CIP-0123? | Bitwise operations over BuiltinByteString

[kozross]

This adds bitwise operations to supplement the ones in CIP-122.

Human-readable version

[rphair]

@kozross exellent Rationale section & useful, precise Path to Active. Have put this on agenda to introduce at next CIP meeting (https://hackmd.io/@cip-editors/89) so it can be discussed along with these likely merging at that meeting:

• CIP-0121? | Integer-ByteString conversions #624
• CIP-0122? | Logical operations over BuiltinByteString #806

@michaelpj @zliu41 @MicroProofs can you review this when you get a chance? p.s. you would all be welcome to attend that meeting to help with the Plutus review in general.

[Crypto2099]

Since CIP-0122 is not yet merged in and "official" wouldn't it make sense to add this logic to that CIP before proceeding? I can see the value add of this but wonder what benefit we gain by breaking it up into two different CIPs rather than a single CIP to be implemented?

[rphair]

@kozross I believe I can see the advantage to keeping this CIP separate from candidate CIP-0122 — based on a layman's reading of the Rationale for this CIP, which considers the possibility of implementing this CIP's primitives based on those of the earlier CIP (technically still a possibility) — but still we will have to consider @Crypto2099's #825 (comment) at tonight's meeting: so if you could answer that question in the next hour (or attend the meeting) it would be very welcome.

[rphair]

p.s. @kozross at the CIP meeting just finished, we were wondering why these newer methods were not either added to CIP-0058 or submitted with an attempt to deprecate CIP-0058 (also asked in #806 (comment)).

[kozross]

p.s. @kozross at the CIP meeting just finished, we were wondering why these newer methods were not either added to CIP-0058 or submitted with an attempt to deprecate CIP-0058 (also asked in #806 (comment)).

These newer methods are modifications (and fairly minor ones) of CIP-58 equivalents. Deprecation of CIP-58 is already functionally the case anyway (as the Plutus team rejected the CIP-58 implementation): is there a more formal process I need to go through for actual deprecation of CIP-58?

[rphair]

@kozross then I believe it would be best to submit a PR on CIP-0058 changing the header text from

Status: Proposed

to whatever best explains how it's "deprecated" or "superseded" (whichever language you think is best), e.g.

Status: Inactive (deprecated by CIP-0121 and CIP-0122)

according to CIP-0001 > Status: Inactive.

BTW when you submit that PR (we can also submit it for you, if you let us know what exact status to change to), I would also contribute some format cleanup before we put it to rest because it was missed in #389 for that effort.

[kozross]

I think 'superseded' is the right term here. Will send that PR ASAP.

[kwxm]

This all looks perfectly fine: most of the design choices are kind of inevitable, so it's a lot more straightforward than the CIP for logical operations.

[kwxm]

I suppose there's a slight danger of confusion here because we index bits in bytestrings from the right (which I agree is a good choice) but if you write down the bytestring in some form such as [0x1a, 0x35] or 0x1a35 or (con bytestring #1a35) this function will actually find what appears to be the last bit set. I think that's an unavoidable consequence of the way we write things down though (and the cultural fact that we read from left to right), and I can't think of a more sensible name. I guess we should just make it clear how it works in the documentation.

[kozross]

Yeah, 'first' here means 'based on our indexing scheme'. This whole left-versus-right situation is quite unfortunate, but as CIP-122 explains, unavoidable.

[rphair]

@kozross if these 3 proposals cannot live happily in an updated CIP-0058, the next best thing is for their numbers to be sequential. However we do need the title also to be made more specific since "Bitwise operations" is far too general for the content of this CIP.

Please also rename the containing directory to CIP-0123 and the proposal filename to README.md. FYI compliments on your work were made all around at the CIP meeting tonight.

[kozross]

@rphair - Done. Great to hear my work is up to standard!

[rphair]

I think this fits in the same box as the recently merged #624 & #806 and it was again discussed (incidentally to the agenda) at tonight's CIP meeting, again without any reservations being expressed.

So @zliu41 @Ryun1 @Crypto2099 I'm going to reflect this in the Last Check tag, which should get it ready to consider merging at the next CIP meeting in 2 weeks' time. (cc @MicroProofs @kwxm)

[Crypto2099]

Seems like this one has appropriately addressed all discussion points and meets the criteria that 121 and 122 also met to be merged in so I see no further blockers to allow this to progress.