feat(ICRC-Ledger): DEFI-1687: implement ICRC-107 fee collector by maciejdfinity · Pull Request #8217 · dfinity/ic

maciejdfinity · 2026-01-05T15:31:01Z

With this change the ICRC ledger now implements the ICRC-107 fee collector standard as described in https://github.com/dfinity/ICRC/blob/main/ICRCs/ICRC-107/ICRC-107.md. The legacy fee collector is not longer supported by the ledger. This results in the following notable changes:

If a fee collector is present, approval fees are also collected.
If an existing ledger has a legacy fee collector configured, the legacy fee collector will be converted to the ICRC-107 fee collector during the upgrade to this or later version.

The fee collector can be configured using the existing fee_collector_account init argument and change_fee_collector upgrade argument.

Since the ICRC-107 fee collector change is not backwards compatible, the ledger upgraded to this version cannot be downgraded to an earlier version.

maciejdfinity · 2026-02-02T14:21:05Z

rs/ethereum/cketh/minter/tests/ckerc20.rs

                        withdrawal_id: cketh_block_index.clone(),
                        burn_in_block: ckerc20_block_index.clone(),
-                        reimbursed_in_block: Nat::from(3_u8),
+                        reimbursed_in_block: Nat::from(4_u8),


If the ICRC ledger is installed with a 107 fee collector, the first block is the fee collector setting block and therefore all block indexes that are tested need to be increased by 1.

maciejdfinity · 2026-02-02T14:22:12Z

rs/tests/cross_chain/ic_xc_ledger_suite_orchestrator_test.rs

            index_status,
            ic_icrc1_index_ng::Status {
-                num_blocks_synced: Nat::from(0_u8)
+                num_blocks_synced: Nat::from(1_u8)


If the ICRC ledger is installed with a 107 fee collector, the first block is the fee collector setting block and therefore even a ledger with no transactions will contain one block.

This reverts commit fe7d54f.

maciejdfinity · 2026-02-04T12:44:36Z

rs/ledger_suite/icrc1/ledger/canbench_results/canbench_u256.yml

      calls: 1
-      instructions: 54416296904
-      heap_increase: 264
+      instructions: 54937060262


u64 diffs:
Only significant changes:

status name calls ins ins Δ% HI HI Δ% SMI SMI Δ%

+/- bench_icrc1_transfers::icrc2_approve 1 19.65B +6.57% 24 -4.00% 128 0.00%

- bench_icrc1_transfers::icrc3_get_blocks 1 7.84M -7.01% 0 0.00% 0 0.00%

u256 diffs:
Only significant changes:

status name calls ins ins Δ% HI HI Δ% SMI SMI Δ%

+/- bench_icrc1_transfers::icrc2_approve 1 20.55B +6.23% 25 -13.79% 128 0.00%

+ bench_icrc1_transfers::icrc2_transfer_from 1 21.03B -1.86% 4 +33.33% 0 0.00%

+/- bench_icrc1_transfers::icrc1_transfer 1 12.64B -2.20% 34 +6.25% 0 0.00%

- bench_icrc1_transfers::icrc3_get_blocks 1 8.22M -7.89% 0 0.00% 0 0.00%

The u64 diffs make some sense, icrc2_approve now also collects fees, so it uses a bit more instructions. The created blocks don't contain the fee collector or fee collector block index, which results in less heap usage for icrc2_approve and less instructions for decoding the blocks in case of icrc3_get_blocks.

For u256, the instruction diffs are similar for icrc2_approve and icrc3_get_blocks. We also see improvements for icrc1_transfer and icrc2_transfer_from (in case of u64 they were significant but below threshold). The latter might be caused by the fact that the fee collector is stored directly as Option<Account> and previously it was stored in Option<FeeCollector<Account>>. The heap diffs are less consistent, though.

mbjorkqvist

Thanks @maciejdfinity! I have a few comments and questions, but overall this looks very good!

mbjorkqvist · 2026-02-05T11:43:19Z

MODULE.bazel

        "ck_btc_ledger": "ic-icrc1-ledger.wasm.gz",
        "ck_btc_ledger_v1": "ic-icrc1-ledger.wasm.gz",
        "ck_btc_ledger_v3": "ic-icrc1-ledger.wasm.gz",
+        "ck_btc_ledger_v5": "ic-icrc1-ledger.wasm.gz",


Would it make sense to rename this to something more descriptive, and/or perhaps add a comment somewhere (I think comments don't work in mainnet-canister-revisions.json, but here would be ok?)? Especially since the v5 doesn't correspond to the LEDGER_VERSION, so it can be confusing.

