## About The Pull Request
As the title says. This should significantly improve the performance of
running lua scripts.
## Why It's Good For The Game
There is a lot of performance overhead in logging each individual
function called by a lua script as can be seen in the following
screenshot:
This is a test done on a local server where I run the
`zombie_controller.lua` script that can be found on the auxlua-cookbook
repository with a lot of AI zombies spawned in.
Logging these calls is not necessary as it doesn't provide any real
information to anyone looking for bad actors. Lua scripts are already
logged when ran and they can be examined to spot if the script being run
is done so in bad faith.
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
Cleaned up the SS13.register_signal and SS13.unregister_signal, removing
the weird list shifting.
Also adds a new lua file that can be included for the use of registering
different signals on various datums and being able to clear them all in
1 function.
Removed the make_easy_clear_function option when registering a signal
via lua because I don't think it's used by anyone and it lacks any sort
of versatility. Users can just create their own function for clearing
signals from a datum.
Also updates the documentation for HARDDELETES.md as
COMSIG_PARENT_QDELETING was renamed to COMSIG_QDELETING
## Why It's Good For The Game
New handler file makes registering signals in batches a lot easier if
you want to clear them in one go without clearing unrelated callbacks on
the same datum. The list shifting in SS13.register_signal had pretty
significant performance problems, so removing that will make registering
and unregistering signals faster.
## Changelog
🆑
admin: LUA - Adds a new library called handler_group. Include it in your
files by doing require('handler_group')
/🆑
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
My lua scripts were hitting the topic byte limit, so this makes file
upload of lua scripts able to bypass the topic limit.
## Why It's Good For The Game
Removes arbitrary restriction on how big a lua file can be in bytes.
## Changelog
🆑
admin: Admins can now run lua files bigger than 36 KB by importing them
directly.
/🆑
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
<!-- Write **BELOW** The Headers and **ABOVE** The comments else it may
not be viewable. -->
<!-- You can view Contributing.MD for a detailed description of the pull
request process. -->
## About The Pull Request
<!-- Describe The Pull Request. Please be sure every change is
documented or this can delay review and even discourage maintainers from
merging your PR! -->
Lua errors don't get logged when `call_function` is called
Timer.start_loop was just straight up broken due to me not properly
testing it, so this fixes that.
## Why It's Good For The Game
<!-- Argue for the merits of your changes and how they benefit the game,
especially if they are controversial and/or far reaching. If you can't
actually explain WHY what you are doing will improve the game, then it
probably isn't good for the game in the first place. -->
Makes debugging lua scripts easier. Also fixes bugs.
## Changelog
<!-- If your PR modifies aspects of the game that can be concretely
observed by players or admins you should add a changelog. If your change
does NOT meet this description, remove this section. Be sure to properly
mark your PRs to prevent unnecessary GBP loss. You can read up on GBP
and it's effects on PRs in the tgstation guides for contributors. Please
note that maintainers freely reserve the right to remove and add tags
should they deem it appropriate. You can attempt to finagle the system
all you want, but it's best to shoot for clear communication right off
the bat. -->
🆑
fix: Fixed lua error logging.
fix: Fixed the SS13.start_loop function not working properly.
/🆑
<!-- Both 🆑's are required for the changelog to work! You can put
your name to the right of the first 🆑 if you want to overwrite your
GitHub username as author ingame. -->
<!-- You can use multiple of the same prefix (they're only used for the
icon ingame) and delete the unneeded ones. Despite some of the tags,
changelogs should generally represent how a player might be affected by
the changes rather than a summary of the PR's contents. -->
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
Lua uses the timer subsystem, which can end up holding the entire timer
subsystem if lua is creating a lot of timers.
The solution to this is to use an internal scheduler instead so that the
load gets placed onto the lua subsystem.
## Why It's Good For The Game
Performance improvement when using lua scripts.
## Changelog
🆑
refactor: Auxlua no longer uses the timer subsystem to get stuff done,
lua scripts shouldn't slow down timers anymore.
/🆑
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
After https://github.com/tgstation/tgstation/pull/80628, these shouldn't
be needed anymore right?
## Why It's Good For The Game
Cleans up some vestigial code
## Changelog
EDIT: Not player-facing.
## About The Pull Request
`SS13.get_runner_client()` and `SS13.get_runner_ckey` will return the
client and the ckey respectively of the user who ran the lua script. Can
be unreliable after the first sleep or yield.
The SS13 module can now be made local as the tables that need to be
accessed globally have been moved to their own global variables.
Added `SS13.start_loop(time, amount, func)`, `SS13.end_loop(id)`,
`SS13.stop_all_loops()` that allow lua scripts to more easily make
loops. Removed the `timer` parameter from these functions, which
specified the timer subsystem to use.
