* Please describe the intent of your changes in a clear fashion.
This PR addresses the `Cannot read null.stat_panel` error occurring in
`SSstatpanels/fire()`.
**Root Cause:**
The primary cause was a race condition during client login. A `client`
object was added to `GLOB.clients` in `client/Login()` before its
`stat_panel` member had been fully initialized. This allowed the
`SSstatpanels/fire()` subsystem, which iterates over `GLOB.clients`, to
attempt to access `target.stat_panel.is_ready()` on a `null`
`stat_panel`, leading to a runtime error. Recent changes increasing the
frequency of `stat_panel` updates exacerbated this issue.
**Solution:**
1. **Reordered Client Initialization:** In
`code/modules/client/client_procs.dm`, the line `GLOB.clients += src`
has been moved to occur *after* `stat_panel = new(src, "statbrowser")`
and `stat_panel.subscribe(...)`. This ensures that a client is only
added to the global list once its `stat_panel` is properly instantiated,
eliminating the race condition.
2. **Defensive Null Checks:** Additional null checks for `stat_panel`
have been added in `code/controllers/subsystems/statpanel.dm`:
* In `fire()`, the condition `!target.stat_panel.is_ready()` was updated
to `!target.stat_panel || !target.stat_panel.is_ready()`.
* In `refresh_client_obj_view()`, an early return
`if(!refresh.stat_panel) return` was added.
These defensive checks provide robustness against any future reordering
issues or unexpected scenarios where `stat_panel` might be null.
* Please make sure that, in the case of mapping changes, you include
images of these changes in the PR's description.
* Please make sure to mark your PR as wip or review required by making a
comment with !wip or !review required
* If you include sprites/sounds/... (assets) that you have not created
yourself specify the license and original author below.
* Ensure that you also credit them in the appropriate location /
changelog as specified in the contributor guidelines
### Asset Licenses
The following assets that **have not** been created by myself are
included in this PR:
| Path | Original Author | License |
| --- | --- | --- |
| icons/example.dmi | ExamplePerson (Example Station) | CC0 |
Fixes
[SERVER-PROD-1M4](https://aurorastation.sentry.io/issues/7547575486/?seerDrawer=true)
---------
Co-authored-by: VMSolidus <evilexecutive@gmail.com>
## About PR
Makes it so when a turf calls the overlay subsystem (if overlay layer
was left as default) it'll add overlays slightly above the turf layer.
With this we'll no longer have the silly issue where turf smoothing
overlays appear under the adjacent turfs depending on turf load order
In short, here's what it fixes
Before:
<img width="396" height="601" alt="Screenshot_10"
src="https://github.com/user-attachments/assets/91708402-c6c6-433f-af9f-5293ce926731"
/>
After:
<img width="391" height="598" alt="Screenshot_9"
src="https://github.com/user-attachments/assets/556d48dd-d463-4f76-b075-f0d77340e837"
/>
---
to-do:
- [x] Figure out whether to change `is_plating()` check method or add
the missing `initial_flooring` for every exoplanet turf type since this
plays a role in deciding decal layer, suffer as both paths are equally
agonizing
The statpanel (alt-click pickup menu) now updates whenever an item in it
is moved.
The statpanel (alt-click pickup menu) previously only updates on the
subsystem firing, every 2 seconds.
This lets you interact with an object after it's already in the contents
of another object.
Eg. Pick up a bullet, load it into a magazine, pick up that same bullet
from the statpanel, load it again.
This is the source of nearly every inventory bug I've encountered.
(vanishing armor attachments, bullet pile duplication, double-click
triggering grenades, modlaser components being in two guns at the same
time).
It also just feels bad and makes the menu much less usable. (Have I
picked up that item? Which one of these 6 identical icons did I click on
last? Better wait 2 seconds for it to update).
This does NOT cause the panel to update when a new item is added, as
that causes an Icon/Image refresh, which is the expensive part of the
proc. (At least I hope it doesn't)
https://github.com/user-attachments/assets/f396d4bb-d9be-407a-ad66-0cbf8cd4490b
With massive thanks to @JohnWildkins for help in touching a lot of
concepts I've never worked with before, as well as helping me get back
into contributing in general.
