Packet generator quests#25
Open
Seltraeh wants to merge 10 commits into
Open
Conversation
- New assets/net/unit.kdl: UnitFavoriteEntry/Req/Resp, UnitEvoReq/Resp (req/resp shapes for the UnitFavorite and UnitEvo handlers) - assets/net/handlers.kdl: import ../net/unit.kdl - assets/mst/unit.kdl: correct numeric fields from i32::int to i32::str, fix comma-separated position fields to plain str, add missing after_image field, replace all TODO docs with descriptive strings
- New assets/net/town.kdl: TownUpdateReq/Resp, TownFacilityUpdateReq/Resp, TownKarmaPayment, TownFacilityUpdateEntry, TownLocationUpdateEntry (handler request/response shapes for the two known town action handlers) - assets/net/handlers.kdl: import ../net/town.kdl TownFacilityUpdate (id 8v43tz7g) receives the complete desired facility + location state from the client along with the total karma cost; this KDL documents those wire keys. TownUpdate (id CuQ5oB8U) request shape is not yet fully reversed; stub documented accordingly.
Five new KDL schema definitions covering the MST tables needed to drive
the quest/campaign/battle systems in the server.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
NEW FILES
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
assets/mst/mission.kdl — MissionMst (hash 2I9V0o6J)
F_MISSION_MST_1 and F_MISSION_MST_2. Covers all story campaign missions
(IDs 1–~1200) and Uncharted Frontier end-game missions (IDs 1005000+).
Key fields:
j28VNcUW id — mission primary key
9C64Qwe0 land_id — land/continent this mission belongs to
VjCY7rX4 area_id — area/region within the land
MHx05sXt dungeon_id — dungeon/node within the area
69vnphig stamina_cost — energy cost to enter
d96tuT2E exp — base EXP reward
Rs7bCE3t zel — base zel reward
HTVh8a65 karma — base karma reward
8f4NYKxb battle_group_ids — colon-sep order:groupId pairs linking to
battle wave data (e.g. '3:10100,1:10101')
3sRN9BPS bgm_map — map/field BGM asset name
nDf28LtU bgm_battle — battle BGM asset name
1e6jxQzf npc_encounter_id — links to MissionNpcUnitMst; 0 = no NPC
assets/mst/mission_ep3.kdl — MissionEp3Mst (hash 2I9V0o6J)
F_MISSION_EP3_MST. Episode 3 (Summoner's Road) supplemental parameters.
Augments MissionMst entries (IDs 20100+) with ep3_params: a KEY:VALUE
string carrying EXP multipliers and ARM_EXP_RATE overrides.
assets/mst/mission_npc_unit.kdl — MissionNpcUnitMst (hash 1e6jxQzf)
F_MISSION_NPC_UNIT_MST. Fully-statted NPC/boss units used in campaign
cutscenes and encounter segments. Links to MissionMst via encounter_id
and to SkillMst via skill_id / extra_skill_id / leader_skill_id.
Carries HP/ATK/DEF/REC, BB/SBB/UBB skill IDs, bb_fill_rate, and
extra_passive_skill_id.
assets/mst/mission_script.kdl — MissionScriptMst (hash N4XVE1uA)
F_MISSION_SCRIPT_MST. Per-mission cutscene / dialogue script entries.
The script_data field is an '@'-separated sequence of scene commands
(phase:command:params) that drives the in-game dialogue engine. The N4XVE1uA
hash matches the dungeon-state field seen in DungeonEventUpdate requests.
assets/mst/skill.kdl — SkillMst (hash 8aiBoHg5)
F_SKILL_MST (parts 1-9, merged). All player BB/SBB/UBB skills, enemy
auto-attack entries, and leader/extra skill effect definitions.
