Just discovered a bunch of nasty bugs in supac. I clearly wasn't testing it to the bone in the recent commits. My bigger issue is that I add an option to one place and forget to add it in another place. I think I'd just make it a point to mark all the places of change first and then make the actual changes, since a level order travelsal requires less contextual overhead for my tangerine sized brain.

My folks at cargo-bin could've just created a JSON array and called it a day, but no, they had to be smartasses and instead just spam multiple JSON objects one after the other in the list. They clearly weren't paying attention on how to remove it either. This is pathetic. Ok, they atleast do a basic check of the binary being there or not, and install it then, so not bad entirely. But still, one can't help but wonder if it's just utter stupidity at their end. I'd have happily used serde-json, but now I have to sit down and do string parsing and then throw the results into the parser, and all of this is, to say the least, extremely annoying.

Shit sucks ass in my code. All Some(X) should be replaced with [X] because that makes much more sense to me in general. Like, you're filtering shit out before concatenating anyways, how does it matter if you use Option to improve readability or whatever. Gosh, pathetic.

Apart from that, STOP using anyhow's methods. I've implemented specific ones for my own usage, I should be using those. WHY AM I NOT?! URGH!

And a neater dry-run should be in order, with that ridiculous prompt, it looks hideous. Might just add owo colors to the deps and use that. Binary sizes are still small anyways. Idk, is 20 MB small?

TOO CLICKY! I just can't. Also, the UI kinda sucks ass, ngl. Inconsistency is really pathetic.

I just saw the unstaged changes, and shiver me timbers! I'm kinda nervous about staging them, but eh, it's fine. It's mostly just me anyways. So I'll go ahead and stage in hunks and let the individual commits be. As long as HEAD compiles.

On a related note, I might be able to just steal the binstall parsing code from metapac.

And I have another extremely messed up idea! I might just keep using the serde-json failures to receive the last number of the Err as the next position to split from. Seems to be a decently good idea, if I'm being honest here. This might be the one to go, yup.