This PR allows crew transfers to be called during red alert and adds
support for weighted voting to help favour active participants in these
particular rounds.
Observers and players still in the lobby are given only half a vote
towards transfer/continue. There is also config support to define an
active 'participation time' in minutes, which is used to stop potential
abuse cases of players respawning/joining from the lobby just before (or
even during) the transfer to gain a full vote. By default, this is set
to 15 minutes, so anyone who's been actively playing in the round for at
least that long is eligible.
This is my first time touching things like the config file or any sort
of subsystem more advanced than, like, the far more basic PRs I'm used
to, but I was encouraged to look at this as a little project. It all
appears to work well locally.
---------
Signed-off-by: Matt Atlas <mattiathebest2000@hotmail.it>
Co-authored-by: Jaraci <mirandatrasen@nt.net>
Co-authored-by: Matt Atlas <mattiathebest2000@hotmail.it>
Co-authored-by: hazelrat <83198434+hazelrat@users.noreply.github.com>
Y'know, probably. It certainly seems to work locally!
changes:
- code_imp: "Exoplanets now cache their main outdoor atmospheric region
instead of trying to search their entire map once per second."
- code_imp: "Exoplanet atmospheric calculations now run only when
necessary using new update_revision counter."
- code_imp: "Exoplanet atmospheres now wait until all random ruins have
been placed before trying to resolve the air everywhere."
- admin: "Adds new admin debug verb, 'Toggle Non-Horizon Temperature
Graphics', which toggles all off-ship gas temp graphics to address lag
in a pinch, if still necessary, without removing actual environmental
hazard."
Fixes https://github.com/Aurorastation/Aurora.3/issues/21944
Spiritual successor of
https://github.com/Aurorastation/Aurora.3/pull/22352
Cherrypicked changes from the C3 Crew Management branch, this PR
contains all NTNet extensibility features, including Field Relay
construction, off-site camera usage (contingent on standing up remote
hardware and being out of jamming), etc. This also fixes a handful of
outstanding telecomms and view/mobeye-related bugs adjacent to where I
was working.
changes:
- rscadd: "Adds NTNet extensibility, requires new machinery constructed
on away-site z-levels."
- rscadd: "Breaks NTNet Relay machinery into Quantum Relay ('core' obj
on Horizon) and Field Relay (deployable off-site constructions)."
- qol: "Embedded cameras (basically every video camera EXCEPT those
which are installed on walls) can now be viewed from NTNet-connected
z-level, not just connected ones."
- rscadd: "Adds several public camera monitor consoles to the Horizon
(near where public sensor grid feed consoles already exist)."
- rscadd: "Adds some spare camera console boards to Horizon technical
storage."
- qol: "Adds action button for 'cancel-camera-view' verb while viewing
cameras or looking up/down z-level."
- bugfix: "Fixes Jockey comms (frequency was missing from ANTAG_FREQS)."
- bugfix: "Fixes look-up/look-down actions intermittently failing
without first using cancel-camera-view."
* Readying up and successfully joining the round now gives you a morale
bonus that lasts for 30 minutes.
AI usage disclosure: The code for this was created in-part using GPT 5.6
Sol.
Complete overhaul of Integrated Circuits Final Edition!
- Adds phoron cost to circuits (not assemblies).
- Printer searching capability.
- Changed names of assemblies to better match their description.
- Adds cloning and auto appending JSON for large builds.
- Adds database, debug, advanced math circuits.
- Reworks the list circuits to finally accept anything other than NUM.
- Adds plenty of list circuits to allow functional building and clearing
of lists.
- Reworks Complexity/Space to allow a bit better flexibility in creation
of devices.
- Adds a "Headroom" concept where space/complexity can go beyond the
maximum to a very slight degree. Concretely, unused space gives `2`
extra complexity per unused component slot, capped at `15%` of
`max_complexity`. Unused complexity gives `1` extra component slot per
`6` unused complexity, capped at `15%` of `max_components`.
- Groups math and logic circuits and adds subgroups.
- Doesn't mess with the circuit tool kit.
- Adds a xenoarch drill assembly with a nice sprite. Thank you
@EnchantedCrocolisk for the sprite!
- Changes activation limit time to 1 second from 2 seconds, adds a 1
second timer that is unique to the INTEGRATED CIRCUIT PRINTER with the
advanced design disk.
