read four seconds later clears the
ordinary case. One per mount, guarded by a ref: a page that retried forever would
be holding the slot it is waiting for. DOWN, DISABLED and NO_IMAGING read exactly
as they did.
## Every spawn-atlas import on an upgraded install has failed since v1.2.0
`shard_spawn_points.unique_id` (Events Phase 12b) was added to the CREATE TABLE
and nowhere else. `CREATE TABLE IF NOT EXISTS` does not add a column to a table
that already exists -- which is what the twenty-odd `ADD COLUMN IF NOT EXISTS`
lines in this same file are for -- so it reached fresh installs and no existing
one, and `replaceAtlas` inserts the column unconditionally:
Unknown column 'unique_id' in 'INSERT INTO'
No bestiary refresh, no spawn map, no champion altars, on every install whose
tables predate 12b. A fresh install cannot reproduce it and neither can a test
whose schema is this file applied to an empty database; it took a rig with old
tables. Org lead, weighing that it is already released: it ships here on edge
rather than as a hotfix to main.
Verified by dropping the column, rebooting, watching the replay put it back, and
importing 6,455 spawners over the bridge.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016wDDVXWMDz82WqE1i969r4
|
|||
| 675e879b48 |
feat(assets): the panel that operates the client-file imports (Phase 8)
Admin -> Client Files: one page over the three things that come out of the
operator's UO client -- creature portraits, item and land pictures, and the
cliloc table. One page rather than three because they are one job: same client
install, same bridge, and all of them change at the same moment, when the
operator patches that client. Boot never asks the shard for any of it, so these
buttons are the only thing that imports.
The cliloc pair had had no UI at all since phase 2. On a bridged install, where
boot deliberately stopped calling the shard, that meant `curl` was the only way
to load 67,496 names.
Update and Re-import everything are section 6's two stages as two buttons rather
than one button and a checkbox, because they cost wildly different things. A
vanished key is reviewed in the page and not in a table -- an asset import only
happens because someone pressed a button here, so the review is already in front
of the person who caused it -- and it shows each key's PICTURE, since
`body/820/a23` names nothing a human recognises. `shard_asset_meta` gained a
`last` block (what the import did, who ran it) so the panel can answer "did last
week's import do anything" without scrolling core's whole activity log.
The live walk against a real shard imported 1,095 portraits in 3.5 s, warmed 313
item pictures in 0.6 s and reloaded 67,496 cliloc rows in 1.7 s -- and found two
DELETIONS that predate this phase and that no test could see, because only a
screen showing the numbers together makes them visible:
* The body import diffed its manifest against every family's rows. Phase 5 put
item and land art in the same table, and a body manifest never mentions
them, so all 313 item pictures were staged for deletion with a sentence
saying the shard had stopped offering them.
* An approved vanish unlinked the sprite and kept the row. The catalogue went
on counting a picture that was gone, the atlas could point a creature page at
a missing file, and the next forced import offered the same key for review
again -- reporting "nothing was changed" about a file it had deleted.
Both fixed here, with the removals now inside `saveAssets`'s own transaction.
The same whole-table read made the panel announce a 1,408-row creature catalogue
on an install holding 1,095 portraits and 313 item pictures.
Protocol stays 8 and EXTRACTOR_VERSION stays 3: nothing on the wire changed.
Refs: docs/link/v8.md sections 12.2, 14, 16 (phase 8)
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016wDDVXWMDz82WqE1i969r4
|
|||
| d4d5989926 |
feat(atlas): the spawn atlas reads the shard, not the shard's filesystem (Phase 7)
`spawnAtlasSource.js` gains a second backend behind its existing interface
(docs/link/v8.md 10). Where a shard is linked and enabled the tree arrives over
the sidecar; where there is none, a local ServUO tree is read exactly as before.
An explicit --servuo path is an instruction and overrules both.
The parsers do not move. spawnAtlasParse.js is still pure, still fs-free and
still CI-covered without a ServUO tree anywhere near it; `buildFromFiles` is now
where the parse starts, and both readers feed it the same shape.
treeBridge.js walks the manifest and then the chunks. Three of its checks are
not decoration -- each is a way this ends in a tree that LOOKS imported, and XML
is forgiving enough that a mis-assembled spawn file parses cleanly and simply
has fewer spawns in it:
- every chunk re-declares its address and carries the hash of its own
uncompressed bytes, and chunks are placed by declared index rather than
arrival order
- the whole file is hashed after reassembly against its manifest row
- the catalog must not move mid-walk, or the import is refused rather than
stitched out of two trees
Boot does not call the shard. The same answer 17.7 gave the cliloc table, and
the same reasoning: a local tree hashes in ~120 ms and skips, while a round trip
in the boot sequence would answer "no" on every restart that did not follow a
map edit. Editing spawn files is an operator action, so importing is one --
Admin -> Spawn Atlas -> Import. What that costs is real and is said out loud in
the panel, the CLI and the log: an install on the bridge has NO automatic
refresh at all.