Key fields:
nj9Lw7mV id — skill primary key (also used as reference key
in MissionNpcUnitMst and unit rows)
h6UL9A1B skill_type — category (1=attack, 2=heal, 3=buff, …)
AR7y0sY4 skill_rank — tier (1=BB, 2=SBB, 3=UBB)
hjAy9St3 process_id — colon-sep effect chain string
eyUo6a8c effect_frame — animation frame data
6Aou5M9r damage_frame — damage-hit frame data
n9h7p02P drop_check_cnt — BC drop-checks per hit
iNy0ZU5M element — elemental override (0=none, 1-6=Fire…Dark)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
USAGE NOTES
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
These schemas are required by the MissionStart battle-engine seed path:
mission.kdl → look up stamina_cost, rewards, battle_group_ids
skill.kdl → look up BB/SBB/UBB frame data and drop-check counts
mission_npc_unit → NPC encounter stat blocks for campaign dialogue
mission_script.kdl → cutscene scripts for the scenario viewer
mission_ep3.kdl → late-game EXP rate overrides
Not yet wired: the server currently uses hardcoded mission-10 battle data.
Hooking MissionStart up to these MST tables is the next task on the
packet-generator-quests branch.
…spatch key t6bRQfln)
… use datetime-unix::str for timestamp fields
…ds, justify the str-kept ones with captured wire data, move Req/Resp into handlers.kdl Audited every contested type-change in PR decompfrontier#21 against real values in deploy/system/unit.json (2291 rows) to decide which `[i32::str]::sep(comma)` → `str` changes were justified and which were not. Reviewer was right on five of the six position fields; we have hard evidence for the remaining three exceptions. ────── assets/mst/unit.kdl — REVERTED to `[i32::str]::sep(comma)` ────── home_img_pos (1W9CxaFK) — all 2291 rows integer detail_img_pos (6z54rgb3) — all 2291 rows integer confirm_img_pos (MYK1fq6c) — all 2291 rows integer skill_cut_image_pos(7hLR6pDN) — all 2291 rows integer hp_disp_pos (3BpHN6VD) — all 2291 rows integer Earlier doc claimed "may contain floats" but the actual data has zero decimals across the full table. The original `[i32::str]::sep(comma)` handles signed integers (including the observed negatives like -1, -100, -244) without issue. Each field now carries concrete example values in the doc string. ────── assets/mst/unit.kdl — KEPT as `str` (justified, with evidence) ────── unknown / 5SUvj4tM (overdriveStats) Values have a TRAILING COMMA: "100,100,100,0,", "100,100,100,5000,", and "" for most units. `[i32::str]::sep(comma)` would lose the trailing comma on round-trip. Kept `str` for byte-faithful passthrough. bad_state_resists / CEeqs63b Values are 6 colon-separated percentages but at least one captured row contains a literal SPACE CHARACTER in a slot: "0:0:0:0: :0". `[i32::str]::sep(colon)` cannot parse the space as i32. Kept `str`. summon_image_pos / KC3Jk8Br 56 of 2291 rows contain DECIMAL FLOATS in the width/height columns, e.g. "41,231,560.5,387.5,0", "175,314,302.5,302.5,0". Integer array codec would truncate or fail. Kept `str`. (This is the only one of the six position fields where the "may contain floats" claim was true.) ────── assets/net/unit.kdl — pared back to shared types only ────── Per reviewer's convention guidance ("request/response network handlers should go in handlers.kdl; keep separate kdl files only for types that might need to be reused"), removed: UnitFavoriteReq, UnitFavoriteResp, UnitEvoReq, UnitEvoResp These now live in handlers.kdl alongside the other request/response wrappers. UnitFavoriteEntry stays in unit.kdl because both the req and resp reference it (shared type). UnitFavoriteEntry doc now carries a concrete example wire-JSON capture: {"edy7fq3L":"12345","5JbjC3Pp":"1","2pAyFjmZ":"0"} ────── favorite field type=i32::str (response to bool suggestion) ────── Reviewer noted favorite (5JbjC3Pp) "sounds as a boolean rather than i32". Semantically true, but the actual wire format is a QUOTED decimal integer: "1" / "0". The available bool modifiers don't match: bool::int → emits unquoted `1`/`0` (number, not string) bool::str → emits unquoted `true`/`false` (the wrong tokens entirely) No `bool::str-num` modifier exists in the schema language, so i32::str is the only modifier that round-trips the literal wire bytes. Doc string updated to explain this explicitly so a future reader doesn't loop on it again. ────── assets/net/handlers.kdl — new entries ────── Added the four moved Req/Resp wrappers. UnitEvoReq/Resp remain placeholders (the old hand-written handler never parsed the body in detail) with docs pointing at `extract_createbody("UnitEvoRequest")` and response_class=UnitOpeEvoResponse for the future implementer.