- Ports several Baystation integrated-circuit manipulation components,
primarily from
Baystation12/Baystation12https://github.com/Baystation12/Baystation12/pull/22458,
with anchoring bolts and hatch lock sourced from
Baystation12/Baystation12https://github.com/Baystation12/Baystation12/pull/25052.
- Changes the electronic ear piece into a workable radio headset with
working encryption keys capability and unique ways the Microphone and
TTS interacts with it.
- Removed the ability for all the electronic wearables to spawn with a
basic cell.
- Combines https://github.com/Aurorastation/Aurora.3/pull/22466,
https://github.com/Aurorastation/Aurora.3/pull/22456,
https://github.com/Aurorastation/Aurora.3/pull/22455,
https://github.com/Aurorastation/Aurora.3/pull/22446 that are now
closed.
# Summary
This PR is the next update to the persistency subsystem. The goal of
this PR is to provide more framework like functions to allow more types
of content to be made persistent.
Currently only volatile game objects created during a round can be (*in
a [clean](https://www.youtube.com/watch?v=rZ3ETK7-ZM8) way*) saved and
made persistent. This update attempts to provide methods to make
*everything*¹ persistent.
This introduces persistent generics and history.
## Database
The following things are going to be changed and added in the database
(open in new tab for better visibility, PNG file includes the drawIO
code):
<img width="1692" height="1041" alt="aurora_persistency_db drawio"
src="https://github.com/user-attachments/assets/ea53f419-f9aa-4592-af8f-3a8d5edf3177"
/>
**Deviations on database implementation from diagram:**
- Removed unique constraint on history table - Prevented adding multiple
records per round per attribute.
## Framework surface changes
- MC/VV: Moved global object track register to subsystem var space.
- MC/VV: Point of interest: Added history_cache and generic_cache to
subsystem var space.
- MC/VV: Updated subsystem stat entry message, now providing information
on cache sizes of new types.
- Added `singleton/persistent_type` defines (Type, clean-up rules,
finalization hook) and macros allowing new definitions of said types.
- Added cache structures that are also used for returns on public procs
in generics and history persistent types.
- Major new framework features: Persistent history (example: Mining
yield records) and persistent generics (example: Persistent Horizon
overmap position). See documentation for more information.
DrawIO diagram for documentation, includes source in it (open in new
tab):
<img width="200" height="200" alt="Persistence-subsystem-flowchart
drawio"
src="https://github.com/user-attachments/assets/51f28331-f999-49a2-a7cc-58278f7ae416"
/>
## Tasks
(These lists are not comprehensive.)
**General**
- [x] Update DB - Write SQL scripts.
- [x] Add subsystem modular files for generics and history.
- [x] Add type definition logic, macros.
- [x] Add type-DB init logic.
- [x] Logging.
- [x] A lot of testing. *A lot.*
- [x] Changelog.
- [x] Self-Review.
- [x] Update documentation on the persistence subsystem.
**"Persistent history"**
- [x] Add init logic.
- [x] Add finalize logic.
- [x] Add framework surface procs.
- [x] Get last record.
- [x] Get last X records.
- [x] Add record.
- [x] Add character ID related validation.
- [x] Add initial example mechanic.
**"Persistent generics"**
- [x] Add init logic.
- [x] Add finalize logic.
- [x] Add framework surface procs.
- [x] Save.
- [x] Load.
- [x] Add initial example mechanic.
## Changes
Too many changes to be listed here - Check changelog and actual changes.
## Warning
There are certain use/test cases that *cannot* be tested locally due to
missing preexisting data in the database. This should only affect new
data structures (new persistent types), not existing data.
¹ _Large scale persistent mapping is excluded for this version._
---------
Co-authored-by: Copilot <copilot@github.com>
These things are ultra shitcode and need to be remade from the ground
up, but for the time being this stops them from exploding the server.
Co-authored-by: Matt Atlas <liermattia@gmail.com>
Cargo orders would have trouble clearing the list of orders shown
without reloading the ui, due to a lack of unique ID if there was
several of the same item. The same applied to the cooking codex. This is
fixed now.
This also fixes the cooking codex search.