Two things the live walk found that the unit tests could not:
- PARSER_VERSION 4 -> 5. The parse is order-sensitive in one place -- the
decoration index keeps the FIRST item id it sees for a type -- and the two
readers agreed on a stock tree by coincidence, since the filesystem reader
walks each directory with localeCompare while the shard sorts whole relative
paths. buildFromFiles now sorts by label, ordinally, once, whatever order
the files arrived in. Identical input, a different answer for a handful of
types: exactly what the version number exists to push through the hash gate.
The parity test asserted deepEqual, which ignores key order; it now asserts
serialised equality too.
- The source fingerprint is taken over RAW BYTES at both ends. Hashing decoded
text hashes a UTF-8 re-encoding -- identical for valid UTF-8, different for a
file that is not, because an undecodable byte becomes U+FFFD and never comes
back. One Latin-1 character in a creature name would have made the drift gate
report a change on every import, forever, with the tree untouched.
A 200 from assets.sources also stopped meaning "the client files are on offer":
a shard may now serve its configuration tree while declining to serve its UO
client. Both client-file readers check `assetsEnabled` and say DISABLED, instead
of reading an empty file list as "your client has no cliloc.enu" and sending an
operator to their client install for a setting that lives on their shard.
Measured end to end against a live shard and the real sidecar: 141 files,
11.9 MB, 158 chunks, 3 pages, 1.33 MB on the wire, 512 ms; every file
byte-identical to disk; and the atlas built over the bridge identical to the one
built off it -- 6,455 points, 800 creatures, 387 regions, 558 landmarks,
25 champions, 309 decoration types.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016wDDVXWMDz82WqE1i969r4
|
|||
| f335531538 |
feat(assets): item pictures on the marketplace and the character sheet (Phase 5)
Both places this site already knew an item's (ItemID, hue) and could only print it as text now show the picture, hued the way the client would draw it. The shard does the hueing: whether a hue repaints every pixel or only the grey ones is a flag in `tiledata.mul`, which a browser has no way to read. **Ingest warms; the route only serves** (org lead, 2026-09-11). A page never waits on the shard and never causes a fetch -- it renders what is stored and leaves out what is not, which is the state every install was in before this phase. Fetching happens behind that, on a timer, from the keys the site's own rows name. The alternative, fetching on first request, was rejected on one number: the shard's asset plane serves ONE request at a time, so a URL that fetched would let any visitor walk 49,152 ids times 3,000 hues through that slot and park an operator's own import behind it. The wanted set is DERIVED (`SELECT DISTINCT item_id, hue`) rather than queued, so it is self-healing: a restart loses nothing, and a key stops being wanted the moment the vendor row naming it is deleted. The in-memory hint set on top is only for the character sheet, which is fetched live from the shard and stored nowhere -- nothing on disk would ever name those keys. Staleness without a manifest (§7): every row records the shard's `catalog` id, a hash of the files that decide its bytes. A client patch changes it and a restart does not, so "is this out of date?" is a per-row question -- and pictures nobody looks at any more are simply never re-fetched, which is why this is lazy rather than a sweep. `shard_asset_meta` is deliberately NOT written here: it is the body catalogue's singleton, and a warm pass touching it would tell the body import that a client it never looked at is unchanged. A key the shard has no art for writes no row at all. An empty row would make the key held and it would never be asked again -- including after the operator patches in the graphic that was missing. `assets.sources` now reports which families an overlay serves, so an overlay older than phase 5 is one reported state with a sentence naming the fix, instead of a refusal per pass forever with no picture ever appearing. 688 server tests pass (14 new); client builds; the frozen manifest regenerates with one added route, all documented, no core URL moved. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016wDDVXWMDz82WqE1i969r4 |
|||
| a194ec68e0 |
feat(assets): creature artwork from the shard's own client (Phase 3)
Until now the only way a creature got a picture on this site was for an
operator to open UOFiddler on a desktop, export sprites by hand, copy them to
the web host and write a spawnAtlas.art.json naming each one. Almost nobody
did, so shard_spawn_creatures.art was NULL on every install.