…rkers
Cross-checked every MST `hash` declaration we authored (mission, mission_ep3,
mission_npc_unit, mission_script, skill) and the upstream-inherited unit.kdl
against three evidence sources:
1. IDA `response_mappings.json` — what hash maps to what response class
2. IDA `typed_responses/*.hpp` — generated readparam structs
3. `deploy/system/*.json` — actual decoded MST data on disk
Found and fixed FIVE wrong hash declarations and made the rest of the
mission/skill KDL hashes explicit about their unverified status.
────── unit.kdl — UnitMst hash corrected (HIGH CONFIDENCE) ──────
Was: json UnitMst { hash "JYFGe9y6" doc "TODO" ... }
Now: json UnitMst { hash "2r9cNSdt" doc "F_UNIT_MST — VERIFIED..." }
`JYFGe9y6` is actually the hash for F_UNIT_EXP_PATTERN_MST — confirmed
both by IDA (response_class=UnitExpPatternMstResponse, is_mst=true) and
by `deploy/system/unit_exp_pattern.json` having `JYFGe9y6` as its
top-level key. `deploy/system/unit.json` has `2r9cNSdt` as its
top-level key, which is the real F_UNIT_MST hash.
`UnitExpPatternMst` definition in the same file kept its original
`JYFGe9y6` hash — that one was correct all along (it's a different
struct in the file; the upstream KDL had two structs sharing the
wrong hash for the wrong one).
────── mission.kdl — hash corrected (per IDA, UNVERIFIED on disk) ──────
Was: hash "2I9V0o6J" → IDA: CampaignMissionInfoResponse, is_mst=FALSE
Now: hash "oXeC1Ak9" → IDA: MissionMstResponse, is_mst=TRUE
The old `2I9V0o6J` was a network response key, not an MST hash —
fundamentally wrong category.
────── mission_ep3.kdl — hash corrected (per IDA, UNVERIFIED on disk) ──────
Was: hash "2I9V0o6J" ← DUPLICATE of mission.kdl's wrong hash
Now: hash "vE0Qdg33" → IDA: MissionEp3MstResponse, is_mst=TRUE
────── mission_npc_unit.kdl — UNVERIFIED marker added ──────
Hash `1e6jxQzf` appears only as a field key in IDA, never as a
top-level response_key for any *MstResponse class. No decoded
`mission_npc_unit.json` on disk to confirm. Marked UNVERIFIED in
the doc string with explicit instructions to verify post-mstdec.py.
────── mission_script.kdl — UNVERIFIED marker added ──────
Hash `N4XVE1uA` is heavily dual-use (LoginInfoResp field, dungeon-state
field in DungeonEventUpdate, arena MST scenario-script slot). Whether
it's also the F_MISSION_SCRIPT_MST top-level key cannot be confirmed
without decoded data. Marked UNVERIFIED.
────── skill.kdl — UNVERIFIED marker added ──────
`8aiBoHg5` IS in IDA but as `UnitSkillMstResponse` (the per-unit skill
assignments table), not a standalone `SkillMstResponse` — IDA has no
class with that name at all. Marked UNVERIFIED.
────── Method ──────
For each KDL `hash`:
1. grep response_mappings.json for the hash → if response_class is
something other than what the KDL claims, it's wrong.
2. Check deploy/system/<file>.json's top-level key directly:
`head -c 30 <file>.json` shows it.
3. If neither check resolves, mark UNVERIFIED in the doc string
with a clear instruction to verify post-mstdec.
