Add req, res, tx to the id-denylist

[Mrtenz]

Description

This pull request adds req, res, tx to the id-denylist. This ensures we use consistent variable naming across repositories. These names are commonly used in some older projects, but should be written out in full for better readability, in my opinion.

Changes

1. Add req to the id-denylist.
2. Add res to the id-denylist.
3. Add tx to the id-denylist.

[legobeat]

I disagree on this one - is there really any room for confusion? Just seems to add friction for something that's primarily an aesthetic preference.

These are commonly recognized patterns. Would the next thing be enforcing identifier over id?

[Mrtenz]

Just seems to add friction for something that's primarily an aesthetic preference.

I don't see how it adds any friction. I agree that it's primarily an aesthetic preference though, but since we're already using error instead of err, message instead of msg, and so on, it feels appropriate to add these as well for consistency.

[legobeat]

Not really a fan of msg being there either but that's not really for here.

A few years in idiomatic Objective-C codebases is what tells me that this road is not one towards productivity. Especially if idiomatic consistency is a driver.

[Mrtenz]

I appreciate your opinion! I will leave this up for others to reply, and we'll see if we merge this or not based on that.

[Gudahtt]

As a style choice, I like what this rule is mandating. I find that I can read and understand code more easily if things are spelled out, even if I recognize the acronym immediately.

But I have always been on the fence about whether to include this rule here. This seems like something MetaMask teams might reasonably disagree on, so better suited for an ESLint config that isn't shared between all teams.

When it comes to style choices, I've tended towards only adopting the rules that make it significantly easier for engineers to contribute across teams. If the consistency win is big enough, style rules can reduce friction when moving between codebases. A rule like this might have a minimal impact here.

[Gudahtt]

Thoughts @rekmarks ? As the person who introduced this rule to begin with