feat: publicly export and document the zip64 threshold constants

[cosmicexplorer]

See zip-rs/zip-old#402. Closes #60.

[cosmicexplorer]

It looks like the added test is failing:

failures:

---- src/spec.rs - spec::ZIP64_BYTES_THR (line 24) stdout ----
Test executable failed (exit status: 101).

stderr:
thread 'main' panicked at src/spec.rs:18:2:
assertion failed: zip.start_file("one.dat", options).is_err()

I'll take a look into why this behavior changed since zip-rs#402 and try to bisect the failure.

[cosmicexplorer]

This may be able to use the work from #67.

[Pr0methean]

Are you still planning to work on this PR, or have you folded it into one of the bigger ones?

[cosmicexplorer]

@Pr0methean this should be ready to review -- I hadn't gotten the hang of the nifty new behavior to rewind any file entries upon write or other errors before, but the test is working now. Fantastic functionality, so incredibly useful.

[Pr0methean]

@cosmicexplorer Sorry if the merge for this PR is delayed. I've been experimenting with new ways to generate and improve the fuzz corpus, and have found more panic-causing bugs tonight than I would've expected in a year -- and that's before cross-pollinating the corpora from different restarts!

[Pr0methean]

Of course, if you're able to fix any of these bugs yourself in separate PRs, I'll merge them very gratefully.

[cosmicexplorer]

Rebased and fixed conflicts; sorry for dropping this one! really want to get a new version of the parallel extraction PR #72 going (since that one is much more straightforward/less error-prone than the rest of #193, and I believe also much more requested by users), but will absolutely be trying to keep an eye out for bugfixes of that sort. I really like reading the code you've added to this crate and fixing bugs is a joy. Please especially feel free to ping me in the future if there's anything in particular you might like an extra hand with.

[cosmicexplorer]

CI is failing on a separate change, I'll push a quick fix for that now in a separate PR.

[cosmicexplorer]

Ping @Pr0methean!

[Pr0methean]

This has a merge conflict, which I'll try to resolve once the queue is empty.

[cosmicexplorer]

The merge conflict mysteriously disappeared!

[Pr0methean]

It also mysteriously reappeared - apparently the conflict was with another PR that got merged today but had been queued before.

[Pr0methean]

The cause of the conflict is that spec.rs no longer imports core::mem::align_of.

[cosmicexplorer]

Fixed!