Five response classes covered (F5Vs19mb / 9Q1Lq5FS / UT1SVg59 / 5PR2VmH1 /
4MCxgS5p) plus the mission MST schema fixes that surfaced during testing:
net/mission.kdl
- Expanded from 16 lines (MissionBreakInfo stub) to 538 lines with
full request + response definitions for MissionStart, MissionEnd,
MissionBreak, MissionReward.
- MissionReward: all 19 fields named (setLvupFlg / setBeforeLv /
setIncExp / setActualZel / setActualKarma / setRewardUnits /
setClearBonus / setRein* / setClear* / setExpUp/ZelUp/KarmaUpType /
setExp/Zel/KarmaGuild / setEToken / setRepeatClearBonus). The
LvupFlg + BeforeLv pair is the actual level-up animation control;
the d96tuT2E "incExp" field is just the displayed value, not a
cascade trigger as previously assumed.
- MissionStartInfo: corrected field semantics — J3stQ7jd is
FriendPoint (not account_id), Z0Y4RoD7 is DeckNum (not
ActiveGuildDeck in this context), 1VagK32J is ReinforceInfo
(was placeholder).
net/user.kdl
- UserClearMissionInfo: was a mock/INVALID stub; now has all 7 fields
with their decoded setters (setUserID / setMissionID / setClearDate /
setClearCnt / setMostTime / setGetBc / setGetHc). All
std::string-typed even when the value is numeric — BF stores them
as quoted strings.
mst/mission.kdl
- Marked 7 fields `optional #true` for the sentinel-row null values
Glaze otherwise rejects (wHN6nfh9, 8f4NYKxb, N4XVE1uA, L8iA9M6c,
3sRN9BPS, nDf28LtU, PYgkvcsy).
- Reverted HSRhkf70 and 1e6jxQzf from i32::str to str — both
heterogeneous across the 3433-row dataset (single-int + comma-pair
/ colon-list mixed shapes that fail single-int parse).
mst/user_level.kdl
- exp doc: per-level chunk semantics confirmed (not cumulative).
mst/gacha.kdl
- Doc cleanup for GachaInfoMst.
net/handlers.kdl
- import "../mst/mission.kdl" so MissionMst gets generated into the
C++ headers (was previously dead KDL).
Audit findings are committed to KDL `doc` strings as setter names +
example values per handbook section 6.13 evidence rules. Audit artifacts
themselves stay private (parent repo's tools/ida/ is gitignored).
IDA readParam audits (server-side handbook §8.32, §8.33, §8.35) decoded three response classes whose schemas were stubs: - net/friends.kdl: new FriendInfo (tojMy68W, 50 fields) — populates FriendInfoList, the singleton the squad-select friend picker / Social tab Friend Info dialog read from. Setter names from the readParam dispatch at 0x13D93C0 plus the GuildRaidFriendSelectRequest::createBody enumeration at 0x1C5B388. Notes the contextual overloads documented in §8.34 (iNy0ZU5M is uint32_t here, NOT string like UserUnitInfo; Ge8Yo32T is EquipItemID here, NOT MissionID like ReinforcementInfo; 2Fh3J7ng is TeamLv here, etc.). - net/reinforcement_info.kdl: ReinforcementInfo (xZH6EIQ7, 47 fields) + FixedReinforcementInfo (T_FIXED_REINFORCEMENT, 21 fields). Renamed formerly-unk_* fields to setTeamLv / setFriendType / setLastLoginDate / setNormalFriendPoint per the new IDA audit. T_FIXED_REINFORCEMENT confirmed as the literal wire key (uppercase with underscores — unusual for BF but real, see audit at 0x13950B8 in GameResponseParser::getResponseObject). - net/handlers.kdl: FriendGetResp now wraps both [ReinforcementInfo] under xZH6EIQ7 (canonical per the captured prod response) AND [FriendInfo] under tojMy68W (for any UI that reads from FriendInfoList). Imports net/reinforcement_info.kdl explicitly. mst/mission.kdl / mst/mission_script.kdl / mst/unit.kdl / mst/user_level.kdl / mst/gacha.kdl: doc + type touch-ups from cross-referencing against the audit findings.
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
2 participants
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.
Rebased to the packet generator
Added: