Files
Bubberstation/code/game/objects/items_reskin.dm
_0Steven aa5eddb99b Fix pride pin reskinning (#82920)
## About The Pull Request

**Edit: Since writing, this pr has been updated to address failing CI
based on code-general suggestions, invalidating the previous
descriptions. The previous descriptions has been included as spoilers
for posterity**

Right, so, this has gone from just a simple pride pin fix to realizing
CI fails with it to doing a more complex lasting fix based on
suggestions.

Recap time. Objects get reskinning set up if they have `unique_reskin`
set when `Initialize(...)` runs.

9145ecb7e1/code/game/objects/items.dm (L267-L269)
Because pride pins use a global list, we set it in `Initialize(...)`...
After we call the parent.

9145ecb7e1/code/modules/clothing/under/accessories/badges.dm (L196-L198)
Obviously this fails.

However, moving this *before* `Initialize(...)`, while fixing the issue,
causes CI to fail due to calling `register_context()` twice.
Why? Well, it's simple. We automatically call `register_context()` if we
have `unique_reskin` set, as seen above, but we *also* call it on
accessory `Initialize(...)` due to it having its own context.

0c562fd742/code/modules/clothing/under/accessories/_accessories.dm (L29-L31)
This causes it to try register the same thing twice, which doesn't
_break_ things, but it sure as hell isn't clean.

So talking about this with San in code general, we decided to try go
with the following:
We add two new procs, `setup_reskinning()` and
`check_setup_reskinning()`, and handle all this fuckery within those.
This lets subtypes override them with their own new checks or
differences in setup.
Then we override `setup_reskinning()` for `/obj/item/clothing/under` and
`/obj/item/clothing/accessory` to not register context again, and do the
same for `/obj/item/clothing/accessory/pride` but while also setting
`unique_reskin`.

This fixes it.

<details>
  <summary>Previous implementation for posterity</summary>
  
Back from my short code break, time to fix some of the things I've been
annoyed by.

Firstly, I noticed pride pins could no longer be reskinned since the
alt-click refactor.
Looking into it, this seems to be because we now only register this on
`Initialize(...)` if `unique_reskin` has been set:

9145ecb7e1/code/game/objects/items.dm (L267-L269)
While due to using a global list we don't set this in the item
definition, but in `Initialize(...)` :

9145ecb7e1/code/modules/clothing/under/accessories/badges.dm (L196-L198)
Where we call the parent proc _before_ setting `unique_reskin`, and thus
not registering our ability to reskin.

So all we do is set this to our global list _before_ we call the parent
proc.
```dm
/obj/item/clothing/accessory/pride/Initialize(mapload)
	unique_reskin = GLOB.pride_pin_reskins // Set before parent proc checks for it.
	. = ..()
```
This fixes it.
  
</details>


## Why It's Good For The Game

Fixes pride pin reskinning.
Theoretically makes it easier to avoid this happening in the future, and
allows `setup_reskinning()` to be manually called in the case of values
being edited post-initialize.

<details>
  <summary>Previous pitch for posterity</summary>
  
Fixes pride pin reskinning.
  
</details>

## Changelog
🆑
fix: Pride pins can be reskinned again with alt-click.
/🆑
2024-05-08 22:18:54 +01:00

87 lines
2.2 KiB
Plaintext

/// Called when alt clicked and the item has unique reskin options
/obj/item/proc/on_click_alt_reskin(datum/source, mob/user)
SIGNAL_HANDLER
if(!user.can_perform_action(src, NEED_DEXTERITY))
return NONE
if(!(obj_flags & INFINITE_RESKIN) && current_skin)
return NONE
INVOKE_ASYNC(src, PROC_REF(reskin_obj), user)
return CLICK_ACTION_SUCCESS
/**
* Checks if we should set up reskinning,
* by default if unique_reskin is set.
*
* Called on setup_reskinning().
* Inheritors should override this to add their own checks.
*/
/obj/item/proc/check_setup_reskinning()
SHOULD_CALL_PARENT(TRUE)
if(unique_reskin)
return TRUE
return FALSE
/**
* Registers signals and context for reskinning,
* if check_setup_reskinning() passes.
*
* Called on Initialize(...).
* Inheritors should override this to add their own setup steps,
* or to avoid double calling register_context().
*/
/obj/item/proc/setup_reskinning()
SHOULD_CALL_PARENT(FALSE)
if(!check_setup_reskinning())
return
RegisterSignal(src, COMSIG_CLICK_ALT, PROC_REF(on_click_alt_reskin))
register_context()
/**
* Reskins object based on a user's choice
*
* Arguments:
* * user The mob choosing a reskin option
*/
/obj/item/proc/reskin_obj(mob/user)
if(!LAZYLEN(unique_reskin))
return
var/list/items = list()
for(var/reskin_option in unique_reskin)
var/image/item_image = image(icon = src.icon, icon_state = unique_reskin[reskin_option])
items += list("[reskin_option]" = item_image)
sort_list(items)
var/pick = show_radial_menu(user, src, items, custom_check = CALLBACK(src, PROC_REF(check_reskin_menu), user), radius = 38, require_near = TRUE)
if(!pick)
return
if(!unique_reskin[pick])
return
current_skin = pick
icon_state = unique_reskin[pick]
to_chat(user, "[src] is now skinned as '[pick].'")
SEND_SIGNAL(src, COMSIG_OBJ_RESKIN, user, pick)
/**
* Checks if we are allowed to interact with a radial menu for reskins
*
* Arguments:
* * user The mob interacting with the menu
*/
/obj/item/proc/check_reskin_menu(mob/user)
if(QDELETED(src))
return FALSE
if(!(obj_flags & INFINITE_RESKIN) && current_skin)
return FALSE
if(!istype(user))
return FALSE
if(user.incapacitated())
return FALSE
return TRUE