`Int` feature tracking ticket

[erik-3milabs]

Related

• Feature wishlist tracking ticket #1 (feature request), and
• Supporting Signed Integer Arithmetic #624 (initial discussion).
• [WIP] Int - sign-and-magnitude #694 (abandoned attempt involving sign-magnitude representation),

Features

• add, sub, mul, div Int - two's complement #695
• and, or, xor, not, shl, shr Int - Bit ops #697
• Int::square operations Int - square operations #734
• vartime division Int - vartime division operations #735
• mul_uint, div_uint Int - Int-Uint interoperability #736
• Encoding support (from_be_slice, from_le_slice)
• inv_mod Int - inv_mod #737
• gcd Int - gcd #698
• xgcd
• prop-tests (see also Int - two's complement #695 (comment))

[ycscaly]

@tarcieri with #695 merged to signed-int, whats the plan for this feature? Specifically, will it be included in the next release? Could be helpful esp. because of comparability other crates like elliptic-curves depending on crypto-bigint

[tarcieri]

@ycscaly yeah, we should be able to get at least a preliminary implementation into the release and then we can additively improve it.

I went ahead and added it to the list in #681

[ycscaly]

@ycscaly yeah, we should be able to get at least a preliminary implementation into the release and then we can additively improve it.

I went ahead and added it to the list in #681

Great thanks!

[tarcieri]

I'm planning on trying to wrap up crypto-bigint v0.6.0 stable, and I'd like to save this for crypto-bigint v0.7, FYI.

Fortunately, all work is happening on the signed-int branch.

[ycscaly]

I'm planning on trying to wrap up crypto-bigint v0.6.0 stable, and I'd like to save this for crypto-bigint v0.7, FYI.

Fortunately, all work is happening on the signed-int branch.

Why can't we have the signed integers on 0.6? There's quite a bit already implemented here, and we are actively using it.
We are planning to lunch our product very soon, and would be better to have this as part of a release if possible.

Thanks!

[tarcieri]

Aah, well then. We can potentially merge the signed-int branch, although there are still a few outstanding PRs and work seems to have stalled.

If it doesn't make it in, we'll be able to merge it in a v0.7 release, which I would expect some time in Q2 2025.

[fjarri]

What breaking changes do signed integers introduce?

[tarcieri]

As far as I'm aware, they're completely additive

[fjarri]

So they can probably be added in 0.6.1 which would be much easier to handle in the dependencies?

[tarcieri]

Sure, no need to block a v0.6 release. It would probably be good to get a PR to merge the branch open first so we can confirm there aren't any breaking changes, though.