| Age | Commit message (Collapse) | Author |
|---|
|
* change Client::component to return a reference
* write docs
* merge main
* remove unused parking_lot feature
|
|
|
|
* move the Client struct out of azalea-client into azalea
* actually add client impls in azalea
|
|
|
|
|
|
might add to ci later, unsure how to do it without adding significant friction for contributors though
|
|
|
|
|
|
|
|
* start implementing sneaking
* fix horizontal_collision being inverted and cleanup
* clippy
* change dimensions and eye height based on pose
* proper support for automatically crouching in certain cases
* fix anticheat issues
* add line to changelog and update a comment
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
functions
|
|
|
|
|
|
|
|
|
|
* basic pathfinder mining poc
* mining descending and autotool
* pathfinder mining descending
* pathfinder fixes
* allow disabling pathfinder miner and other fixes
* small optimization to avoid chunk vec iter lookup sometimes
* seeded rng in pathfinder bench
* consistently use f32::INFINITY
this brings performance much closer to how it was before
* astar heuristic optimization from baritone
* add downward_move
* fix downward move execute
* avoid liquids and falling blocks when mining
* fix COST_HEURISTIC
* fix to not path through flowing liquids
* only reset pathfinder timeout while mining if the block is close enough
* cache mining costs of block positions
* fix mine_while_at_start and move PathfinderDebugParticles to its own module
* add ReachBlockPosGoal
in other news: azalea's sin/cos functions were broken this whole time and i never noticed
* clippy and add things that i accidentally didn't commit
* improve wording on doc for azalea::pathfinder
|
|
* replace bevy FixedUpdate with azalea GameTick
* Update to Bevy 0.12.1
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
|
|
|
|
* add Client::attack
* partially implement attack cooldowns
* attack speed modifiers
* don't care clippy
---------
Co-authored-by: mat <git@matdoes.dev>
|
|
* more mining stuff
* initialize azalea-tags crate
* more mining stuff 2
* mining in ecs
* well technically mining works but
no codegen for how long it takes to mine each block yet
* rename downloads to __cache__
it was bothering me since it's not *just* downloads
* codegen block behavior
* fix not sending packet to finish breaking block
* mining animation 🎉
* clippy
* cleanup, move Client::mine into a client extension
* add azalea/src/mining.rs
---------
Co-authored-by: mat <git@matdoes.dev>
|
|
* update to bevy 0.11
* clippy
---------
Co-authored-by: mat <git@matdoes.dev>
|
|
|
|
* start adding azalea-inventory
* design more of how inventories are defined
* start working on az-inv-macros
* inventory macro works
* start adding inventory codegen
* update some deps
* add inventory codegen
* manually write inventory menus
* put the inventories in Client
* start on containersetcontent
* inventory menu should hopefully work
* checks in containersetcontent
* format a comment
* move some variant matches
* inventory.rs
* inventory stuff
* more inventory stuff
* inventory/container tracking works
* start adding interact function
* sequence number
* start adding HitResultComponent
* implement traverse_blocks
* start adding clip
* add clip function
* update_hit_result_component
* start trying to fix
* fix
* make some stuff simpler
* clippy
* lever
* chest
* container handle
* fix ambiguity
* fix some doc tests
* move some container stuff from az-client to azalea
* clicking container
* start implementing simulate_click
* keep working on simulate click
* implement more of simulate_click
this is really boring
* inventory fixes
* start implementing shift clicking
* fix panic in azalea-chat i hope
* shift clicking implemented
* more inventory stuff
* fix items not showing in containers sometimes
* fix test
* fix all warnings
* remove a println
---------
Co-authored-by: mat <git@matdoes.dev>
|
|
* replace 0.9.1 with 0.10.0
* start migrating to bevy .10
* well it compiles
* doesn't immediately panic
* remove unused imports
* fmt
* delete azalea-ecs
* make RelativeEntityUpdate an EntityCommand
* fix a doc test
* explain what FixedUpdate does
|
|
* start fixing stuff where systems run in the wrong order
* fix ordering ambiguity
* add debugging guide
* some fixes
* fix panic for swarms
* fix some warnings
|
|
|
|
* add EntityData::kind
* start making metadata use hecs
* make entity codegen generate ecs stuff
* fix registry codegen
* get rid of worldhaver
it's not even used
* add bevy_ecs to deps
* rename Component to FormattedText
also start making the metadata use bevy_ecs but bevy_ecs doesn't let you query on Bundles so it's annoying
* generate metadata.rs correctly for bevy_ecs
* start switching more entity stuff to use ecs
* more ecs stuff for entity storage
* ok well it compiles but
it definitely doesn't work
* random fixes
* change a bunch of entity things to use the components
* some ecs stuff in az-client
* packet handler uses the ecs now
and other fun changes
i still need to make ticking use the ecs but that's tricker, i'm considering using bevy_ecs systems for those
bevy_ecs systems can't be async but the only async things in ticking is just sending packets which can just be done as a tokio task so that's not a big deal
* start converting some functions in az-client into systems
committing because i'm about to try something that might go horribly wrong
* start splitting client
i'm probably gonna change it so azalea entity ids are separate from minecraft entity ids next (so stuff like player ids can be consistent and we don't have to wait for the login packet)
* separate minecraft entity ids from azalea entity ids + more ecs stuff
i guess i'm using bevy_app now too huh
it's necessary for plugins and it lets us control the tick rate anyways so it's fine i think
i'm still not 100% sure how packet handling that interacts with the world will work, but i think if i can sneak the ecs world into there it'll be fine. Can't put packet handling in the schedule because that'd make it tick-bound, which it's not (technically it'd still work but it'd be wrong and anticheats might realize).
