`RLP` redesign #137

letypequividelespoubelles · 2025-03-28T11:40:41Z

No description provided.

Signed-off-by: F Bojarski <[email protected]>

cla-assistant · 2025-04-01T07:39:16Z

All committers have signed the CLA.

spec/_all_spec.tex

Signed-off-by: F Bojarski <[email protected]>

letypequividelespoubelles · 2025-07-31T12:03:19Z

rlp_patterns_v2/_all_rlp_patterns_v2.tex

delete all the pattern_v2 folder now I guess ?

letypequividelespoubelles · 2025-07-31T12:03:30Z

rlp_txn/.DS_Store

letypequividelespoubelles · 2025-07-31T12:04:36Z

rlp_txn_v2/columns/shared.tex

+\end{enumerate}
+The following are \textbf{counter-constant}, \textbf{binary} columns used in particular in the \rlp{}-ization of \textbf{access lists},
+see section~(\ref{rlp txn v2: phase constraints: access list: approach}).
+\begin{multicols}{2}


all this batch of columns should be in compt perspective

That's what I did initially when I was presenting this to you, but I decided against it after consideration: you won't save on columns since the COMPUTATION perspective is already the widest + you get would get extra CONTEXT[i] = 1 pre-conditions all over the place.

letypequividelespoubelles · 2025-07-31T12:05:30Z

rlp_txn_v2/columns/computation.tex

+	$k=1, \dots, \numberOfCounterConstantAuxiliaryDataColumns$:
+	\ccc;
+	used to store auxiliary data;
+    \item


I guess this has to disappear now ? seems a duplicate of the previous batch of columns

There is a difference: one batch of columns is counter constant, the other isn't. So I split the old TMP_k column batch in two.

If you compile it you'll find that there is a CCC in the column names. The column description also contain c.c.c. which stands for "counter constant column".

letypequividelespoubelles · 2025-07-31T12:06:48Z

rlp_txn_v2/generalities/binarities.tex

we have to add the sub phase columns of the accessList ?

Yes, that's an omission. Fixed it.

letypequividelespoubelles · 2025-07-31T12:07:40Z

rlp_txn_v2/generalities/constancies.tex

+
+Given the above definitions we require that
+\begin{enumerate}
+	\item \CFI, \typeZeroTx{}, \typeOneTx{}, \typeTwoTx{} be transaction-constant


up to type4

we want replayProtection and yParity to be tx-constant too ?

I didn't yet add the extra transaction types, mostly because they weren't included in the initially but also because we have time before including account delegation transactions.

BTW for account delegation transactions we may have to add more binary columns or re-purpose the newly introduced binary columns.

letypequividelespoubelles · 2025-07-31T12:22:08Z

rlp_txn_v2/specialized/processing/address.tex

+\begin{description}
+	\item[\underline{\underline{Setting \maxCt{}:}}] 
+		we impose $\maxCt _{i + \relof} = 2$
+	\item[\underline{\underline{Setting \rlpTxnSharedColumnIsPrefix{}:}}]


why we want this ?

letypequividelespoubelles · 2025-07-31T12:23:07Z

rlp_txn_v2/specialized/processing/byte_string_prefix.tex

+				nBytes    = \locRlpPrefixByteSize ,
+			}
+		\]
+	\item[\underline{\underline{Imposing nontriviality:}}] 


I'm not sure we need to impse nontrivialoty somewhere ... checking

letypequividelespoubelles · 2025-07-31T12:23:36Z

rlp_txn_v2/specialized/processing/bytes_32.tex

+\begin{description}
+	\item[\underline{\underline{Setting \maxCt{}:}}]
+		we impose $\maxCt _{i + \relof} = 2$
+	\item[\underline{\underline{Setting \rlpTxnSharedColumnIsPrefix{}:}}]


why do we want this ?

The whole purpose of \rlpTxnSharedColumnIsPrefix has to be re-discussed I believe. In the cases of the INTEGER, BYTES32 and ADDRESS we deal with RLP prefixes simultaneously with the data.

letypequividelespoubelles · 2025-07-31T12:28:08Z

rlp_txn_v2/phase_constraints/to_address.tex

setting phase_end is missing

letypequividelespoubelles · 2025-07-31T12:30:04Z

rlp_txn_v2/phase_constraints/data/payload_size_analysis.tex

+\]
+We impose the following constraints
+\begin{enumerate}
+	\item we impose


this is true for all the data phase, not only the prefix

LX and LT are phase constant except for the first phase of a transaction. This was added in a recent commit.

letypequividelespoubelles · 2025-07-31T12:33:40Z

rlp_txn_v2/phase_constraints/data/payload_size_analysis.tex

+				anchorRow        = i                          ,
+				relOffset        = 0                          ,
+				byteStringLength = \locTransactionPayloadSize ,
+				firstByte        = \locFirstByteOrZero        ,


we shouldn't constraint to be zero in case of empty.
In case it's empty, the next row will be a transaction row, so exo_data9_(i+1) might me the txn column of a shared register. The easier is to constraint the value only if payload is non empty

