MyPageTask.Parse (Wizard/MyPageTask.cs:155-163) does
`_userItemDict.Clear();` the moment `user_item_list` is present in
the response body — not when it's non-empty — then re-populates
from the wire. Our /mypage/index was emitting [] by default (the
field initializer on the DTO), which wiped the inventory that
/load/index had just populated.
Downstream consequence: the client's PackChildGachaInfo.CostGoodsCount
reads from _userItemDict, so a wiped dict makes every ticket-cost
pack report CostGoodsCount=0, PackConfig.EnableBuyPack returns false,
and is_hide=1 packs (including the tutorial legendary starter 99047)
disappear from the rotation pack list — even though the gift bundle
just granted the ticket and the DB row exists. The tutorial then
auto-selects whatever non-tutorial pack happens to be at index 0 of
the filtered list, the user can't afford it, and the flow is stuck.
Fix:
- MyPageController.Index now sets UserItemList from viewer.Items
(already loaded by GetViewerByShortUdid's home-screen graph).
- DTO docstring rewritten to call out the presence-sensitive semantics
and the load-bearing path through PackConfig.EnableBuyPack, so the
next developer doesn't get the "empty is fine" hint the old comment
implied.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
LastPlayTime and MissionChangeTime were typed as DateTime, which STJ
serialised as "0001-01-01T00:00:00.0000000Z" for a fresh viewer
(DateTime.MinValue). Prod's wire shape is "yyyy-MM-dd HH:mm:ss"
(no T, no Z, no fractional seconds) when present and null when
absent — verified against data_dumps/traffic_prod_tutorial.ndjson.
The .NET default format has a real chance of crashing the client's
DateTime.Parse path on any code that reads either field, and the
fields are presence-sensitive (NetworkTask-family Keys.Contains
followed by ToDateTime), so emitting the .NET default reaches the
client as a stale-but-present value.
Switching the properties to string? + FormatProdDateTime helper:
- non-default DateTime -> "yyyy-MM-dd HH:mm:ss"
- DateTime.MinValue -> null (omitted from wire via global
WhenWritingNull policy in Program.cs)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
A wiped/fresh client (NukeIdentityOnStartup, new install, or any path
that clears PlayerPrefs) defaults its stored RES_VER to "00000000"
per Cute/SavedataManager.GetResourceVersion. The client builds the
Akamai manifest URL as dl/Manifest/<RES_VER>/<lang>/<Platform>/, and
Akamai 404s the "00000000" path -> Toolbox.AssetManager.InitializeManifest
fails -> the title screen shows "Connection Error / Reconnect"
before any tutorial UI loads.
Fix:
- New ResourceConfig [ConfigSection] in SVSim.Database — single
field RequiredResVer defaulting to "4670rPsPMVlRTd2" (the value
prod returned in data_dumps/traffic_prod_tutorial.ndjson and was
still returning at 2026-05-28 21:00 UTC). Lives in GameConfigs so
it can be tuned via DB / appsettings without code edits.
- ShadowverseTranslationMiddleware injects IGameConfigService and
emits required_res_ver in data_headers ONLY on /check/game_start
responses. NetworkTask.Parse opens a "new data is available" popup
whenever required_res_ver is present and the URL is anything other
than GameStartCheck (NetworkTask.cs:128-138); the suppression on
game_start is what lets us silently bump PlayerPrefs["RES_VER"]
before ResourceDownloader runs.
- DataHeaders gains a nullable RequiredResVer field. DataWrapper.DataHeaders
is now Dictionary<string, object?> instead of the typed DataHeaders POCO
directly — the construction site stays type-safe (the middleware builds
the typed POCO, then projects through the same STJ +
ConvertJsonTreeToPlainObject pipeline that DataWrapper.Data uses) so
null-valued optional fields are absent from the wire instead of being
written as "key":null. Without this, MessagePack's ContractlessStandardResolver
walked the typed properties and wrote required_res_ver=null on every
non-game_start response, tripping the popup on every boot.
- GameConfigurationJsonbTests updated to expect the 9th config section.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
After /tool/signup the client switches to SID-only headers (no UDID), so
the next request's body can't be decrypted unless the server already
knows the SID's UDID. ShadowverseSessionService now mirrors the client's
Cute/Cryptographer.MakeMd5(viewerId + udid) formula (salt
"r!I@ws8e5i="), and ToolController.Signup prestores the mapping at the
end. Verified against a live signup capture: viewerId=1 +
udid=62747917-93bc-454c-abb4-ef423b3c9317 produces the captured SID
dc4aac79d35fe15dfb6262e0071bb03c.