Fixes several runtimes from the past few weeks, some additional material
repath regressions, and the fact that if I shoot you with a stream of
ionized deuterium particles, it doesn't riddle you with cancer.
changes:
- bugfix: "Fixes emergency shields runtiming when hit by thrown
objects."
- bugfix: "Fixes several stale global list runtimes."
- bugfix: "Fixes several health percentage checks that could use unset
initial health values."
- bugfix: "Fixes additional materials regressions in INDRA code."
- bugfix: "Fixes invalid fuel injector depletion math."
- bugfix: "Makes accelerated particles apply radiation damage to living
mobs they pass through."
Soft-ish port of https://github.com/NebulaSS13/Nebula/pull/540. Except
we call them singletons.
Repaths all materials as singletons instead of datums, and replaces
material defines from strings to paths so that we can just run
GET_SINGLETON instead of needing to use SSMaterials. This is Step One.
This PR has no player-facing changes.
changes:
- refactor: "Repaths /material to /singleton/material."
- refactor: "Replaces all material string defines to path defines,
replacing SSmaterials procs w/ GET_SINGLETON instead."
- refactor: "Removes all material var edited objects from all maps,
adding new presets where necessary."
- refactor: "Updates recipes unit test to run all recipes against all
material singletons."
---------
Signed-off-by: Batrachophreno <Batrochophreno@gmail.com>
Co-authored-by: kano-dot <bhutanlikanoxy@gmail.com>
Life() was wasting a bunch of processing time on "Every hardsuit
forever" regardless of if it was equipped or not. Additionally Life()
was responsible for double-processing hardsuits, which is extremely
improper. Since hardsuits are being moved to SSprocessing instead of
SSmobs, which has a different tickrate, I went ahead and tick
differentiated them to make sure that nothing will break.
# Summary
This PR adds a new grim spatial sound track to the lobby track list +
lemurian sea sector and adds an ambience loop tracks to vents+scrubbers
as well as SMES's.
## Asset Licenses
The following assets that **have not** been created by myself are
included in this PR:
| Path | Original Author | License |
| --- | --- | --- |
| sound/music/lobby/spatial_audio.ogg | [spatial audio 2 by
BadWolf23/freesound.org](https://freesound.org/people/BadWolf23/sounds/723957/)
| CC BY 4.0 |
| sound/machines/ventilation_humming.ogg | [humming machinery by
Lewooz/freesound.org](https://freesound.org/people/Lewooz/sounds/545952/)
| CC0 1.0 |
| sound/machines/electrical_humming.ogg |
[SFX_GEOFON-HUM-ELECTRICAL-BOARD_4824_01 by
pblzr/freesound.org](https://freesound.org/people/pblzr/sounds/785942/)
| CC0 1.0 |
**Changes made to original sounds:**
- Added fadein/out to spatial_audio, reduced volume, corrected file
metadata (author, trackname, comment).
- Cut background noise in ventilation_humming, shortened soundfile,
corrected file metadata (author, trackname, comment).
- Shortened electrical_humming soundfile, corrected file metadata
(author, trackname, comment).
Unused vars take up memory and worsen proc overhead. By far the worst
offenders to this are vars on /atom and /mob since they more or less
globally stack lots and lots and lots and lots of proc overhead. So I'm
cleaning up unused vars on these to save resources. I also noted several
vars that are excellent candidates for making components or elements.
When I'm not drowning in school work anymore, their time is numbered.
Mfw every space carp forever has a CCIA record, no wonder mobs are
laggy.
I've had extended logging for Incomplete Destroys for about 2 weeks now,
so there's the promised PR that fixes each and every single poisoned
destroy that was recorded in the past 2 weeks by the sentry logs.
Mind blankers and psi-protect pills now display messages roughly every 2
minutes and 1 min respectively.
Mind blankers have a random chance for the message to also display on
the screen, as they're more invasive, being implants, but also safer.
A warning timer was also added for seizures from psi-protect, so people
have a short moment to react in.
Thanks to more extensive data from Sentry logging, I've discovered that
there's actually two distinct causes for Hard Dels.
1. **Unaccounted For References**: Any potential references that are not
nulled out during a Destroy() proc will eventually trigger a hard
delete. These are the classic "Memory Leaks" and were previously the
most common form of hard del.
2. **Poisoned Destroy()**: Some destroy procs are prone to runtime
errors. When a runtime error occurs, the destroy() proc returns .
instead of ..(), meaning all parents of an object will skip any
reference cleanup they would have wanted to do. Since there's ref
cleanup performed at every single step of a parent hierarchy all the way
down to /datum itself, these failed Destroy() procs are nearly
guaranteed to trigger a hard delete.
For example, nearly all current /human related hard deletes are caused
by this specific runtime error:
https://github.com/Aurorastation/Aurora.3/pull/22680
So this PR adds a try catch to the garbage collector to log runtime
errors generated by Destroy() procs, allowing me to far more easily
track down this specific variant of harddel.
---------
Signed-off-by: VMSolidus <evilexecutive@gmail.com>
I accidentally flipped a sign on the guard clause, making it so that
APDI was only sending messages to corpses.
---------
Signed-off-by: VMSolidus <evilexecutive@gmail.com>
ADPI wasn't obeying the standard rules for psionics, and was incorrectly
bypassing arbitrary sources of psionic protection. There was a lot of
unecessary hardcoding in the checks as well that meant it couldn't
account for edge cases. Luckily I made helper procs for handling every
single one of these edge cases automatically about 6 months ago.
changes:
- refactor: "Refactors Gravity Generator UI from HTML to TGUI."
- rscadd: "Adds sector-based modifier var 'overmap_hazards_multiplier'
for generating total overmap hazard count."
- soundadd: "Adds new Lemurian Sea 'welcome' announcements."
- rscadd: "Added Lemurian Sea gravitic anomalies as overmap hazards and
random events."
- rscadd: "Adds ADPI/The Voices."
- rscadd: "Adds ADPI Panel for modmins to send ADPI messages to a single
character or all characters from their respective pools, or custom
messages."
- balance: "Vaurca Hivenet communication, the Hivenet Manifest, and
Hivenet Echoes are now disabled while the ship is in Lemurian Sea
sectors."
Part 3 of the Mob Destroy Refactor, this time going through the entire
list of every ref stored up to a mob/human, and (attempting) to verify
and cleanup every possible circular ref. This PR also fixes some
mistakes made with cleaning up UI elements, namely that tgui's really
don't like it when you qdel them, and screen objects also weren't always
clearing their own references if Qdel'ed directly. There were several
niche situations too where circle refs might be retained by a mob. I
also found an issue where static lights were only cleaned up on
/atom/movable/ but actually existed farther up the chain on /atom. I
don't know if any /atoms that aren't /atom/movable ever get static
lights, but the fact that they can be needs to be correctly accounted
for.
I can't possibly have gotten all of them, but this is every single one I
could find after 3 hours of work.
Rewrites Saycode and Langchat to add support for multiple languages in
one message, including audible emotes.
<img width="1139" height="338" alt="image"
src="https://github.com/user-attachments/assets/25e26932-7a6e-4c54-ab74-56fffb92ecad"
/>
Here's some samples to explain how to mix languages.
<img width="422" height="26" alt="image"
src="https://github.com/user-attachments/assets/e1b176cc-8625-4dc9-83c8-a053d3f310e6"
/>
`Languages ,2 can be mixed ,3 like this.`
<img width="540" height="21" alt="image"
src="https://github.com/user-attachments/assets/19156c67-4670-4d7a-84d7-26e527de2676"
/>
`!explains, ,2Emotes work too. ,0The text will get auto-quoted.`
<img width="592" height="18" alt="image"
src="https://github.com/user-attachments/assets/cfc31c5c-2383-41c8-82be-b36836339321"
/>
`,3Most languages ,0can be ,2mixed ,1arbitrarily, ,3any number of
,0times.`
<img width="636" height="20" alt="image"
src="https://github.com/user-attachments/assets/388b4f9d-192a-4374-ac31-bbd4e4e5dfe4"
/>
`,2Emotes. ,eAnd he nods. ,3They don't have to come first anymore.`
There are some exceptions. The exceptions are currently anything with
any of the flags `SIGNLANG`, `HIVEMIND`, `PRESSUREPROOF`,
`KNOWONLYHEAR`. Exceptions work the same as current languages do: they
must be the first language in the message. If so, they prevent switching
into any other language mid-message; if they're not first, they just
wont trigger.
