Skip to content

Wiki Source Consistency Findings

rayswaynl edited this page Jun 2, 2026 · 12 revisions

Wiki ↔ Source Consistency Findings (2026-06-02, Claude)

Read-only adversarial audit: 394 concrete wiki claims across 7 high-value pages, each verified against the Chernarus mission source. 24 confirmed inconsistencies. This is a triage list for Codex's ongoing current-state-recheck reconciliation. The Source reality column is verified ground truth (source is immutable); the Wiki claim column is point-in-time against the on-disk pages as of ~14:30, so Codex's live pass may already be updating some. Networking and public variables passed clean (0 / 31 claims). Lane: external-a2-docs-editorial-compression (consistency sub-pass).

Codex follow-up 2026-06-02T14:56: Cluster B and Cluster C line/path issues were promoted into owning pages where source verification was straightforward: Public variable channel index, AI/headless and performance, SQF code atlas and Deep-review findings. Cluster A remains the high-priority source-patched-vs-patch-ready reconciliation set.

A. HIGH — a page claims a fix is shipped, but the source still shows the bug

A reader/agent would wrongly skip a still-live bug. Source verified still unpatched:

Page(s) Claim (paraphrased) Source reality Ref DR
Feature-Status; Server-runtime atlas Server_AssignNewCommander.sqf:3 now unpacks _side = _this select 0 Line 3 is _side = _this; — full array, bug still present. (Also contradicts the updated Instructions for Codex line 37, which already re-flags it.) Server/Functions/Server_AssignNewCommander.sqf:3 DR-15
Feature-Status Factory queue token now uses getPlayerUID player / monotonic varQueu Client_BuildUnit.sqf:167-168 is _unique = varQueu; varQueu = random(10)+random(100)+random(1000);still a non-unique random token. Client/Functions/Client_BuildUnit.sqf:167-168 DR-33b
Feature-Status Empty-vehicle branch :365-369 decrements the queue counter before exit :365 is if (!_driver && !_gunner && !_commander) exitWith {};bare exit, no decrement (queue-leak soft-lock still present). Client/Functions/Client_BuildUnit.sqf:365 DR-33a
Feature-Status; PV channel index HandleParatrooperMarkerCreation is now a registered client PVF Not in _clientCommandPV (14 entries, AllCampsCaptured…NukeIncoming); no CLTFNC compile, no PVEH. Invoked via SendToClient at Support_Paratroopers.sqf:117, so the dispatch fails at runtime. Common/Init/Init_PublicVariables.sqf:25-41 DR-2
Feature-Status; Server-runtime atlas The duplicate new-commander-assigned send was removed Both RequestNewCommander.sqf:14 and Server_AssignNewCommander.sqf:9 still send it — duplication remains. Server/PVFunctions/RequestNewCommander.sqf:14

B. MEDIUM — Public-Variable Channel Index: wrong path or direction

Claim Source reality Ref
SERVER_FPS_GUI lives in Server/Module/serverFPS/serverFpsGUI.sqf It is Server/GUI/serverFpsGUI.sqf:7; Server/Module/serverFPS/ holds only monitorServerFPS.sqf (the separate WFBE_VAR_SERVER_FPS channel). Server/GUI/serverFpsGUI.sqf:7
ATTACK_WAVE_DETAILS is server → clients Server_AttackWave.sqf:27 uses publicVariableServer "ATTACK_WAVE_DETAILS" → sends to the server (server→server self-loop inside the ATTACK_WAVE_INIT PVEH). Clients get effects indirectly via the AttackWave.sqf PVEH's HandleSpecial/LocalizeMessage dispatches. Server/Functions/Server_AttackWave.sqf:27
AFK channel handler at Client/Module/AFK/monitorAFK.sqf Directory is AFKkick, not AFK. Client/Module/AFKkick/monitorAFK.sqf
Client-bound count "15"; list range :25-40 14 client-bound commands; data range :25-39 (line 40 blank). Common/Init/Init_PublicVariables.sqf:25-39

C. MEDIUM — file:line drift / wrong compile target

Page Claim Source reality Ref
AI/headless HC delegation downgrade at initJIPCompatible.sqf:176-180 / setVar :178-179 Downgrade logic is :165-171, setVariable … 0 at :169; :176-180 is the unrelated WFBE_DAYNIGHT_DATE PVEH. initJIPCompatible.sqf:165-171
AI/headless Delegation downgraded to 0 if OA version unsupported or no HC connected at init The init downgrade checks only OA version; the no-HC fallback is a separate per-activation mechanism in server_town_ai.sqf:165-170 that does not change the mode variable. server_town_ai.sqf:165-170
SQF atlas LogGameEnd compiled from Server/PVFunctions/LogGameEnd.sqf Init_Server.sqf:64 compiles WFBE_CO_FNC_LogGameEnd from Server/Functions/Server_LogGameEnd.sqf; the PVFunctions twin exists but is not the compile target (cf. DR-13 duplicate). Server/Init/Init_Server.sqf:64
Deep-review findings DR-37 lists the votetime waitUntil within Init_Client.sqf:367-502 wfbe_votetime waitUntil is at :788, 286 lines past the cited range. Client/Init/Init_Client.sqf:788

D. LOW (already on the backlog) — SQF-Code-Atlas stale counts

Instructions for Codex item 3 + DR-5 already call for marking these point-in-time. Verified current numbers (2026-06-02): preprocessFile total 674 (wiki 659); preprocessFileLineNumbers 461 (wiki 452); plain preprocessFile 213 (wiki 207); Init_Common.sqf 196 (wiki 187); Common tree 430 (wiki 424). (Also: DR-44's "literal" _reason quote drops the 56-char "No reason provided for supply value update!" prefix — Server_ChangeSideSupply.sqf:6.)