Documentation on all new added functions have been added in the
lua/README.md
## Why It's Good For The Game
Getting the client who ran the script and the ckey that ran the script
is useful for self contained scripts that are looking for an entrypoint
(e.g. location to spawn some item/mob/structure). `dm.usr` is a special
variable used for other purposes and can be unreliable when used this
way even if there haven't been any sleeps or yields yet, so having a
dedicated variable and function to handle it makes things easier.
Being able to make the SS13 module local allows for more self-contained
scripts. Although this isn't super helpful because `require` will still
load the same object for all lua scripts loaded on the same state.
Basic looping helpers allow for lua scripts to more easily create loops
without needing to recursively do a `set_timeout` or to do a
`while(true) do SS13.wait(1) end` loop.
## Changelog
🆑
admin: Added SS13.get_runner_ckey() and SS13.get_runner_client() which
stores the ckey and returns the client of the user who ran the lua
script. Can be unreliable if accessed after sleeping.
admin: Added timer loop helpers to the SS13.lua module, check the docs
admin: The SS13.lua module can now be made local without causing any
errors.
/🆑
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
Fixes some more lua harddel problems with lingering refs on the gc_guard
variable. This variable has been changed to a list instead and will get
cleared every time the SSlua subsystem fires so that lua instantiated
objects that are not tracked by the subsystem will essentially delete
themselves on the next tick, aka whenever the lua script sleeps.
Also removed the unnecessary and not completely functional
lua_reference_cleanup proc which wasn't even being called reliably
because the signal handler was the datum itself.
## Why It's Good For The Game
Fixes more harddel bugs, increases consistency.
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
See title. Callbacks were getting deleted in SS13.lua which is
non-ideal. They should be automatically collected by the garbage
collector.
Also adds some utility lua functions like SS13.is_valid,
SS13.stop_tracking, SS13.new_untracked, SS13.type and SS13.qdel
## Why It's Good For The Game
Fixes some harddel issues, as well as callbacks clogging up the garbage
subsystem.
## Changelog
🆑
admin: Added new lua functions: SS13.is_valid, SS13.stop_tracking,
SS13.new_untracked, SS13.type and SS13.qdel
/🆑
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
Lua created atoms hard delete because they get added to a references
list without ever being cleared. This fixes that by registering a
qdeleting signal on them and removing them from the list on delete.
## Why It's Good For The Game
Harddel fixes
## Changelog
🆑
fix: Fixed a hard delete that would occur with lua-created atoms.
/🆑
---------
Co-authored-by: Watermelon914 <3052169-Watermelon914@users.noreply.gitlab.com>
## About The Pull Request
Signals were initially only usable with component listeners, which while
no longer the case has lead to outdated documentation, names, and a
similar location in code.
This pr pulls the two apart. Partially because mso thinks we should, but
also because they really aren't directly linked anymore, and having them
in this midstate just confuses people.
[Renames comp_lookup to listen_lookup, since that's what it
does](102b79694f)
[Moves signal procs over to their own
file](33d07d01fd)
[Renames the PREQDELETING and QDELETING comsigs to drop the parent bit
since they can hook to more then just comps
now](335ea4ad08)
[Does something similar to the attackby comsigs (PARENT ->
ATOM)](210e57051d)
[And finally passes over the examine
signals](65917658fb)
## Why It's Good For The Game
Code makes more sense, things are better teased apart, s just good imo
## Changelog
🆑
refactor: Pulled apart the last vestiges of names/docs directly linking
signals to components
/🆑
## About The Pull Request
Damn that's a long title.
Admin Verbs can be used in the verb bar with hyphens instead of spaces
again.
## Why It's Good For The Game
Admin muscle memory
## Changelog
## About The Pull Request
Mothblocks recently talked about wanting an auxlua function to check if
the lua state was about to overrun its execution limit, and that was
something I had been thinking about for a long time prior, but never
bothered to actually implement until it was brought up. So I did just
that. The documentation has also been updated to include a description
of how the execution limit works and how to use the new function,
`over_exec_time`
## Why It's Good For The Game
Allows for lua loops of indeterminate length to do as much work as they
reasonably can in a single tick without setting off the execution
limiter and erroring out.
## Changelog
🆑
admin: Adds a new function for admin lua scripting, "over_exec_time",
for checking if lua code is running close to the execution limit.
Details are available in the lua editor's help menu.
admin: The execution limit is described in detail in the lua editor's
help menu.
/🆑
Co-authored-by: Mothblocks <35135081+Mothblocks@users.noreply.github.com>
Makes the code compatible with 515.1594+
Few simple changes and one very painful one.
