Derive P2TR input weight from witness data #1200

Mshehu5 · 2025-11-21T19:50:11Z

Replaces: #880 as contibutor is currently busy

I aslo followed comments from previous PR: #880 (comment)

closes: #788

Previously, P2TR weight estimation assumed key-path spends only, which was incorrect for script-path spends and non-default sighash.
This change derives weight from actual witness data when available, matching P2WSH handling.
When witness data is unavailable, applications must provide expected_weight manually, ensuring accurate fee calculation.

Also Adding new helpers for P2tr so key-spend inputs get the default weight and script-path inputs take caller-supplied weight.

Pull Request Checklist

Please confirm the following before requesting review:

I have disclosed my use of
AI
in the body of this PR.
I have read CONTRIBUTING.md and rebased my branch to produce hygienic commits.

coveralls · 2025-11-21T19:52:59Z

Pull Request Test Coverage Report for Build 20275523512

Details

85 of 86 (98.84%) changed or added relevant lines in 3 files are covered.
1 unchanged line in 1 file lost coverage.
Overall coverage increased (+0.1%) to 83.402%

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
payjoin/src/core/receive/mod.rs	76	77	98.7%

Files with Coverage Reduction	New Missed Lines	%
payjoin/src/core/psbt/mod.rs	1	67.38%

Totals
Change from base Build 20245341060:	0.1%
Covered Lines:	9728
Relevant Lines:	11664

💛 - Coveralls

arminsabouri

Approach looks correct. Had a suggestion and question for breaking up p2tr key spend and script spend input pair constructors. Did not review the tests. Will circle back once this is out of draft. Thanks

payjoin/src/core/receive/mod.rs

payjoin/src/core/psbt/mod.rs

arminsabouri · 2025-11-24T15:59:56Z

payjoin/src/core/psbt/mod.rs

+                match witness.len() {
+                    0 => Err(InputWeightError::NotSupported),
+                    1 => {
+                        let signature =
+                            witness.iter().next().ok_or(InputWeightError::NotSupported)?;
+                        if !(signature.len() == SCHNORR_SIGNATURE_SIZE
+                            || signature.len() == SCHNORR_SIGNATURE_SIZE + 1)
+                        {
+                            return Err(InputWeightError::NotSupported);
+                        }
+                        Ok(InputWeightPrediction::new(0, vec![signature.len()]))
+                    }
+                    _ => Ok(InputWeightPrediction::new(
+                        0,
+                        witness.iter().map(|el| el.len()).collect::<Vec<_>>(),


Is this match on length necessary? Once we have a non-empty witness can we not just calculate expected satisfaction weight?

InputWeightPrediction::new(0, witness.iter().map(|el| el.len()).collect::<Vec<_>>())

payjoin/src/core/receive/mod.rs

arminsabouri · 2025-12-04T16:55:06Z

@Mshehu5 I'll do another review once this is out of draft

arminsabouri

Some trivial nits in the tests and suggested clean up of duplicated code. The main issue remaining is a possible regression in psbt/mod.rs

payjoin/src/core/receive/common/mod.rs

arminsabouri · 2025-12-11T17:08:36Z

payjoin/src/core/psbt/mod.rs

                    Ok(iwp)
                },
-            P2tr => Ok(InputWeightPrediction::P2TR_KEY_DEFAULT_SIGHASH),
+            P2tr => {


I believe there is a regression here. For key spends we know the weight prediction even though we do not have the witness. We may need to internally distinguish between keyspends and scriptpath using the new ctors introduced by this PR.

Thank you for the review and for sharing the links to help clarify the necessary changes.

I’ve pushed updates that set the weight to key spend when the witness is not available.
Script path weights can still be set explicitly using the constructor or derived from witness

I believe this will have the same issue as the original issue. I.e if the witness is not known and its a taproot output we assume the key spend path is being used -- even though that may not be the case; which is what prompted the new ctors and user provided weight.

Perphaps the best way to solve this is to set expected weight to the key spend weight in the key spend ctor. And return a error here as we cannot accurately determine the input weight prediction for a taproot output that has no witness. What do you think?

payjoin/src/core/receive/mod.rs

arminsabouri · 2025-12-11T17:22:02Z

payjoin/tests/integration.rs

+        let expect_weight = psbtin
+            .witness_utxo
+            .as_ref()
+            .filter(|txo| txo.script_pubkey.is_p2tr())
+            .map(|_| Weight::from_wu(P2TR_INPUT_WEIGHT));
+        InputPair::new(txin, psbtin, expect_weight).expect("Input pair should be valid")


Can you explain this change?