The shard has had those files the whole time. Admin -> Shard -> Import now
walks its asset manifest, fetches only the sprites whose hash changed, writes
them under uploads/atlas/, asks the shard for a body id per atlas creature
(§8: it CONSTRUCTS the creature and reads Body.BodyID, which is the only thing
that is right for a shard's own custom creatures) and points each creature at
its picture. On a stock client that is 787 portraits, about a megabyte.
**The one thing v8.md §12 got wrong, and it is not cosmetic.** It says
`shard_spawn_creatures.art` "starts being filled by the import". That table is
emptied and refilled by replaceAtlas on EVERY atlas refresh, and a refresh runs
on every boot -- so a filename stored there would be destroyed by an ordinary
re-parse of the ServUO tree, with the next Update finding the client files
unchanged, reporting "nothing to do", and never restoring it. Nothing would
report a fault; the pictures would just be gone.
So the assets and the body map live in their own tables outside that blast
radius, and applyAtlas re-derives `art` on the way past as
`{ ...derived, ...operatorMap }` -- which is also the one place "the operator's
own artwork wins" is enforced, on every rebuild rather than only at import.
Smaller decisions worth not rediscovering:
- The derivation joins on the catalogue KEY, not on the body id. The simpler
join is correct today and stops being correct the moment phase 6 adds
body/400/a2/f0, at which point one slug matches dozens of rows.
- Filenames are content-addressed. A stable name overwritten in place leaves
every browser and CDN serving the previous client's sprite, with the database
row perfectly correct.
- An unchanged key whose FILE is missing is fetched again. The row and the disk
can disagree (a wiped uploads volume, a restore from a dump), and a broken
image on a creature page is worse than one re-fetched sprite.
- A key the shard cannot render is not a failure. Two thirds of the playable
ghost and gargoyle bodies have no art on a stock client, and an import that
reported eight failures every time would teach an operator to ignore the panel.
- A key that VANISHED from the manifest needs review before anything changes:
an unmounted client volume and a deliberate downgrade look identical here.
23 new tests; 674 server and 42 client tests pass. The SQL was also run against
a real MariaDB, which is what proved the CONCAT join and the singleton CHECK.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016wDDVXWMDz82WqE1i969r4
|
|||
| 1a13f680f5 |
feat(guilds): name the core contribution each declared slot wants
Core no longer fills a slot by name - it offers a contribution and the module that owns the page says where each one goes (MODULE_API 1.6.0, amended). The three slot names are unchanged and stay this module's own vocabulary; what is new is the second argument saying which of core's three contributions belongs in each place. Nothing here worked differently before. The change is for every game that is not this one: core used to fill the literal name uo.guild.detail, so a second module declaring a place under its own id got an empty page and no error. The registration fake gained the same validation core does, including the contribution catalogue - written down rather than imported, since this suite runs against the built chunk with no core in the process, which makes it a claim about core that has to be re-read when core's list changes. 42 client tests, 437 server tests. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| c57310c505 |
feat(guilds): a third place on the guild page, and where that page lives
Two lines only core cannot supply for itself.
`uo.guild.header` is a third declared slot, at the top of the page, for core's
per-Team notification control. A third rather than a corner of the feed because a
slot holds one component and the first fill wins: the control is an action ON this
page and the other two are content IN it, and separate slots are what let this
module say so.
`pageUrlTemplate` tells core where a guild page actually is. Teams are a contract
primitive with no core surface — core owns the tables and the access rules, this
module owns the word "guild" and therefore the page — which leaves core unable to
write a link to one. A notification email that cannot take you to the thread it is
about is most of the way to useless. Core substitutes `{externalId}` and does
nothing else with it; a template naming its own host is refused at registration.
