CIP-0058? | Bitwise primitives #283

kozross · 2022-06-23T19:23:07Z

This improves upon, and supercedes, the original. Specifically, the formatting has been designed to be clear and readable on Github, and concerns vis-a-vis semantics and wording raised here.

A few issues await @michaelpj's responses; in particular, the following questions need to be addressed:

See rendered markdown.

…bitwise

michaelpj

Looking good!

michaelpj · 2022-06-24T09:27:59Z

CIP-?/README.md

+```
+Performs a bitwise shift of the first argument by a number of bit positions
+equal to the absolute value of the second argument, the direction of the shift
+being indicated by the sign of the second argument.


This still doesn't say how the direction is indicated!

Clarified in the text, also for rotations.

michaelpj · 2022-06-24T09:35:04Z

CIP-?/README.md

+of the represented byte sequence being treated in little-endian,
+least-significant-bit-first encoding. For example, consider the byte sequence 
+$s = 110110011100000$; the `BuiltinByteString` literal corresponding to this would
+be `"\217\224"`.


Aha, are you referring to how the IsString instance works? I think it would be clearer to avoid that. For the purposes of this document BuiltinByteStrings are just bytestrings.

Rewritten to avoid completely.

michaelpj · 2022-06-24T09:36:17Z

CIP-?/README.md

+all $i \in \\{ 0,1,\ldots,n \\}$, $s_i \in \\{ 0, 1 \\}$. A bit sequence 
+$s = s_n s_{n-1} \ldots s_0$ is a *byte sequence* if $n = 8k - 1$ for some 
+$k \in \mathbb{N}$. We denote the *empty bit sequence* (and, indeed, empty byte 
+sequence as well) by $\emptyset$.


How about this:
"...by $\emptyset$. We do not need to make any assumptions about how bytestrings are actually represented in Plutus, we give the semantics of the operations here, they can be implemented on whatever representation is actually used."

I've gone for a slightly different rephrasing: the ordering is still specified, but a little more clearly.

michaelpj · 2022-06-24T09:37:21Z

CIP-?/README.md

+
+We assume that `BuiltinByteString`s represent byte sequences, with the indexes
+of the represented byte sequence being treated in little-endian,
+least-significant-bit-first encoding. For example, consider the byte sequence 


endianness is about integer encoding!

This has now been moved to the relevant section.

michaelpj · 2022-06-24T09:38:36Z

CIP-?/README.md

+We describe the translation of `BuiltinInteger` into `BuiltinByteString` which 
+is implemented as the `integerToByteString` primitive. Informally, we represent
+`BuiltinInteger`s with the least significant bit at bit position $0$, using a 
+twos-complement representation. More precisely, let $i \in \mathbb{N}^{+}$. We 


"This corresponds to a little-endian, least-significant-bit first encoding."

michaelpj · 2022-06-24T09:39:34Z

CIP-?/README.md

+We describe the semantics of `shiftByteString` and `rotateByteString`.
+Informally, both of these are 'index modifiers' for bit sequences: given a
+positive $i$, the index of a bit in $s$ 'increases' in the result; given a
+negative $i$, the index of a bit in $s$ 'decreases' in the result. This can mean


Now you specify it. I just think this should be in the brief description too.

rphair · 2022-06-24T17:37:41Z

applying same labelling as held by deprecated version (#268 (comment))

CIP-?/README.md

KtorZ · 2022-10-25T07:53:24Z

CIP-?/README.md

+```
+
+Convert a non-negative number to its bitwise representation, erroring if given a
+negative number.


I reckon that the conversion uses big endianness. It could be nice to clarify it even though the interface described below seems pretty much agnostic to this.

Pretty sure it's supposed to be little endian.

CIP-?/README.md

KtorZ · 2022-10-25T08:04:41Z

CIP-?/README.md

+We observe that `byteStringToInteger` 'undoes' `integerToByteString`:
+
+```haskell
+byteStringToInteger . integerToByteString = id


Suggested change

byteStringToInteger . integerToByteString = id

byteStringToInteger . integerToByteString = identity

(because readers may not be familiar with Haskell, let's not force the bad choices of the base Prelude on them).