Note: this only fixes the fresh-signup path. Clients restarting with a
cached viewer_id (which skip /tool/signup entirely) still hit the same
issue — separate follow-up.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The client's PlayerStaticData.UpdateHaveUserGoodsNumByJsonData does direct
assignment on each reward_list entry's reward_num, so currency/item totals
must be the new viewer balance — not the gift delta. Fresh accounts were
seeing their cached crystal/rupy balances clobbered down to the gift counts
until the next /load/index. Matches the project_wire_reward_list_post_state
memory and the prod capture (which shows 120 rupy = baseline 20 + gift 100).
Stack [HttpPost("/tutorial/pack_open")] alias on PackController.Open. Detect
isTutorialPath via HttpContext.Request.Path; gate the type_detail rejection,
currency switch, open-count tracking, and currency reward_list entries behind
!isTutorialPath so the starter legendary pack (99047/990047, type_detail=5)
bypasses the purchasable-pack code path. After grant, set MissionData.TutorialState=100
and emit tutorial_step=100 in PackOpenResponse — this is the sole END transition,
per live-traffic capture. Add pack 99047 to test-fixture packs.json.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Stacks a second [HttpPost("/tutorial/pack_info")] absolute route on
PackController.Info so the tutorial flow resolves to the same action
and returns the same pack_config_list as /pack/info (no filtering in v1).
Adds PackControllerTests.cs with TutorialPackInfo_returns_same_list_as_pack_info
verifying byte-identical responses from both URLs.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements POST /account/update_name — writes Viewer.DisplayName and
returns an empty array per the prod capture. Includes TDD test covering
the persist side-effect.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
POST /tutorial/update echoes tutorial_step back and saves it to
Viewer.MissionData.TutorialState. is_skip=1 is handled server-side
by honoring whatever tutorial_step value the client sends (client
already sends 100 when skipping). Adds TutorialUpdateRequest DTO,
TutorialUpdateResponse DTO, injects SVSimDbContext into
TutorialController, and adds GetViewerTutorialStateAsync helper to
SVSimTestFactory.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Returns an empty data object (result_code=1 from middleware envelope).
Client uses SkipAllNetworkChecks so the response body is never read.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Client's LitJson serializer emits the C# property name verbatim — the
SetParameter param classes in Wizard/GenerateDeckCodeTask.cs use cardID /
phantomCardID, and the matching Parse() reads jsonData["cardID"]. Snake-case
keys bound to empty in msgpack deserialize, the controller saw 0 cards, and
returned INVALID_DECK — surfaced as a blank deck code in the in-game UI.
Repro lived in data_dumps/traffic.ndjson #19-20. Existing tests pass through
the same JsonPropertyName on both serialize and deserialize, so they happily
round-tripped any consistent key — adding a wire-shape regression test that
posts the literal client JSON would be the right way to catch this class of
bug in the future (out of scope here).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds the portal pair (shadowverse-portal.com deck-builder endpoints) as
anonymous routes on the app server. The translation middleware learns a new
[NoWireEncryption] attribute that skips both AES calls but keeps the rest of
the msgpack + base64 + envelope pipeline intact, matching prod's portal wire
profile observed in data_dumps/traffic_prod_deckcode.ndjson.
Storage is a 3-minute IMemoryCache — codes are anonymous-global, 4-char
lowercase alphanumeric (matches the shortest prod sample). Foil bit is
stripped on mint to match prod's normalize-on-encode behaviour.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Two issues caught in a real-client smoke run against the freshly
bootstrapped DB:
1. NRE in ShadowverseTranslationMiddleware for parameterless actions.
Five new actions (Sleeve.Info, LeaderSkin.{Ids,Products},
ItemPurchase.Info, SpotCardExchange.Top) took no parameters, but
the middleware does
`endpointDescriptor.Parameters.FirstOrDefault().ParameterType`
to discover the request DTO — `FirstOrDefault` returns null on a
zero-param action and `.ParameterType` NREs. Tests didn't catch it
because the test client POSTs plain JSON, bypassing this path.
Fix: each action now takes `BaseRequest _` matching the codebase
convention (PuzzleController.Info, BattlePassController.Info, etc.),
plus the middleware throws an actionable
InvalidOperationException pointing at the convention so the next
contributor doesn't repeat the mistake.
2. Leader-skin set sale showed up as "FREE / Claim" with empty
Includes panel after the viewer bought every skin in a series
with no configured bonus items. Root cause: ComputeRewardStatus
emitted status=1 (not_got) when set_sales_status != 0 regardless
of whether rewards.items was empty, and SkinPurchaseInfoTask.
CreateSetSaleInfo flags `is_free=true` on (is_completed &&
not_got). Prod ships status=0 when items is empty even with
set_sales_status==1 — we now mirror that.
504 tests still pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
(1) RegisterAnonymousViewer now catches the unique-violation
race (SQLSTATE 23505 on Postgres / code 19 on SQLite) and
re-reads by UDID, returning the existing row instead of
surfacing 500 to the second concurrent /tool/signup caller.
New repo test exercises the back-to-back register path.