Co-Authored-By: Claude <noreply@anthropic.com>
|
|||
| 9d0a197008 |
feat(guilds): declare a second place on the guild page, for core's forum
The mirror of the activity feed, one phase later. Core owns the Team forum — membership, manual grants and the member/guest split are all core's rules, and a module reimplementing any of them would be reimplementing a security boundary — but core publishes no Team page, because it does not own the word "guild". So this module declares the place and core puts the forum in it. TWO declarations rather than one, and that is the interesting part. A slot holds one component and the first fill wins, so folding the forum into `uo.guild.detail` alongside the feed would hand core the decision about where each of its two contributions sits on a page this module owns. Separate slots also keep them independent: with the forum switched off, the feed renders exactly as before. The registration test now asserts the set of declared slots and that EVERY one of them is rendered by the page that owns it, rather than naming a single slot twice. A slot nothing renders is a slot core fills into the void. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| dda0e32dd3 |
feat(guilds): a guild detail page, and the slot core puts the feed in
The module's half of the org lead's correction: Teams is the contract, guilds are the presentation, and the presentation is this module's. Adds `/uo/guilds/:id` — the detail view the board never had — with the roster from this module's OWN board, which is the same data it answers core's Team provider from. Reading core's projection of our own answer back would be a round trip through a staler copy of it. The page declares `uo.guild.detail` and core fills it with the Team activity feed. That is the one part of this page core cannot hand over: only core can resolve whether the viewer is inside the Team, and the public/members split on that feed is a security boundary. The guild is named in OUR terms — core maps its own Team from the module id and the external id — so this module never holds core's row id or slug. `TeamOverviewStrip` is deleted with the core Team page it filled. `team.member.row` is not declared here either: the useful thing to put in a roster row is a link to the character behind it, and nothing core could supply identifies one. `GET /public/shard/guilds/:id` backs the page, gated and projected through the same `guilds` feature as the board — so an operator who raises that audience raises this too, and the locked acct/webId fields never survive below admin. A roster is where those appear in bulk, which makes this the endpoint where getting the projection wrong would matter most. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| d4aa5ade12 |
feat(teams): project rosters by audience rung, and add to the Team page
The module's half of TEAMS.md phase 3. `projectRoster` is the optional fourth provider method and the only one core calls on a request path. Core holds the roster and owns its public shape; the question that is this module's is who is allowed to look, because the audience rungs and their configuration live here. The answer is all-or-nothing, which is the honest translation rather than a shortcut: a rung is a property of the FEATURE, and there is no configuration in which some members of a guild are public and others are not. The refusal semantics INVERT here, and the tests say so. For the other three methods a refusal means "change nothing" and an empty array would be destructive. Core fails CLOSED on this one, so the dangerous answer is the opposite — returning every key because the config could not be read would publish a roster an operator gated to staff. Every path that cannot reach a confident answer refuses, including the catch. The anonymous case is answered directly rather than by handing `viewerLevel` a synthetic request. Given one with no `req.user` it falls through to `auth.getUserFromRequest`, which expects real cookies and throws on a fake — and that throw would have become a refusal, so every anonymous visitor would have been served an empty roster on a shard whose guilds are public. Caught by the tests, not by reading. `team.overview` gets a live population reading beside core's stored one. Core's number comes from the last roster sync and is coarse by construction; this is the `presence.online` feed this module already holds. It is explicitly not a per-Team presence figure — the shard publishes a global aggregate and no per-guild breakdown exists on the wire, so claiming one would be inventing a number — and it renders nothing at all when it has nothing true to say. `team.member.row` is left unfilled. The useful thing to put there is a link to the character behind a row, and the props core can supply do not identify one: the member key and the site account id are withheld from every public roster. An empty cell beats a guess. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| d70e5e10d0 |
test(client): assert the URLs this module calls
Five assertions that stayed behind in core's `apiClient.test.js` when the bindings moved in slice 1 — the atlas-vs-shard path split, the query-string filtering, the slug encoding, the admin atlas methods — plus a new one pinning the seven admin URLs the shipped Android app calls by name. They were asserting UO URLs from inside core's suite, which is the boundary Phase 3 removes, and core's slice-4 deletion of those bindings would otherwise have deleted the coverage with them. The fake `window.__rg` carries the REAL react/react-dom/router rather than stubs: `src/core.js` compares its imported bindings against the published ones and logs a "bundled its own copy" error when they differ, so stubs make every run of this file print the exact wording of a real defect. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| 9d559091c5 |
fix(client): give the portal row an icon, and assert every gated nav has one
The §7.7 smoke, running the pair together: registering the player row without an `icon` blanked the whole portal with React error #130, because core's PlayerPortalLayout rendered `<n.icon />` unguarded. Core is guarded now (website), and this is the other half — the row had an icon before it moved and should have kept one. A second glyph rather than reusing IconShard: these two rows sit in different navs and each matched its neighbours before the extraction. The admin sidebar's UO rows were gems; the portal's Characters row was a person beside Appeals' shield and Account's gear. Matching the nav a row lands in is the whole reason `icon` is in the contract. `registration.test.js` now asserts it for admin AND player rows, which is the cheap place to catch the next one. The public header is text buttons and is deliberately excluded. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| e4af7dd9a8 |
test(client): check what the chunk registers, and fix two defects in the checks
Three things, all about checks that had never met a real chunk.