This assumes instead that they are familiar with one of the preludes that renames id to identity, which is even more unlikely than someone working on Cardano or Plutus not knowing Haskell.

No, it assumes they are familiar with word 'identity' which is pervasive in mathematics.

I mean seriously, for anyone outside of the the tiny Haskell bubble, 'id' means 'identifier'.

Except the laws are specifically written in Haskell terms throughout. Furthermore, I really don't think anybody reading this won't be familiar with Haskell.

This is merely a suggestion. Do as you please, but we aware that most people building on Cardano and participating in CIPs aren't actually that familiar with Haskell.

CIP-?/README.md

rphair

@kozross this also mainly looks good to me... I'll keep an eye out for a commit making the changes in @KtorZ's last review (especially updating to Proposed status) and then approve & merge it as well.

rphair · 2022-10-25T10:49:36Z

p.s. oops, I accidentally approved by ticking the wrong box... so we will wait for that commit before merging 🤪

perturbing · 2023-01-11T08:35:13Z

For what it is worth, I made a similar implementation that this cip aims to achieve building on other builtins (this is thus not as efficient as the solution of this cip). Its still a work in progress.

I mostly followed the interface of the modules Crypto.Numbers (1) and Data.Bits (2).

Examples, of some basic functions

Prelude > import Plutus.Crypto.Number.Serialize as S
Prelude S > bs = S.i2ospOf_ 32 (2^255)
Prelude S > bs
"\128\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL"
Prelude S > import Plutus.Data.Bits as B
Prelude S B > bs2 = B.setBit (S.i2ospOf_ 32 0) 255
Prelude S B > bs2
"\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\NUL\SOH"
Prelude S B > bs == B.reverseBS bs2
True

I am yet to implement the benchmarking of these functions (and optimize them). But if someone needs them, they can.

rphair · 2023-01-11T08:54:27Z

@kozross marking "Waiting for author" because as far as I can tell this is ready to merge once the corrections from @KtorZ's last review are done (just to assign the official number & status & correct a half-dozen spelling mistakes).

Co-authored-by: Matthias Benkort <5680256+KtorZ@users.noreply.github.com>

L-as · 2023-02-22T20:22:10Z

@rphair I've committed all suggestions except the controversial one about id (which I don't agree with either). I think this can be merged now?

L-as · 2023-03-09T18:00:43Z

PR: IntersectMBO/plutus#4733

michaelpj · 2023-03-10T11:44:00Z

@L-as might be helpful to add an Implementation Plan and Implementors as per the new CIP-001, since I assume you're now the implementor?

L-as · 2023-03-10T12:43:29Z

Yes, indeed, I will do that.

nielstron · 2023-07-18T15:06:35Z

There is also an implementation of bitmaps in the OpShin programming language and several builtins (such as chr, converting ints to corresponding unicode characters) that would benefit from these primitives

https://github.com/OpShin/opshin/blob/main/opshin/std/bitmap.py

* Add CIP * Notation fix-ups * Fix heading levels on use cases * Rewrite how shifts work * Clarify findFirstSet description * Clarify representation * Clarifications * Replace FFS with TZCNT and LZCNT, add finite field arithmetic example * Rewrite * Fix formatting * Apply suggestions from code review Co-authored-by: Matthias Benkort <5680256+KtorZ@users.noreply.github.com> --------- Co-authored-by: goverthrow <goverthrow@protonmail.com> Co-authored-by: Las Safin <me@las.rs> Co-authored-by: Matthias Benkort <5680256+KtorZ@users.noreply.github.com>

goverthrow and others added 7 commits May 27, 2022 16:53

Add CIP

7a1ecf4

Merge branch 'master' of https://github.com/goverthrow/CIPs into koz/…

147194b

…bitwise

Notation fix-ups

415e893

Fix heading levels on use cases

1ea1062

Rewrite how shifts work

ea8fe5c

Clarify findFirstSet description

115bfb1

Clarify representation

0b990c6

kozross mentioned this pull request Jun 24, 2022

Bitwise operations IntersectMBO/plutus#4733

Closed

58 tasks

michaelpj approved these changes Jun 24, 2022

View reviewed changes

rphair mentioned this pull request Jun 24, 2022

CIP-0058? | New CIP for bitwise primitives #268

Closed

rphair added Candidate CIP labels Jun 24, 2022

kozross added 4 commits June 27, 2022 14:22

Clarifications

0766678

Merge branch 'master' into koz/bitwise

b714f27

Replace FFS with TZCNT and LZCNT, add finite field arithmetic example

76ab90e

Rewrite

fd308ab

kozross force-pushed the koz/bitwise branch 13 times, most recently from 3bdad5a to f8057e5 Compare September 14, 2022 23:26

