major rewrite

[Techatrix]

I've been planning to experiment with adding a different fuzzing mode but the code seemed very hard coded into markov.
These changes should make extending sus with new fuzzing modes much easier.

I wanted to test this with the fuzzing website but I'm not smart enough to get it running.
professional JS Developer, @SuperAuguste could you help me with this?

Here are some of the notable changes:

• revampt cli interface (every argument is named instead of positional)
• well specified interface between fuzzer and modes (see mode.zig)
• no temporary files while fuzzing, everything is stored in memory until a crash occurs
• replaced uri.zig in favor of std.zig
• compression with deflate is now configurable
• the best behavior mode is now function (but its useless anyway) (it found some files that instantly crash zls)
• fixes Switch to std.json #17
• uses transition from tres to std.json ziglibs/zig-lsp#6

[SuperAuguste]

LOL hello @Techatrix, professional JavaScript developer here; the current website / website code is kinda totally dead because it was designed when we had no money and no fuzzing results (we now have 3.6GBs of fuzzing data so it OOMs really badly) so I'll need to rewrite the website as well as how we encode fuzzing data (so it's easily storable in a bucket environment / indexable in some sort of database - yes we actually need one of those now lol).

I can still help you get the current version of the website running (should still work for a few thousand results) tonight over VC if you'd like! :) (and maybe we can play some CS:GO too :^))

[SuperAuguste]

I'm gonna open a PR into this PR for my new website/backend work and we can merge it into this, then merge this into main :) - this should ensure that all our fuzzing is 💯

[SuperAuguste]

(Sorry, didn't want to mess with this branch directly, but Git forced me to with its unique handling of case-sensitive filetrees)

[SuperAuguste]

tysm!