Repository navigation
Conversation
SecurityCheckAction.isUseful() passed !ai->GetMaster() (a bool) to PlayerbotAIStorage::GetAI, which resolves to a guid-0 lookup that never finds a bot. The engine therefore never ran its Execute(). Removed the action, its creator registration and its "security check" entry on the "often" trigger. The "check mail" action on that trigger is kept. No behaviour change: the action never ran. Fixes Sagiroth#919
List Living Poison (16027, map 533) in the creature keep-out table as an aura-less row: pulse 5 yd (Explode 28433), keep-out 8 yd. Fixes Sagiroth#906
Add a per-row tankExempt (true for the sludge, false for Living Poison) and pass the row to CreatureAuraChaseExempt and CreatureAuraCombatExempt. End points, the shared hazard snapshot and gap closers skip only exempt rows for a tank. Docs note that Living Poison is not selectable.
… battleground reach limit (Sagiroth#904)
Sagiroth#906) Raise the Living Poison keep-out from 8 to 10 yd. A lethal row (not tank-exempt) skips the 2 s no-landing back-off of the step-out and, when every landing is refused, steps straight away from the creature.
…emoved sEluna global
…rify_all checks Eluna guards
… code calls into Eluna
…epout fix: bots keep out of Naxxramas Living Poison
fix(combat): set behind only in melee reach, give up after 6 s (Sagiroth#911)
…ered fix: healers stay put when another healer covers the ally (Sagiroth#904)
…check fix: remove never-running SecurityCheckAction (Sagiroth#919)
Eluna ON builds: fix the bot XP hook, check Eluna guards, add Eluna CI jobs
A 39-companion raid took about 6 minutes of one-by-one .bot hire calls and a manual raid conversion. .bot hire raid <preset|composition> [tier] plans the roster (role counts such as 3t 12h 24d drawn from a class spread, an explicit class list, or a config preset), cuts it to HireMaxBotsPerPlayer minus the current companions, quotes one total fee, converts an overflowing party to a raid, then hires each slot through the existing Hire() path. It stops at the first failed hire and reports one summary line plus a TBM:HIRE_RAID line. New keys AiPlayerbot.HireRaidSpread and AiPlayerbot.HireRaidPresets. Hire() now shares its gates, fee formula and race/gender default with the raid path. Refs Sagiroth#900.
The raid command called Hire() up to 39 times inside the chat command, and each call rescans pool accounts with synchronous queries. Plan, quote and refusals stay in the command; an accepted raid is queued and worked off two hires every 250 ms from Update(), with a closing summary and the TBM line when the job ends. A running job reuses the previous slot's account. The cap now also respects free raid seats (group members plus companions still mustering), the party is converted by eventual size (members + mustering + planned > 5) and the reply says when it became a raid. Only one raid hire runs per player at a time. Refs Sagiroth#900.
A master who started .bot hire raid solo had no group to convert when the job began, so once four companions joined, Hire() refused the next seat as GroupFull and the job stopped near 16 of 39. StepRaidJobs now converts the master's party to a raid (leader only) before each hire when the planned raid outgrows five, and tells the player. A far teleport took the master out of the world for a moment and the job dropped silently. The job now looks the master up NotInWorld and waits through the loading screen; only a master who is gone drops it.
…ecurity) Rebased onto feat/dungeon-self-clear: a raid quote now uses the same HireFeeCharged rule as a single hire, so a staff account is not refused for a fee it would not pay.
…ou leave the group
Player::EquipNewItem never rolls a random property, and the seed equip loop only stamped the best-for-spec suffix. Resistance-only jewelry groups (e.g. 3301) score 0 for every spec, so those rings and necklaces equipped stat-less. Fall back to a loot-style GenerateItemRandomPropertyId roll; same for the <30 fallback and bot quest objective items.
In an LFT/pool group every usable drop is usable by someone, so the 'never Need against a player' fallback (unseen player + CanUseItem OK = competes) demoted every upgrade NEED to GREED even when no player had voted. Unseen players now compete only via a NEED the bot actually saw; a bot the player directly drives (alt / hire with its master) keeps the old yield-on-usable behaviour.
…anches ci: delete the branch of a PR closed without merge
docs: restructure the documentation for players and server admins
…cs/server/dashboard.md)
Dashboard: read-only Agent maps tab for turtle-agent run maps (Sagiroth#1102)
Working branch only (pjerra/turtle-agent#259). Brings dev 64d6a44 into feat/dungeon-self-clear 48c7ba8. - WandPolicy/PlayerbotAI: keep the run leader's heal-only healer and dev's pull-hold gate and pet stance (Sagiroth#973) side by side. - Dashboard: dev's live Dungeons tab (Sagiroth#895) keeps its names; the run trace tab becomes "Dungeon runs" (/api/v1/dungeon-runs, tab-dungeon-runs, dungeon-runs.js); its poll goes through the refresh gate. - Docs: dev's new layout; the run leader, boss tactics and run trace pages move to docs/play and docs/server, their settings and commands into the new pages; docs/PROVENANCE.md stays dropped as on dev.
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.
Brings the dungeon-clear branch up to current dev: #1050, #904, #1057, #1064, #1075, #872 alt-bots, #1109 and everything else merged to dev since, so the next test build has the fixes 19z16 lacked. One merge commit (parents: feat 48c7ba8, dev 64d6a44); 25 conflicted files resolved by keeping both sides (WandPolicy healOnly + pullHold, PlayerbotAI, pull/meal fix made twice now once, BotCommandContext comments, docs moved into dev's new layout, PROVENANCE.md dropped as on dev). Dashboard: dev's live Dungeons tab is unchanged; the run-trace tab is renamed Dungeon runs (/api/v1/dungeon-runs, #/dungeon-runs) with a test that mounts both routes. Checks: verify_all 1-6, the standalone policy tests (incl. 8 not in the list), run-leader wiring, plan and boss-tactics SQL tests, Go and node checks pass; compiled on a VM against the core (rc=0, no warnings). Reviews: Codex SHIP, Opus cold review SHIP (no HIGH/MEDIUM). Not run live; dashboard not opened in a browser. Follow-ups noted by the review (comment/doc nits and a missing pullHold unit test) go in a small follow-up.