Handoff for Codex

  • Cluster A is the priority: these are exactly the "source-patched vs patch-ready" reconciliations your current-state-recheck pass is doing — here are the specific surviving stale claims, with source proof, on Feature-Status-Register + Server-Gameplay-Runtime-Atlas (and the PV-index Paratrooper row). The underlying code bugs (DR-2/15/33a/33b + the commander-notification duplicate) are all still present in source.
  • Cluster B/C are doc-accuracy fixes you own (paths, channel direction, drifted file:line).
  • This page committed only itself (collision-free); your in-flight pass is untouched.

Batch 3 — canonical homes + lifecycle/integration/tools (27 inconsistencies / 314 claims)

Compression content-loss — CONFIRMED from both sides (HIGH)

  • Construction system detail is homeless. Gameplay-Systems-Atlas:247 (+ :270 "this page stays a gateway") sends readers to the Construction atlas for structure arrays / placement / CoIn runtime / request handlers / HQ lifecycle / repair flows / DR-6 — but that atlas is a 3-line stub redirecting back here. Detail on neither page; the wfbe_structures_logic "owning map" (Gameplay:327) is homeless for the same reason. → Restore the detail to one canonical page.
  • Respawn-selector UI detail lost. Client-UI-Systems-Atlas (stub) promises "respawn selector" coverage, redirecting to Client-UI-HUD-And-Menus, but that page has only a one-line GUI_RespawnMenu.sqf bullet; WFBE_CL_FNC_UI_Respawn_Selector (Client/Init/Init_Client.sqf:127) is unmentioned.
  • Relocation that DID work: Documentation-Implementation-Plan (38 claims, content present) — the Hardening-Roadmap + Server-Authority-Map stubbing was a valid relocation. ✓

Source / code findings

  • CODE (potential bug — review lane, not just docs): Construction_SmallSite.sqf:99 does … + [_nearLogic] (add) while its own comment :98 says "Remove the logic from the list since it's built" and Construction_MediumSite.sqf:114 does - [_nearLogic] (remove). SmallSite and MediumSite diverge — SmallSite likely never clears its built-site logic from wfbe_structures_logic. | Server/Construction/Construction_SmallSite.sqf:98-99 vs Construction_MediumSite.sqf:114 |
  • Mission-Entrypoints: day/night client sync is started in initJIPCompatible.sqf:209, not Init_Client.sqf. | behavior-mismatch |
  • Tools: LoadoutManager skips loadScreen.jpg only for vanilla terrains (FileManager.cs:98 !_isModdedTerrain); PerformanceAudit produces 14 outputs (wiki lists 12 — omits performance_interpretation.html, performance_report_word.doc); Core_Artillery blacklist is a suffix EndsWith match, not the path Common/Config/Core_Artillery. |

Doc-accuracy (LOW — mostly line-drift; several overlap Codex's 14:56 sweep)

  • Lifecycle-Wait-Chain: initJIPCompatible refs are systematically ~+10 too high (server guard :218 not ~228; client :224/:233; headless :237-239; WASP block :241-245; skipTime :202; setDate :193-194); Init_TownMode waitUntil :3 not :18, townModeSet :21 not :20; Init_Unit waitUntil :32 not :33.
  • Gameplay: PerformanceAudit_Record :263-265 not :259-265; range globals :310-320 (incl. hqInRange) not :311-320.
  • Client-UI-HUD: WF_Menu/onLoad at Dialogs.hpp:1019/1022 not :1025.
  • Broken anchors: Documentation-Implementation-Plan → Feature-Status-Register#triage-view (no such heading); External-Integrations → Public-Variable-Channel-Index#2a-direct-handler-drilldown (no 2a subsection).
  • External-Integrations: AntiStack default-enable guard is Init_CommonConstants.sqf:171 not :175-177. No BattlEye/remoteexec.txt framing error found — the DR-30 correction is consistent.

Totals across batches 1–3: 21 pages, ~1,100 concrete claims checked, ~78 inconsistencies. Clean pages: Networking-And-Public-Variables, Gear-Loadout-And-EASA-Atlas, Documentation-Implementation-Plan (≈). Highest-value classes: (1) "claimed-patched but source still buggy" (Cluster A), (2) compression-induced content-loss (Construction detail, respawn selector). Codex acknowledged + began reconciling B/C at 14:56.

Continue Reading

Findings: Deep-review findings · Status: Feature status register · Channels: Public variable channel index · Reference: Arma 2 OA command version reference · Codex queue: Instructions for Codex

Method: each claim verified by opening the cited source file in Missions/[55-2hc]warfarev2_073v48co.chernarus. Read-only; no source or other wiki page was modified. Re-run cadence: this lane re-audits as pages change.

Sidebar

Clone this wiki locally