Simian is deprecated in favour of spitroast, which originated in a Discord client modification, and has now been split out into a separate lib.
It is currently the most featureful patcher and is a collaborative effort between many contributors, including myself. It is also smaller.
Please use it instead, though [email protected] provides a wrapper around it
if you need multiple patchers with separate cleanupAll()s.
Flexible JS monkey-patching simple enough for monkey brains!
Monkey-patches JS functions on-the-fly.
So if you have a function that is being used, and you want to hijack it to run your code, Simian can do that.
You can apply multiple patches to any given function, and can undo the patches in any order.
Originally cause I wanted to try make a patcher without looking at existing patchers for the challenge, but also now to be used in one of my friends' projects.
// 1: import it
import Patcher from "simian";
// 2: create a patcher
const patcher = new Patcher();
// 3: some functions to patch!
function wintest(a) { return a * a }
const obj = { objtest: (a, b) => a / b };
wintest(5); // 25
obj.objtest(10, 3); // 3.3333333333333335
// 4: get patching
const unpatchWin = patcher.after("wintest", window, (args, returnVal) => args[0] + returnVal);
const unpatchObj = patcher.after("objtest", obj, ([a, b]), ret) => (ret + b) * a)
// 5: the functions are now patched
wintest(5); // 30
obj.objtest(10, 3); // 63.33333333333334
// 6: remove a patch
unpatchObj();
obj.objtest(10, 3); // 3.3333333333333335
// 7: clean up all patches
// also removes any data stores attached to objects not part of the patcher
patcher.cleanupAll();
wintest(5); // 25after patches attach on the end of functions, and allow you to modify the return value of the function (or not! - simply don't return anything!). The original function is first called as usual, then your code is passed the args as an array, and the return value, which you can modify.
patcher.after: (funcName: string, obj: unknown, patch: (args: unknown[], ret: unknown) => unknown) => () => voidbefore patches attach on the front of functions, and allow you to modify the arguments of the function, before sending them on to the original function. Your code is first given the args as an array, and you simply return the args to pass to the original function, again as an array. Then the original function is called as usual with your new args.
patcher.before: (funcName: string, obj: unknown, patch: (args: unknown[]) => unknown[]) => () => voidinstead patches replace the function entirely, taking control over the args, return value, and side effects. However, if you wish to use the original function for anything (including calling it!), you are also passed it, though of course you don't need to take it as an arg - (args, orig) => {}, args => {} are both valid.
patcher.instead: (funcName: string, obj: unknown, patch: (args: unknown[], func: Function) => unknown): () => voidYou can also pass a string to new Patcher() to modify the label of the patcher's data store symbols.
This is NOT necessary for using multiple patchers at once, as every call of Symbol() returns a globally unique value.
This is not really useful directly, but recomended: it does mean that if others are going to see Simian stores while debugging theirs via devtools, its going to be obvious that they came from your software.
Simian is written in TS, so providing type decls is a no-brainer. Just import as usual and go, Typescript decls included.
No flow types unless I get bored at some point, but probably just use flowgen.
Cause monkey patching, and monkeys are simians, and, uh, yeah I'm not good at names... 😅
And as for monkey patching itself, according to wikipedia, it comes from the earlier term "guerrilla patch", as in stealth, and which sounds like "gorilla".