If the byteStringLength is set to 0 what we do with firstByte will not matter. I'd rather provide it with a value that is indeed a byte value rather than have it undefined, which the above achieves. Am I missing something ?

letypequividelespoubelles · 2025-08-01T01:53:52Z

rlp_txn_v2/phase_constraints/access_list/countdowns/addresses.tex

+		\[
+			\locStartRlpizingNewAddress _{i} \define
+			\left[ \begin{array}{cl}
+				\cdot & \rlpTxnSharedColumnIsAccessListAddress _{i} \\


wouldn't it make more sense to decrement by one when starting the rlp of an access list tuple as we require the address constancy fpor the all tuple ?

letypequividelespoubelles · 2025-08-01T01:58:00Z

rlp_txn_v2/phase_constraints/access_list/countdowns/access_list_RLP_length.tex

+\]
+The purpose of \locAccessListLengthCountdown{} is to \textbf{initially} hold $\| s( T_\textbf{A} ) \|$
+and then count down to $0$ as the access list gets processed.
+\begin{description}


I don't see the initialization ?

Same idea as for rlp_txn_v2/phase_constraints/access_list/countdowns/access_list_item_RLP_length.tex

letypequividelespoubelles · 2025-08-01T01:58:12Z

rlp_txn_v2/phase_constraints/access_list/countdowns/access_list_item_RLP_length.tex

+for the various access list items that make up $T_\textbf{A}$,
+and then count down to $0$ as the access list item gets processed.
+\begin{description}
+	\item[\underline{\underline{Updates:}}]


I don't see the initialization ?

IIRC there is no explicit initialization: you set the value during the "access list tuple prefix" subphase of IS_ACCESS_LIST and it decrements throughout the "access list tuple" processing.

letypequividelespoubelles · 2025-08-01T01:59:24Z

rlp_txn_v2/phase_constraints/access_list/subphases/prefix_of_full_access_list.tex

+We therfore impose that
+\begin{enumerate}
+	\item
+		$\limbBelongsToBothLtAndLx {


this is the case of all the access list, not only this subphase

I made some changes to the LT / LX constraints so that this setting remains constant throughout every processing phase (except the initial one.)

cursor

Bug: Macro Argument Labeling Mismatch

The wcpGeneralizedCallToIszero macro displays incorrect and inconsistent argument labeling. Although the ISZERO operation takes a single argument, the macro's output labels imply two arguments ("1st argument hi/lo" and "2nd argument hi/lo"). Specifically, the labels "2nd argument hi/lo" are erroneously associated with variables cmdWCP@var@argOneOne and cmdWCP@var@argOneZero, which are components of the first argument. All underlying variables (argOneThree through argOneZero) consistently indicate a single, multi-limb argument.

pkg/xkeyval_macros/wcp_calls.sty#L146-L148

linea-specification/pkg/xkeyval_macros/wcp_calls.sty

Lines 146 to 148 in b02d5bb

    
           \utt{1st argument lo:} & \cmdWCP@var@argOneTwo    \\ 
        
           \utt{2nd argument hi:} & \cmdWCP@var@argOneOne    \\ 
        
           \utt{2nd argument lo:} & \cmdWCP@var@argOneZero   \\

pkg/wc3.sty#L142-L149

https://github.com/Consensys/linea-specification/blob/b02d5bb412d7c2ef18ec750223d33355bd176faa/pkg/wc3.sty#L142-L149

cursor · 2025-08-01T20:54:45Z

pkg/rlp_log.sty

@@ -18,8 +18,8 @@
 \newcommand{\subPhaseIdDataSize}{\red{83}}
 \newcommand{\subPhaseIdTopicDelta}{\red{96}}

-\newcommand{\subPhaseIdWeightIsPrefix}{\red{6}}
+\newcommand{\subPhaseIdWeightisPrefix}{\red{6}}


Bug: Macro Naming Inconsistency

The macro \subPhaseIdWeightIsPrefix was incorrectly renamed to \subPhaseIdWeightisPrefix in pkg/rlp_log.sty. This introduces a naming inconsistency, as the lowercase 'i' in isPrefix deviates from the established IsPrefix convention used by similar macros (e.g., \subPhaseIdWeightIsOt, \subPhaseIdWeightIsOd) and throughout the codebase. This inconsistency is propagated to files like log_info/verticalization.tex and rlp_txnrcpt/global_constraint.tex, and appears to be a typo.

feat: RLP redesign

899b8cf

Signed-off-by: F Bojarski <[email protected]>

letypequividelespoubelles self-assigned this Mar 28, 2025

letypequividelespoubelles linked an issue Mar 28, 2025 that may be closed by this pull request