They're exceptions currently because there's not really a clean way that
I or the people I asked for help on this one to make them look nice.
`SIGNLANG` for example doesn't scramble text, it shows it's own `
gestures a lengthy message.` text for those that don't understand. Could
we just replace every instance of sign language with that if somebody
doesn't understand? Probably. It would look pretty awful though. e.g.
`Alina Eskelinen says, "Hello." Alina Eskelinen gestures a short
message.`
This definitely needs testmerging because it more or less rewrites the
entire pipeline surrounding `say`. The `say` code itself had to be
rewritten to support the multiple languages, as well as all the existing
plumbing for listeners receiving messages. In return though, it's
significantly more straightforward and hopefully by extension easier for
people to add to in the future.
Primarily, instead of having four different `hear_say`, `hear_radio`,
`hear_sleep`, etc., routes for messages to come through, every single
audible message is received by `hear_message`, which is responsible for
figuring out how clear the message is (is the radio damaged? is it a
whisper we're eavesdropping on?), who needs to receive it in their
chatbox, formatting it correctly for each listener, and finally if any
npc or object within range needs to react to it in some way, like a
parrot or a mech.
changelog:
- rscadd: "Adds code-switching: you can now speak in multiple languages
in the same message."
- rscadd: "Adds audible emotes to the language list. They can be
triggered with ,e."
- rscdel: "Removes SSrunechat."
- refactor: "Rewrote langchat in order to support multiple languages and
partial comprehension."
- refactor: "Rewrote a vast majority of all saycode and the code
responsible for displaying saytext to clients."
- bugfix: "Sleeping mobs are no longer able to understand all
languages."
- bugfix: "Langchat now correctly shows the appropriate comprehension
for all viewers rather than all viewers sharing the comprehension of the
last viewer."
- bugfix: "Languages which are supposed to be invisible when not
understood no longer appear as scrambled overhead text."
Forgive me whoever has to review this. Biggest areas that have room for
error is stuff like a borer inside someone's head, and Dionae stuff. Old
langchat had odd exceptions for those and I was forced to rewrite it
entirely, but I think I got it all back to how it was working before.
changes:
- bugfix: "Restores the Modular Computer access cable header button for
connecting to IPCs and machinery."
- bugfix: "Fixes display issues in the Robotics interface's
<collapsible> elements."
- bugfix: "Fixes in-chat feedback text provided by the Cyborg Analyzer
(organ name and status now both inline)."
- bugfix: "Crew Manifest PDA app now filters out off-ship personnel."
- bugfix: "Fixes Autolathe TGUI appending search- and category- filtered
recipes instead of recomputing the list with those filters."
- bugfix: "Fixes runtime when ghost opens dedicated camera console UI."
- code_imp: "Update some defs referencing 'supply' department to
'operations'."
- refactor: "Standardizes all interface search bars to use the native
'SearchBar' component."
- refactor: "AtmosAlarmControl interface now correctly auto-sorts alarms
by alert status."
changes:
- bugfix: "Fixes LateJoin and Manifest interfaces not displaying
department colors by creating a shared component for both interfaces to
use."
- bugfix: "Fixes the Manifest interface not displaying colors for crew
status."
- bugfix: "Fixes status composer not allowing AIs to view TGUIs."
- bugfix: "Fixes inline icons in alt-click atom viewer."
- bugfix: "Fixes PDA light toggle icon not appearing."
- bugfix: "Fixes autolathe category view not filtering items."
---------
Signed-off-by: Batrachophreno <Batrochophreno@gmail.com>
Re-creation of https://github.com/Aurorastation/Aurora.3/pull/21046 to
skip merge conflict hell. Brings us modern TGUI.
**ALTERNATE TITLE: TGUI HELLSCAPE PR
ABANDON ALL HOPE YE WHO ENTER HERE**
- [x] Migrate build tools (javascript -> typescript, bun for package
management).
- [x] Upgrade all TGUI dependencies and associated root files to
TG-congruent versions (axios, babel, dompurify, eslint, highlight,
marked, prettier, sass, source-map, stacktrace-parser, typescript).