Let's start with the easy:
* puts call behind `LIBCALL` define, so call_ext is properly used in 515
* Adds `NAMEOF_STATIC(_,X)` macro for nameof in static definitions since
src is now invalid there.
* Fixes tgui and devserver. From 515 onward the tmp3333{procid} cache
directory is not appened to base path in browser controls so we don't
check for it in base js and put the dev server dummy window file in
actual directory not the byond root.
* Renames the few things that had /final/ in typepath to ultimate since
final is a new keyword
And the very painful change:
`.proc/whatever` format is no longer valid, so we're replacing it with
new nameof() function. All this wrapped in three new macros.
`PROC_REF(X)`,`TYPE_PROC_REF(TYPE,X)`,`GLOBAL_PROC_REF(X)`. Global is
not actually necessary but if we get nameof that does not allow globals
it would be nice validation.
This is pretty unwieldy but there's no real alternative.
If you notice anything weird in the commits let me know because majority
was done with regex replace.
@tgstation/commit-access Since the .proc/stuff is pretty big change.
Co-authored-by: san7890 <the@san7890.com>
Co-authored-by: Mothblocks <35135081+Mothblocks@users.noreply.github.com>
* Moves spawners and decals to a different init/delete scheme
Rather then fully creating and then immediately deleting these things,
we instead do the bare minimum.
This is faster, if in theory more fragile. We should be safe since any
errors should be caught in compile since this is very close to a
"static" action. It does mean these atoms cannot use signals, etc.
* Potentially saves init time, mostly cleans up a silly pattern
We use sleeps and INVOKE_ASYNC to ensure that handing back turfs doesn't
block a space reservation, but this by nature consumes up to the
threshold and a bit more of whatever working block we were in.
This is silly. Should just be a subsystem, so I made it one, with
support for awaiting its finish if you want to
* Optimizes garbage/proc/Queue slightly
Queue takes about 1.6 seconds to process 26k items right now.
The MASSIVE majority of this time is spent on using \ref
This is because \ref returns a string, and that string requires being
inserted into the global cache of strings we store
What I'm doing is caching the result of ANY \ref on the datum it's
applied to. This ensures previous uses will never decay from the string
tree.
This saves about 0.2 seconds of init
When I changed the syntax of SS13.lua to account for the ability to properly index tables with datums, it turns out that the callbacks created for signal handlers and timeouts had circular references, resulting in hard-deletes.
My first solution was to make it so signal handler and timeout callbacks use weakrefs, which get resolved in lua_state/call_function, but it turns out that, when calling the signal cleanup function, a qdeleted-but-not-yet-garbage-collected datum's weakref resolves to null because datum.gc_collected is set to GC_CURRENTLY_BEING_QDELETED before COMSIG_PARENT_QDELETING gets sent.
To resolve this issue, Potato and Oranges both recommended that I make a snowflake variant of resolve which ignores whether the datum a weakref points to is qdeleted - only that it is null or it's weakref isn't the very weakref resolve was called on. This proc was given a lengthy autodoc comment describing when it should or shouldn't be used, to ensure it is only used in cases similar to the one I needed to create it for (needing to resolve a weakref to a datum in a COMSIG_PARENT_QDELETING handler registered on that very datum).
When trying to run getFlatIcon using lua scripting, I discovered that it was so expensive that, for certain atoms like complex humans, there is no way it will complete within the lua execution limit. While Mothblocks would suggest just raising the execution limit, the idea leaves a bad taste in my mouth due to the possibility of a script causing extreme lag by consistently overrunning the BYOND tick.
I instead elected to make /datum/auxtools_promise sleep once before invoking its assigned proc, thus immediately returning execution to lua, even if the proc being awaited wouldn't sleep. This allows for awaiting extremely expensive non-sleeping procs (like the aforementioned getFlatIcon)
* preparations for self-referential list conversion
* additional changes for the auxlua 1.1.1 update
* fixed a type in `SS13.await`
* bumps auxlua to 1.2.0
* bumps auxlua to 1.2.1
Atomized from #68816, with a little addition. Fixes some dumb formatting issues with the lua editor, adds a "jump to bottom" button when viewing the state log, and paginates the state logs.
This PR fixes this issue by making sure every proc called in the stack of /proc/wrap_lua_print which could sleep is called using INVOKE_ASYNC, and to prevent such problems in the future, marks all the wrappers as SHOULD_NOT_SLEEP(TRUE). I also figured out how to fix the dumb overflowing problem of the lua editor ui.
Due to lag concerns regarding lua states with a large number of global variables (including fields within global tables), I have made it so the global table and state log are hidden by default - they can be shown using a toggle button in the editor ui.