`checkExternals.js` rejected slice 3's build outright, naming a fragment of
minified JSX as an imported specifier: a button reading "Approve and import"
puts the token immediately before a quote, and no regexp can tell that from a
statement. Same wall the server's `checkImports.js` hit, answered the same way —
a character walk. A mask rather than a rewrite, because a real import has its
keyword outside a string and its specifier inside one.
Writing the test for that false positive found the false NEGATIVE underneath
it: the pattern required whitespace after `import`, so it could not see
`import{useState}from"react"` — the one shape a minified build actually emits,
and the most likely way for a missed alias to reach production. It has never
been able to see it.
`registration.test.js` is new: stand up a fake `window.__rg` with a recording
registry and the real React, import the BUILT chunk, and read back what it
asked for. No DOM, because nothing renders. It holds the agreement that rots
quietly — every nav row points at a route this module actually registered —
rather than restating both lists.
CI now builds before it tests, because both of those read `dist/entry.js` and
skip without it. Run the other way round they are green and asking nothing.
Co-Authored-By: Claude <noreply@anthropic.com>
|
|||
| 28f4b9afe2 |
feat(client): the whole client half (phase 3, slice 3)
The 35 files behind twelve public pages, seven admin views, two player views and three core-page extensions, ported onto `window.__rg`. Every one of them imports exactly the seven kit members plus `lib/format.js`, which is the finding §2.7.1 predicted and this confirms. `client/src/core.js` is the port mechanism, and unlike the server's it is a plain read: `window.__rg` is published before any module chunk evaluates, so there is no gap to defer around and a ported component keeps its ordinary import shape. `client/src/api.js` rebuilds the UO namespaces over the request primitive — same URLs, because §1.2 freezes the API surface. SPA paths changed and API paths did not. `/site/shard` is `/uo/shard`, and the admin paths lost their now-redundant `shard-` prefixes (`/admin/uo/ops`), a clean break being the only moment that is free. `shim/rg.js` becomes the single reader of the global, so the "core did not publish its dependencies" message is reachable from whichever module the bundler happens to touch first rather than from whichever one is imported first — a guarantee that used to last until someone sorted the imports. Co-Authored-By: Claude <noreply@anthropic.com> |
|||
| 5d7668d5ea |
feat(module): the bundle skeleton (phase 3, slice 0)
The first real module. It registers nothing, deliberately: what slice 0 proves
is the delivery path itself, end to end, before a single UO file moves into it.
Server half: module.json, an entry point that takes (ctx, api) and registers
nothing, a test suite built on a fake ctx, and scripts/checkImports.js -- the
MODULE_API.md §5.1 boundary check. Client half: the Vite library build, four
shims re-exporting react / react-dom/client / react-router-dom / jsx-runtime
from window.__rg, an entry that verifies each is identity-equal to core's copy,
and scripts/checkExternals.js. 29 server tests, 9 client tests, both new.
Verified against a real core: the module loads, mounts its zero routes, runs to
`started`, and is published by /api/v1/public/modules. Its chunk serves from
the entry's directory with `Cache-Control: no-cache` while the module's server
source, module.json and package.json all 404. In Chrome, under the enforced
`script-src 'self'`, the chunk evaluates and reports all four shared
dependencies OK, with zero CSP reports and no console errors.
Three findings, each of which had produced a green build that was wrong.
MODULE_API.md §3.6 shows `external` alongside the aliases and they do not
compose. Rollup asks `external` BEFORE Vite's alias resolver runs, so a
specifier in both is marked external and never aliased -- the chunk then ships
bare `import "react"`, which no browser can resolve without an import map, and
CSP forbids one. Built cleanly and emitted exactly that; checkExternals caught
it. So: alias only, `external` empty, and vite.config.js grows a resolution-time
guard that fails the build if a shared dependency resolves into node_modules.
That guard was wrong twice before it worked. Written against Rollup's `load`
hook it never ran -- `load` is first-wins and an earlier plugin had already
claimed the module -- so a deliberately-broken alias produced a 24 kB chunk with
react-router welded in, and a green build. And its forbidden-package list was
derived from the alias list "so the two cannot disagree", which meant deleting
an alias also deleted the guard against what that alias prevented. It states the
contract now, and a test asserts the aliases stay inside it.
checkImports failed on its own documentation the first time it ran: the comment
naming require("../../etc/passwd") as an example of what to catch, and index.js
explaining why the module must never require("express"). A boundary check that
cannot survive being described is one people stop writing comments around. It
strips comments and template literals with a character walk rather than a
regexp, because a URL in a string contains a comment opener and a comment
contains quotes -- and it has its own test suite, since a check never shown to
fail is a check nobody knows the state of.
Co-Authored-By: Claude <noreply@anthropic.com>
|