- [x] InfernoJS -> React migrations
- [x] React cleanup and polish (migrate all remaining .js files to
appropriate .ts or .tsx filetype, all remaining hooks, linting, error
corrections, etc.)
- [ ] Test all remaining TGUI interfaces
Makes ship combat significantly more destructive (and a bit deadlier),
especially with non-explosive projectiles.
**Additions:**
- A shell that penetrates a solid (dense) structure will cause spalling.
This spall does 2-30 damage, depending on range and has a high embed
chance. These embedded chunks are large enough to be ripped out by hand.
(But you probably shouldn't). The pilot's suit has 30 ballistic armor,
so a prepared bridge crew can survive a lot of spall.
- Adds a gun that shoots any ship weapon that admins can spawn. Useful
for testing, or for adminbus where they don't want to have to spawn a
whole overmap ship.
- Made several structures destructible, vending machines, computer
frames, grilles (these were just bugged) and watertanks (and their
children). These structures could just block infinite shots and were
common enough to serve as effective armor for the horizon. These now
spawn shrapnel (of their material if it's set), in addition to sheets of
steel when destroyed.
- Added support for negative maim chance. This allows chosen projectiles
to be prevented from decapitating. No more accidentally gibbing the
bridge crew's heads. (Unless they get hit with something really really
big).
**Balance:**
- All non-explosive anti ship rounds got significant buffs. Typically
2-3x Anti-Material numbers.
- Armour piercing anti-ship rounds got massively increased penetration
stats. Typically 4-8 from 1-2.
- Doors now get destroyed in less hits, if they are hit by anti-material
projectiles. Hits to destroy -= Antimaterial / 2
- Francisca frag shells now shoot less fragments but deal more damage.
394 projectiles for every shot was a bit laggy, now it's only 280.
**Shields:**
Shields now take more power. 13-14 megawatts for a 10 strength shield.
Just inside the ship's generation capacity with both the INDRA and
Supermatter running.
Instead of deleting any projectile they are hit by, shields now have
different failure modes.
- Firstly, a piercing projectile has it's penetration reduced by the
shield strength. Eg A shot that would go through 10 walls, only
penetrates 4 after going through a 6 strength shield.
- Secondly, a projectile has it's damage multiplied by 1 - shield
strength / 10, then has shield strength subtracted from it's damage. Eg.
A damage 100 projectile, hitting a 5 strength shield, has it's damaged
reduced to 50, (100 * 0.5), then to 45. (50-5)
- Thirdly, an explosive projectile (that does not penetrate per the
first step) does not explode, but instead deals it's full damage to the
shield.
Hits to the shield now also damage the field itself, not just the shield
tile. As such, a very big hit on the shield can bring the whole thing
down, requiring an engineer to reset it.
Shields can now be upgraded with research components. Additionally,
shield generators can now support more than one capacitor. The
roundstart capacitor on the Horizon is only good enough to support a 6
strength shield.
The upgrades increase efficiency, strength loss over time and charge
speed.
**Overmap Targetting**
Previously, shells would spawn within 20 tiles of their target and aim
directly forwards. This was bugged for shots with burst, which would
instead all aim inwards and mostly miss. This is the reason the Grauwolf
(in addition to the explosion bug below) and Fransisca felt so useless,
only one of every shell in a burst would hit.
Shots now spawn at the map edge, based on the orientation of the
projectile and ship. Diagonals are now accounted for.
Shots are then aimed directly at the target selected on the console,
with 3 degrees of dispersion.
**Bugfixes:**
- Fixes lots of explosions that didn't pass a turf to the explosions
subsystem. This caused the explosion not to happen if the thing hit got
qdeled before the subsystem could get the turf to spawn the explosion
on.
- Embedded shrapnel no longer gets the initial projectile name, this
prevents it getting named shrapnelshrapnel.
- Fixes a bunch of projectile piercing bugs, mostly with grilles and
windows. Shots now go past window frames without hitting them.
- Fixes shields not having a click delay when attacked.
---------
Signed-off-by: FenodyreeAv <fenodyree.av@gmail.com>
Co-authored-by: Matt Atlas <mattiathebest2000@hotmail.it>
Refactoring the entire destroy proc path from Mob Human all the way down
to Atom while trying to find the causes for the damn mob human hard
deletes. This PR comprehensively reorganizes every single stray
snowflake var used by /atom/ all the way to /mob/living/carbon/human,
and makes sure that every var that COULD store a reference, is now
cleared during the entirety of the Mob Destroy() parent hierarchy.
