ref(core): Make hint callback argument non-optional

[timfish]

Now we allow mutating of properties in the hint to pass through things like attachments in beforeSend and Event processors, the hint is always defined, even if it's just an empty object.

This PR removes the optional from the types which means users will no longer need to do unnecessary things like this to work around the incorrect types:
https://github.com/getsentry/sentry-javascript/pull/5140/files#diff-51dd65d366cb1690b4b8b45f0f835ae66f2a20b4d7ee77c0b6c1f2efd28dcb55R48-R54

Because theses are callback parameters and users can just leave off parameters they're not using, this is not a breaking change and means the types more accurately reflect what we're passing to the callback.

Even if a user has code with an optional parameter type defined:

addGlobalEventProcessor(async(event: Event,hint?: EventHint)=>{

their code will potentially be left with unnecessary null checks but will compile without errors.

I guess there could be lints I'm unaware of that check signatures match?

[timfish]

Fixing these types to actually represent what is passed has highlighted that this would have always been logging an empty object for an empty hint.

[lobsterkatie]

Man, sometimes TS makes my brain hurt.

typeOneOrTwoParamCallback=(a: string,b?: string)=>void;typeTwoParamCallback=(a: string,b: string)=>void;typeOneOrTwoParamAccepter=(cb: OneOrTwoParamCallback)=>void;typeTwoParamAccepter=(cb: TwoParamCallback)=>void;functiononeParamCallback(a: string): void{}functiononeOrTwoParamCallback(a: string,b?: string): void{}functiontwoParamCallback(a: string,b: string): void{}// Currently, we have a OneOrTwoParamAccepterconstiTakeAOneOrTwoParamCallback: OneOrTwoParamAccepter=(cb: OneOrTwoParamCallback)=>{cb("first param")cb("first param","second param")};iTakeAOneOrTwoParamCallback(oneParamCallback);iTakeAOneOrTwoParamCallback(oneOrTwoParamCallback);iTakeAOneOrTwoParamCallback(twoParamCallback);// After, we'll have a TwoParamAccepterconstiTakeATwoParamCallback: TwoParamAccepter=(cb: TwoParamCallback)=>{cb("first param","second param")};iTakeATwoParamCallback(oneParamCallback);iTakeATwoParamCallback(oneOrTwoParamCallback);iTakeATwoParamCallback(twoParamCallback);

My linter hates this,

but it does compile, weirdly.

In any case, you're correct that we don't break anyone by doing this - that erroring case is under the current scheme, not the new one. And it is nice not to have to (((thingWhichIsMarkedOptionalButIActuallyKnowIsThere || {}).anotherSuchThing || {}).yetAnotherOne || {}).ughFinally = thisIsAPain all the time. So I think this is a good change.

[timfish]

Whichever of #5140 or this PR gets merged first, I'll clean the other one up before merging because it'll simplify the null checking.