Title. The clickon proc calls params2list only after the component signal was sent notwhitstanding clickon comsig procs also needing to access these parameters in order to work properly, which means every one of these comsig procs has its own params2list(params) call. This problem isn't exlusive to these comsig procs. For some reason the params text string is passed down to the melee click chain, afterattack and afterattack_secondary instead of the params/modifiers list... buuut we'll get there later. TL;DR params2list is being called more times than necessary because of a suboptimal old code.
Another issue: Some of these comsig procs ignore parameters such as shift, which is used to examinate things and access popup menus, alt, ctrl and/or middle mouse button. This is the case of the honorbound mutation datum, shy and shy in room components, mechas and the clown car. This is of course also getting fixed.
I'm adding a circuit component that can print text string on a paper object in a variety of colors and font typefaces (currently only web-safe ones are available, maybe i'll add some fancy ones in the future but they'd need to be imported either through @import of @font-face in a separate CSS not imported by every tgui UI).
It's important to note that because the UI sanitizes new text inputed by users and not what's already written on the paper (so the pen_color and pen_font don't be purged in the process), we can't safely have these strings "printed" into the info variable directly, because of that these values will be stored in two new list variables, one for the text and one for font color, face and the signature. When the paper sheet UI is opened, these will be sanitized and then parsed into the text, so the next time the paper is edited we can clear these two lists.
Obviously better than a hacky byond proc - parsemarkdown() is outdated af -, albeit a bit messy... like the rest of paper code.
Requires #62033.
Components like the MMI one can't be added to circuits more than once since they may register signals with same proctype and similar things which make for some tangled up race conditions if more than one is present.
Unfortunately this safety can be bypassed - with little gain alas, an MMI can't be inserted by attacking the component with it. it needs a shell - by using a module component. That's no good. So I'm adding a flag that can be used to stop certain components from being added to module components.
Made all say()s deal with encoding, audited all uses of say() to prevent double encoding or like, manually inserting span().
I left some stuff without sanitize that only draws from the code, since it's hell to clean up otherwise. That
and I let admins do whatever the fuck they want
if you create a bunch of signallers then spam the send signal button you can grind the server to a halt at low numbers and crash it at high numbers of signallers
This is an imperfect fix as a proper fix would refactor signal datums to use a subsystem but I dont feel like rewriting all of signal code thank you very much
Co-authored-by: Zephyr <12817816+ZephyrTFA@users.noreply.github.com>
Co-authored-by: Mothblocks <35135081+Mothblocks@users.noreply.github.com>
* tgui bsod
* debug disconnections
* prelim
* recomment
* set_value -> put ._.
* DAMN IT
* reinsert subsystem
* prepare
* unditch signals
* remove combiner
* remove combiner some more
* how did router.dm get here? deleting.
* These two COMSIGS should be one.
* critical typo
* inline cast
* have your signals
* Have your set_input & set_output.
* make compile
* upgrade save/load to n-to-n-wires
* have your documentation
* have your unsafe proc
* pay no attention to the compile errors
* unlist the ref
* paste my for block back in ._.
* fix manual input
* oops pushed too soon
* Have your !port.connected_to?.length
Co-authored-by: Watermelon914 <37270891+Watermelon914@users.noreply.github.com>
Co-authored-by: Watermelon914 <37270891+Watermelon914@users.noreply.github.com>
Fixes to the cooldowns. Also circuits now use an input trigger/output trigger system which the speech component hadn't yet adopted because it was made before that system was introduced and left out when changing most components to use this system.
Also temporarily disables input/output signal ports for the prebuilt speech relay circuit until I can properly code in a way to load circuits.
Adds the airlock shell. The circuit has full control over the airlock.
Refactors USB code to be easier to use for less experienced coders.
Implements USB cables for the binary valve to be able to open/close the valve.
Adds a private channel for radios that only lets circuits with the same owner's ID to interact with it.
Adds the MMI component which outputs signals whenever they click somewhere/try to move.
Allows inputting signals into ports, making it easier to debug integrated components.
Adds the multiplexer circuit component - en.wikipedia.org/wiki/Multiplexer
Circuit components can now be directly inserted into shells rather than having to take the integrated circuit out.
Special information can be accessed from components now through the "Info" button besides the eject button on a component.