Fix formatting

a1b9ff0

kozross force-pushed the koz/bitwise branch from f8057e5 to a1b9ff0 Compare September 14, 2022 23:29

Benjmhart mentioned this pull request Oct 2, 2022

Priority CIP's for late 2022/Early 2023 input-output-hk/Developer-Experience-working-group#20

Closed

KtorZ approved these changes Oct 25, 2022

View reviewed changes

rphair approved these changes Oct 25, 2022

View reviewed changes

rphair added the Waiting for Author Proposal showing lack of activity or participation from their authors. label Jan 11, 2023

Apply suggestions from code review

e86732f

Co-authored-by: Matthias Benkort <5680256+KtorZ@users.noreply.github.com>

rphair merged commit 82e77b9 into cardano-foundation:master Mar 1, 2023

L-as deleted the koz/bitwise branch March 1, 2023 17:45

L-as mentioned this pull request Mar 1, 2023

Fully homomorphic encryption circuits in Cardano... #415

Open

rphair removed the Waiting for Author Proposal showing lack of activity or participation from their authors. label Mar 2, 2023

effectfully mentioned this pull request Jun 29, 2023

Add support for PubKey into PubKeyHash conversion on-chain IntersectMBO/plutus#5369

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

CIP-0058? | Bitwise primitives #283

CIP-0058? | Bitwise primitives #283

kozross commented Jun 23, 2022 •

edited by KtorZ

michaelpj left a comment

michaelpj Jun 24, 2022

kozross Jun 27, 2022

michaelpj Jun 24, 2022

kozross Jun 27, 2022

michaelpj Jun 24, 2022

kozross Jun 27, 2022

michaelpj Jun 24, 2022

kozross Jun 27, 2022

michaelpj Jun 24, 2022

kozross Jun 27, 2022

michaelpj Jun 24, 2022

kozross Jun 27, 2022

rphair commented Jun 24, 2022

KtorZ Oct 25, 2022

L-as Feb 22, 2023

KtorZ Oct 25, 2022

kozross Oct 25, 2022

KtorZ Oct 25, 2022

KtorZ Oct 25, 2022 •

edited

kozross Oct 25, 2022

KtorZ Oct 25, 2022 •

edited

rphair left a comment

rphair commented Oct 25, 2022

perturbing commented Jan 11, 2023

rphair commented Jan 11, 2023

L-as commented Feb 22, 2023 •

edited

L-as commented Mar 9, 2023

michaelpj commented Mar 10, 2023

L-as commented Mar 10, 2023

nielstron commented Jul 18, 2023

	byteStringToInteger . integerToByteString = id
	byteStringToInteger . integerToByteString = identity

CIP-0058? | Bitwise primitives #283

CIP-0058? | Bitwise primitives #283

Conversation

kozross commented Jun 23, 2022 • edited by KtorZ

michaelpj left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rphair commented Jun 24, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

KtorZ Oct 25, 2022 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

KtorZ Oct 25, 2022 • edited

Choose a reason for hiding this comment

rphair left a comment

Choose a reason for hiding this comment

rphair commented Oct 25, 2022

perturbing commented Jan 11, 2023

rphair commented Jan 11, 2023

L-as commented Feb 22, 2023 • edited

L-as commented Mar 9, 2023

michaelpj commented Mar 10, 2023

L-as commented Mar 10, 2023

nielstron commented Jul 18, 2023

kozross commented Jun 23, 2022 •

edited by KtorZ

KtorZ Oct 25, 2022 •

edited

KtorZ Oct 25, 2022 •

edited

L-as commented Feb 22, 2023 •

edited