This may very well be the end of the lag war.
In total, I've hunted down and cleared 39 hanging references between
/atom and /mob/living/carbon/human
---------
Signed-off-by: VMSolidus <evilexecutive@gmail.com>
The typical hard del costs 300ms of time, but Hard Dels only get
reported to the Sentry logs if they exceed 500ms, which only an
extremely tiny minority of them ever get past. In actual practice we
only need to ignore the hard dels that are under 200ms because only the
two "Roundstart subsystem hard dels" will ever be below that number and
can be ignored. ALL other hard dels should be reported with the details
required to fix them, especially so that I can have actually usable
information about what is causing them and what I need to fix.
* Please describe the intent of your changes in a clear fashion.
This PR addresses SERVER-PROD-75, where the `set_turf_examine_tab` and
`return_object_images` procs in the statpanels subsystem would crash due
to attempting to access `client.obj_window.atoms_to_show` when
`client.obj_window` was null.
The root cause was a lifecycle desync: `client.obj_window` is primarily
initialized within `/mob/set_listed_turf` and can be explicitly nulled
by `/datum/object_window_info/Destroy()`. However, the statpanels
subsystem would still attempt to update the turf examine tab if
`mob.listed_turf` was set, even if `client.obj_window` had become null
(e.g., after a mob transfer, relogin, or `obj_window` destruction
without a subsequent `set_listed_turf` call).
The fix involves lazy-initializing `client.obj_window` within both
`set_turf_examine_tab` and `return_object_images`. This ensures that
`client.obj_window` is always a valid `/datum/object_window_info`
instance before its properties are accessed, preventing the null
dereference crash.
* Please make sure that, in the case of mapping changes, you include
images of these changes in the PR's description.
* Please make sure to mark your PR as wip or review required by making a
comment with !wip or !review required
* If you include sprites/sounds/... (assets) that you have not created
yourself specify the license and original author below.
* Ensure that you also credit them in the appropriate location /
changelog as specified in the contributor guidelines
### Asset Licenses
The following assets that **have not** been created by myself are
included in this PR:
| Path | Original Author | License |
| --- | --- | --- |
| icons/example.dmi | ExamplePerson (Example Station) | CC0 |
Fixes SERVER-PROD-75
---------
Co-authored-by: sentry[bot] <39604003+sentry[bot]@users.noreply.github.com>
Co-authored-by: VMSolidus <evilexecutive@gmail.com>
This PR fixes a bunch of skills related bugs, the biggest of which were
the result of the system being overly trusting of the database, when in
reality due to a bunch of unpredictable edge cases, the database is not
guaranteed to always have what I think it has. To fix these bugs, I've
had to slightly refactor how skills are generated on player characters
and antagonists, such that the burden of proof for skills is with the
Skills Subsystem rather than the Database.
Skills generated for a fresh character that has NO preferences saved
(Worst case scenario):
<img width="1491" height="849" alt="image"
src="https://github.com/user-attachments/assets/683fb538-106a-4679-8e2a-30dd45f456a1"
/>
Promoting that same character to Antagonist now increases certain skills
to a minimum baseline:
<img width="1315" height="830" alt="image"
src="https://github.com/user-attachments/assets/94bcc69e-cbbe-45fa-956e-f53b7f5d2779"
/>
By Mel's request, Bluespace Technicians spawn with all skills fully
maxed out for debugging purposes:
<img width="1909" height="985" alt="image"
src="https://github.com/user-attachments/assets/db4f3f65-3d87-47fc-907e-c68eae6d4cdb"
/>
fixes https://github.com/Aurorastation/Aurora.3/issues/22409
The reason for the increased elevator timer was due to how rarely it
came up that the supplier time would be longer than the base 30 second
timer, so instead the two timers were combined to trial how a longer
timer would feel.
This is a fix for https://aurorastation.sentry.io/issues/7405385361
Which I believe might actually also be one of the elusive "niche human
hard dels" that are so hard to track down since this runtime would also
prevent a key memory deallocation from occurring.