This change is no longer needed as key spend is now automatically derived when no witness is present unlike before

payjoin/src/core/psbt/mod.rs

arminsabouri · 2025-12-15T16:37:56Z

payjoin/src/core/psbt/mod.rs

                    Ok(iwp)
                },
-            P2tr => Ok(InputWeightPrediction::P2TR_KEY_DEFAULT_SIGHASH),
+            P2tr => {


I believe this will have the same issue as the original issue. I.e if the witness is not known and its a taproot output we assume the key spend path is being used -- even though that may not be the case; which is what prompted the new ctors and user provided weight.

Perphaps the best way to solve this is to set expected weight to the key spend weight in the key spend ctor. And return a error here as we cannot accurately determine the input weight prediction for a taproot output that has no witness. What do you think?

arminsabouri

Cack
I had a few nits. Otherwise looks good 👍
Thanks

arminsabouri · 2025-12-16T14:01:26Z

payjoin/src/core/receive/mod.rs

+        let key_spend_weight =
+            InputWeightPrediction::P2TR_KEY_DEFAULT_SIGHASH.weight() + NON_WITNESS_INPUT_WEIGHT;


Nit: this could be a const.

when trying to do this I get an error from rust analyzer:

const P2TR_KEY_SPEND_WEIGHT: Weight = InputWeightPrediction::P2TR_KEY_DEFAULT_SIGHASH.weight() + NON_WITNESS_INPUT_WEIGHT;

cannot call non-const operator in constants calls in constants are limited to constant functions, tuple structs and tuple variants

weight( ) is a non constant function

I meant the whole expressions can be evaluated and expressed as a const
const DEAFULT_SIGHASH_KEY_SPEND_INPUT_WEIGHT: Weight = ...,

payjoin/src/core/receive/mod.rs

Previously, P2TR weight estimation assumed key-path spends only, which was incorrect for script-path spends and non-default sighash. This change derives weight from actual witness data when available, matching P2WSH handling. When witness data is unavailable, applications must provide expected_weight manually, ensuring accurate fee calculation.

Use the new helpers so key-spend inputs get the default weight and script-path inputs take caller-supplied weight.

arminsabouri

Re-AcK a7dbe56

arminsabouri reviewed Nov 24, 2025

View reviewed changes

Mshehu5 force-pushed the P2TR_estimation branch from 7b85692 to 81eb642 Compare November 25, 2025 20:51

Mshehu5 force-pushed the P2TR_estimation branch from 81eb642 to 0d49a8e Compare December 9, 2025 21:51

Mshehu5 marked this pull request as ready for review December 9, 2025 22:00

arminsabouri requested changes Dec 11, 2025

View reviewed changes

Mshehu5 force-pushed the P2TR_estimation branch from 0d49a8e to 017feb4 Compare December 13, 2025 18:06

arminsabouri reviewed Dec 15, 2025

View reviewed changes

Mshehu5 force-pushed the P2TR_estimation branch 2 times, most recently from a4c1812 to ce2be55 Compare December 15, 2025 20:11

arminsabouri approved these changes Dec 16, 2025

View reviewed changes

Mshehu5 force-pushed the P2TR_estimation branch 2 times, most recently from 6368515 to ab2f027 Compare December 16, 2025 16:21

Adopt P2TR keyspend/scriptpath constructors

a7dbe56

Use the new helpers so key-spend inputs get the default weight and script-path inputs take caller-supplied weight.

Mshehu5 force-pushed the P2TR_estimation branch from ab2f027 to a7dbe56 Compare December 16, 2025 16:38

arminsabouri approved these changes Dec 16, 2025

View reviewed changes

arminsabouri merged commit 4c91534 into payjoin:master Dec 16, 2025
9 checks passed

arminsabouri mentioned this pull request Dec 16, 2025

Fix weight estimation for P2TR inputs #880

Closed

		let key_spend_weight =
		InputWeightPrediction::P2TR_KEY_DEFAULT_SIGHASH.weight() + NON_WITNESS_INPUT_WEIGHT;

Derive P2TR input weight from witness data #1200

Derive P2TR input weight from witness data #1200

Uh oh!

Conversation

Mshehu5 commented Nov 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

coveralls commented Nov 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pull Request Test Coverage Report for Build 20275523512

Details

💛 - Coveralls

Uh oh!

arminsabouri left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

arminsabouri Nov 24, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

arminsabouri commented Dec 4, 2025

Uh oh!

arminsabouri left a comment

Choose a reason for hiding this comment

Uh oh!

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!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

arminsabouri left a comment

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!

Uh oh!

arminsabouri left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Mshehu5 commented Nov 21, 2025 •

edited

Loading

coveralls commented Nov 21, 2025 •

edited

Loading

arminsabouri Nov 24, 2025 •

edited

Loading