RLP_TXN redesign #62

Open

3 tasks

letypequividelespoubelles and others added 10 commits March 28, 2025 12:40

Merge branch 'main' into 62-rlp_txn-redesign

b1dffaa

Merge branch 'main' into 62-rlp_txn-redesign

567bc93

make it compile

367a75a

Signed-off-by: F Bojarski <[email protected]>

must enforce wcp result

460b28f

Signed-off-by: F Bojarski <[email protected]>

add more patterns

9b5a6fd

Signed-off-by: F Bojarski <[email protected]>

rlp utils module

eb34c21

Signed-off-by: F Bojarski <[email protected]>

partlyinteger inst

2156327

Signed-off-by: F Bojarski <[email protected]>

finish integer + feat in wcp

605ba55

Signed-off-by: F Bojarski <[email protected]>

finish rlp utilms modukle

69dbb39

Signed-off-by: F Bojarski <[email protected]>

Merge branch 'main' into 62-rlp_txn-redesign

c3419e3

OlivierBBB added enhancement New feature or request hardfork cancun EIP's for the Cancun hardfork arithmetization labels Apr 6, 2025

OlivierBBB added this to the Cancun milestone Apr 6, 2025

Merge branch 'main' into 62-rlp_txn-redesign

fa74edc

OlivierBBB reviewed Apr 8, 2025

View reviewed changes

spec/_all_spec.tex Outdated Show resolved Hide resolved

OlivierBBB added 10 commits April 8, 2025 11:45

feat: add RLP_UTILS module flags etc ...

78aed41

feat: remove the commenting of files from the _all_spec.tex file

824be2e

feat: add means to build the RLP_UTILS pdf in Makefile

7c6b50c

feat: add packages to make it compile

ec0988e

ras: labels & renaming

f308877

ras: formatting

db191ae

ras: counter constant -> counter-constant

0200030

ras: \resultColumn{k} macro

f7e0b48

feat: further uncommenting _all_spec.tex

6aca1ba

fix: using \resultColumn{k} + adding underscores (_)

f95b87f

delete unused old file

196e6db

Signed-off-by: F Bojarski <[email protected]>

This comment was marked as outdated.

Sign in to view

letypequividelespoubelles commented Jul 31, 2025

View reviewed changes

rlp_patterns_v2/_all_rlp_patterns_v2.tex Outdated

Copy link

Contributor Author

letypequividelespoubelles Jul 31, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

delete all the pattern_v2 folder now I guess ?

letypequividelespoubelles commented Jul 31, 2025

View reviewed changes

rlp_txn/.DS_Store Outdated

Copy link

Contributor Author

letypequividelespoubelles Jul 31, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

?

letypequividelespoubelles commented Jul 31, 2025

View reviewed changes

rlp_txn_v2/phase_constraints/to_address.tex Outdated

Copy link

Contributor Author

letypequividelespoubelles Jul 31, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

setting phase_end is missing

letypequividelespoubelles commented Jul 31, 2025

View reviewed changes

OlivierBBB added 4 commits July 31, 2025 16:24

small refactoring

e244145

some refactoring

2c170b9

fix: added missing binary constraints

bcc20b0

feat: missing constancy condition for LT and LX

fd04116

This comment was marked as outdated.

Sign in to view

letypequividelespoubelles commented Aug 1, 2025

View reviewed changes

OlivierBBB added 3 commits August 1, 2025 17:21

review with FB

1c6d505

feat: remove most LT / LX setting constraints

4906d79

extraction of RLP_L?_BYTESIZE constraints + finalization + ras

b02d5bb

cursor bot reviewed Aug 1, 2025

View reviewed changes

OlivierBBB added 2 commits August 1, 2025 22:20

fix: using compound constraints in the RLP_PREFIX phase

9f834ae

renaming

71fb1dd

cursor bot reviewed Aug 1, 2025

View reviewed changes

	\utt{1st argument lo:} & \cmdWCP@var@argOneTwo \\
	\utt{2nd argument hi:} & \cmdWCP@var@argOneOne \\
	\utt{2nd argument lo:} & \cmdWCP@var@argOneZero \\

RLP redesign #137

Are you sure you want to change the base?

RLP redesign #137

Uh oh!

Conversation

letypequividelespoubelles commented Mar 28, 2025

Uh oh!

cla-assistant bot commented Apr 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

This comment was marked as outdated.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment was marked as outdated.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cursor bot left a comment

Choose a reason for hiding this comment

Bug: Macro Argument Labeling Mismatch

Uh oh!

cursor bot Aug 1, 2025

Choose a reason for hiding this comment

Bug: Macro Naming Inconsistency

Uh oh!

Uh oh!

`RLP` redesign #137

`RLP` redesign #137

cla-assistant bot commented Apr 1, 2025 •

edited

Loading