(2) Add unique index on SocialAccountConnection (AccountType,
AccountId). The auth handler's find-or-link path claimed
this index existed as the dedup backstop; the claim was
accurate as design intent but the schema was missing. Now
matched. Comment in handler updated.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
After /tool/signup, the client has a viewer_id but no Steam social row.
The first authenticated request (typically /check/game_start) carries
the Steam ticket; if the SteamId lookup misses but the UDID resolves
to a viewer, attach the Steam social now. Subsequent requests hit the
fast SteamId path. Closes the CheckController.GameStart TODO that was
blocking fresh-client boot.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
POST /tool/signup upserts a Viewer keyed on the resolved request UDID
(via the existing SID->UDID dict). Stashes the viewer on HttpContext so
the translation middleware emits viewer_id/short_udid/udid in
data_headers. Empty data payload -- all signup outputs flow in
data_headers per spec. Idempotent: repeat signups for the same UDID
return the existing viewer.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Request mirrors LoginPostParams (device telemetry); response is empty
because all signup outputs live in data_headers (viewer_id, short_udid,
udid). MessagePackObject + Key mirrors JsonPropertyName per project
convention.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
SignUpTask.Parse validates data_headers.udid against Certification.Udid;
mismatch discards the response. Sourced from the same mappedUdid the
translation middleware uses to decrypt — never controller state. Other
endpoints carry the extra key; SignUpTask is the only reader.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Mirrors how the translation middleware resolves the per-request UDID;
needed by ToolController.Signup and the SteamSession find-or-link
branch.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Story emits story_chapter_finish:<main|limited|event>:<story_id>.
Practice emits practice_win:<difficulty>:<enemy_class_id> on win only.
Practice catalog rows use opponent NAMES (e.g. practice_win:elite:arisa)
not numeric class_ids, so captured catalog rows won't match yet. The
infrastructure is in place; bridging numeric→name is a follow-up.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
EnsureCurrentAsync now takes viewerId (was Viewer), so it works with
LoadController's AsNoTracking-loaded detached viewers.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three endpoints + 9 integration tests. Captured-data-is-catalog: viewer's
achievement Level starts at MIN(Level) per type from the catalog (not 1),
so the assembler always has a row to render against.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Tests intentionally deferred to controller integration tests (Tasks
18-21) which exercise the assembler end-to-end via the wire.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reads existing state from DB on each call (don't trust navigation
property — caller may pass it stale or double-tracked). Adds via DbSet
only, not via navigation property, to avoid EF double-tracking.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Also wires IMissionCatalogRepository + IViewerMissionRepository +
IMissionProgressService into DI. Task 17's separate DI step is now
subsumed by these registrations.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Crystal synthesis in BuyPremiumAsync is now unconditional: always remove
any crystal entry ApplyAsync may have added, then append the fresh
post-deduction total. Prevents stale on-screen balances when a retroactive
grant also touches crystal (or when no grants fire and the conditional
guard would have been the only crystal entry).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds BattlePassSalesPeriodInfoDto, BattlePassProductDto, BattlePassItemListResponse DTOs,
GetItemListAsync on BattlePassService (one product if not premium + CanPurchase, empty if
already premium or off-season), and the /battle_pass/item_list controller action.
2 new integration tests; all 408 pass.
GetOrCreateProgressAsync now persists the new row itself and catches
DbUpdateException on unique-constraint violations — concurrent /info
calls no longer throw 500s. BattlePassService no longer calls
SaveChangesAsync after the get-or-create. FormatWireDate uses a named
JstOffset constant instead of an inline TimeSpan.FromHours(9).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Also fixes BattlePassRepository.GetActiveSeasonAsync to use client-side
DateTimeOffset filtering (SQLite provider cannot translate DateTimeOffset
comparisons in LINQ WHERE/ORDER BY clauses).
- IndexResponse.BattlePassLevelInfo widened to IReadOnlyDictionary<string,BattlePassLevel>?
so any IReadOnlyDictionary impl (FrozenDictionary, wrapper, etc.) serializes correctly
instead of silently null-ing via a failed as-cast
- LoadController.Index now takes CancellationToken ct and threads it to GetLevelCurveAsync
instead of CancellationToken.None
- BattlePassRepository.ResetLevelCurveCache changed from public to internal; added
InternalsVisibleTo("SVSim.UnitTests") to SVSim.Database.csproj (was absent)
Wire IBattlePassService.GetLevelCurveAsync into LoadController so /load/index
emits the 100-entry battle_pass_level_info dict when levels are seeded.
Also adds BattlePassRepository.ResetLevelCurveCache() to bust the process-level
static cache in tests that seed levels after earlier HTTP calls have primed it
with an empty list, and updates SVSimTestFactory.SeedGlobalsAsync + the stale
Index_surfaces_seeded_globals_after_bootstrap assertion accordingly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