* packet handling
now it runs the schedule only when we get a tick or packet :smile:
also i systemified some more functions and did other random fixes so az-world and az-physics compile
making azalea-client use the ecs is almost done! all the hard parts are done now i hope, i just have to finish writing all the code so it actually works
* start figuring out how functions in Client will work
generally just lifetimes being annoying but i think i can get it all to work
* make writing packets work synchronously*
* huh az-client compiles
* start fixing stuff
* start fixing some packets
* make packet handler work
i still haven't actually tested any of this yet lol but in theory it should all work
i'll probably either actually test az-client and fix all the remaining issues or update the azalea crate next
ok also one thing that i'm not particularly happy with is how the packet handlers are doing ugly queries like
```rs
let local_player = ecs
.query::<&LocalPlayer>()
.get_mut(ecs, player_entity)
.unwrap();
```
i think the right way to solve it would be by putting every packet handler in its own system but i haven't come up with a way to make that not be really annoying yet
* fix warnings
* ok what if i just have a bunch of queries and a single packet handler system
* simple example for azalea-client
* :bug:
* maybe fix deadlock idk
can't test it rn lmao
* make physicsstate its own component
* use the default plugins
* azalea compiles lol
* use systemstate for packet handler
* fix entities
basically moved some stuff from being in the world to just being components
* physics (ticking) works
* try to add a .entity_by function
still doesn't work because i want to make the predicate magic
* try to make entity_by work
well it does work but i couldn't figure out how to make it look not terrible. Will hopefully change in the future
* everything compiles
* start converting swarm to use builder
* continue switching swarm to builder and fix stuff
* make swarm use builder
still have to fix some stuff and make client use builder
* fix death event
* client builder
* fix some warnings
* document plugins a bit
* start trying to fix tests
* azalea-ecs
* azalea-ecs stuff compiles
* az-physics tests pass :tada:
* fix all the tests
* clippy on azalea-ecs-macros
* remove now-unnecessary trait_upcasting feature
* fix some clippy::pedantic warnings lol
* why did cargo fmt not remove the trailing spaces
* FIX ALL THE THINGS
* when i said 'all' i meant non-swarm bugs
* start adding task pool
* fix entity deduplication
* fix pathfinder not stopping
* fix some more random bugs
* fix panic that sometimes happens in swarms
* make pathfinder run in task
* fix some tests
* fix doctests and clippy
* deadlock
* fix systems running in wrong order
* fix non-swarm bots
|
|
* make azalea-pathfinder dir
* start writing d* lite impl
* more work on d* lite
* work more on implementing d* lite
* full d* lite impl
* updated edges
* add next() function
* add NoPathError
* why does dstar lite not work
* fix d* lite implementation
* make the test actually check the coords
* replace while loop with if statement
* fix clippy complaints
* make W only have to be PartialOrd
* fix PartialOrd issues
* implement mtd* lite
* add a test to mtd* lite
* remove normal d* lite
* make heuristic only take in one arg
* add `success` function
* Update README.md
* evil black magic to make .entity not need dimension
* start adding moves
* slightly improve the vec3/position situation
new macro that implements all the useful functions
* moves stuff
* make it compile
* update deps in az-pathfinder
* make it compile again
* more pathfinding stuff
* add Bot::look_at
* replace EntityMut and EntityRef with just Entity
* block pos pathfinding stuff
* rename movedirection to walkdirection
* execute path every tick
* advance path
* change az-pf version
* make azalea_client keep plugin state
* fix Plugins::get
* why does it think there is air
* start debugging incorrect air
* update some From methods to use rem_euclid
* start adding swarm
* fix deadlock
i still don't understand why it was happening but the solution was to keep the Client::player lock for shorter so it didn't overlap with the Client::dimension lock
* make lookat actually work probably
* fix going too fast
* Update main.rs
* make a thing immutable
* direction_looking_at
* fix rotations
* import swarm in an example
* fix stuff from merge
* remove azalea_pathfinder import
* delete azalea-pathfinder crate
already in azalea::pathfinder module
* swarms
* start working on shared dimensions
* Shared worlds work
* start adding Swarm::add_account
* add_account works
* change "client" to "bot" in some places
* Fix issues from merge
* Update world.rs
* add SwarmEvent::Disconnect(Account)
* almost add SwarmEvent::Chat and new plugin system
it panics rn
* make plugins have to provide the State associated type
* improve comments
* make fn build slightly cleaner
* fix SwarmEvent::Chat
* change a println in bot/main.rs
* Client::shutdown -> disconnect
* polish
fix clippy warnings + improve some docs a bit
* fix shared worlds*
*there's a bug that entities and bots will have their positions exaggerated because the relative movement packet is applied for every entity once per bot
* i am being trolled by rust
for some reason some stuff is really slow for literally no reason and it makes no sense i am going insane
* make world an RwLock again
* remove debug messages
* fix skipping event ticks
unfortunately now sending events is `.send().await?` instead of just `.send()`
* fix deadlock + warnings
* turns out my floor_mod impl was wrong
and i32::rem_euclid has the correct behavior LOL
* still errors with lots of bots
* make swarm iter & fix new chunks not loading
* improve docs
* start fixing tests
* fix all the tests
except the examples i don't know how to exclude them from the tests
* improve docs some more
|
|
|
|
Pathfinding is very much not done, but it works enough and I want to get this merged.
TODO: fast replanning, goals that aren't a single node, falling moves (it should be able to play the dropper), parkour moves
|
|
Arc<Event> -> Event, Arc<Mutex<State>> -> State
Items in State now need to have interior mutability (i.e. Arc<Mutex<T>>), but it's a worthwhile tradeoff since it allows the user to customize it for each field
|
|
|
