Improved emote parsing in PrivateMessage

[kevinrpb]

Hi there!

I've been using your package for a personal project and it is great! Thank you very much for all the work. Since it has been very helpful, I thought I'd give back by improving something I found was not as straight forward as other aspects of the message.

Up to now, the parsing of the emotes in the message was a bit odd, specially for handling messages where the same emote repeats multiple times.

I made a small utility function that takes care of better parsing the information from the message into an array containing a new struct EmoteReference. This struct includes the id of the emote, the name (aka the text used to represent the emote), and the indices in the message that correspond to the emote. The idea is that having all this information it is easier for a user of the package to handle emotes.

Please, let me know if there's anything out of place or if I should include some more test cases (although I feel like the one included already handles all cases).

Cheers!

[MahdiBM]

Thanks for the PR.
It looks good but I would like some fixes/improvements.
I've left some comments.

[MahdiBM]

We cannot use this kind of variable binding because we need to support older swift versions than 5.7.
change this to let lastID = lastID,

[kevinrpb]

Got it! That's my bad, I could only run the tests on Xcode 14 :)

[kevinrpb]

I made the changes, thank you for taking the time to review the PR.

However, I feel like Equatable conformance is at least nice since it allows automatic synthesis. Let me know if there are any downsides to this that I might be missing.

Cheers 😃

[MahdiBM]

Thanks for the changes, I'll take a look.

I do agree that the automatic synthesized conformance for Equatable is nice, but then if I want to allow it, one could argue "Why the other structs don't have the conformance?".
So its a question of if I want to go all in for the conformance or not.
I also like to keep the library minimal. I've been in situations that i would have preferred TwitchIRC's stuff be Codable so its easier for me to encode them, but it wouldn't make sense to add it to the library because its unrelated to the library and it was just my needs.

[MahdiBM]

Thanks again. This looks good.
I'm going to merge this and do an extra commit for some nit-picky cleanup.
You / I can utilize Emote in other files as well. For now this is enough and can act like a test to see if Emote works as expected.

[MahdiBM]

This will be tagged in a few days. This also contains a breaking change, so hopefully its worth it.

[kevinrpb]

I do agree that the automatic synthesized conformance for Equatable is nice, but then if I want to allow it, one could argue "Why the other structs don't have the conformance?".

Yes, I can see that. It's really not a big problem, and "keeping it simple" is a good argument for it. If only the compiler could auto-synthesize from an extension 😄.

You / I can utilize Emote in other files as well. For now this is enough and can act like a test to see if Emote works as expected.

I've been using it in my own app with good success to load up the emotes. The only thing I am missing now is getting emote sets that are not global or the room's. I will look into that and perhaps send another PR.

This will be tagged in a few days. This also contains a breaking change, so hopefully its worth it.

Understood 👍🏽