Good point! Maybe we could change it to correspond to the ledger versions, i.e.:
v1 -> no_mem_mgr
v3 -> v1
v5 -> v3
WDYT?

mbjorkqvist · 2026-02-05T12:10:32Z

mainnet-canister-revisions.json

+  "ck_btc_ledger_v5": {
+    "rev": "512cf412f33d430b79f42330518166d14fc6884e",
+    "sha256": "901bc548f901145bd15a1156487eed703705794ad6a23787eaa04b1c7bbdcf48"
+  },


This points to ledger-suite-icrc-2025-04-14. Is there a particular reason for this specific version, or could we use ledger-suite-icrc-2025-10-27, which is the minimum version we point to from the currently most recent version ledger-suite-icrc-2026-02-02?

No reason, I just used the same version as current mainnet ledger :) Changed!

rs/ledger_suite/icrc1/index-ng/tests/tests.rs

mbjorkqvist · 2026-02-05T12:53:45Z

rs/ledger_suite/icrc1/index-ng/tests/tests.rs

Would it make sense to move all the fee collector-related tests to a separate file? rs/ledger_suite/icrc1/index-ng/tests/tests.rs is already over 2000 lines.

Also, would it make sense to add tests for the following (unless they already exist and I missed them):

Trying to set the fee collector account to the minting account, the anonymous account, and an invalid account (e.g., an account with an invalid principal). For the init, upgrade, and endpoint fee collector setting paths. IIUC, there's a check for the minting account in the init and upgrade paths, but not for the endpoint?

mbjorkqvist · 2026-02-05T13:00:05Z

rs/ledger_suite/tests/sm-tests/src/icrc_106.rs

+    // Downgrade to mainnet is currently not possible. The rest of the test can
+    // be uncommented again, once the PR that introduced this line is on mainnet.
+
+    // // Downgrade the ledger to the mainnet version that does not support ICRC-106
+    // env.upgrade_canister(canister_id, mainnet_ledger_wasm, encoded_empty_upgrade_args)
+    //     .expect("should successfully downgrade ledger canister");
+    // assert_index_not_set(&env, canister_id, false);
+
+    // // Upgrade to a ledger version that supports ICRC-106, but do not set the index principal
+    // let encoded_empty_upgrade_args = Encode!(&encode_empty_upgrade_args()).unwrap();
+    // env.upgrade_canister(canister_id, ledger_wasm, encoded_empty_upgrade_args)
+    //     .expect("should successfully upgrade ledger canister");
+    // assert_index_not_set(&env, canister_id, true);


Would it help to have a ticket in the ICRC-107 epic to remember to uncomment this, and the other snippet in rs/ledger_suite/icrc1/tests/upgrade_downgrade.rs, once this has been released and deployed to mainnet? Or will the tests fail anyway once that happens, and uncommenting cannot be forgotten?

Co-authored-by: Mathias Björkqvist <mathias.bjorkqvist@dfinity.org>

maciejdfinity added 17 commits December 15, 2025 13:04

add 107 schema

1b98a38

add block encoding decoding tests

71e4b6a

Merge branch 'master' into maciej-schema-feecol

c07af35

typo

5dda884

typo

561005a

smaller range

8973bab

build fix

76e35cb

Merge branch 'master' into maciej-schema-feecol

816719f

build fix

48a948f

changelog

5237a5d

Merge branch 'master' into maciej-schema-feecol

b4b2a1e

make tx required

1a8f5cb

handle endpoint initiated tx

5ac5de8

use constants

b59a42d

make tx kind a separate constant

df4cc15

Merge branch 'master' into maciej-schema-feecol

cea1742

add set fee collector arg and error

539ef8c

maciejdfinity changed the title ~~add set fee collector arg and error~~ feat(ICRC-Ledger): DEFI-1687: implement ICRC-107 fee collector Jan 5, 2026

github-actions bot added the feat label Jan 5, 2026

maciejdfinity added 3 commits January 5, 2026 15:36

add comments

ee265bb

add endpoint

cf62736

fix did, not completely

9a2483b

Base automatically changed from maciej-schema-feecol to master January 7, 2026 14:03

maciejdfinity added 7 commits January 8, 2026 11:52

Merge branch 'master' into maciej-ledger-feecol

a67b6e0

check if endpoint called by the controller

a24cfb7

init args always set 107 fee col

172d1a3

credit fc107 when applying the transaction

ad12027

Merge branch 'master' into maciej-ledger-feecol

806f889

handle fc107 and test

13feab0

Merge branch 'master' into maciej-ledger-feecol

5cb8a43

maciejdfinity added 2 commits February 2, 2026 13:07

fix xc suite orchestrator test

85ab3b8

update canbench

fe7d54f

maciejdfinity commented Feb 2, 2026

View reviewed changes

maciejdfinity added 6 commits February 2, 2026 16:12

remove fee collector from apply_transaction

a50e892

build fix

8516089

clippy

c0e9cec

Merge branch 'master' into maciej-ledger-feecol

6e1a69f

move fc tests to fee_collector.rs

f577ca9

Merge branch 'master' into maciej-ledger-feecol

d857983

maciejdfinity marked this pull request as ready for review February 3, 2026 13:29

maciejdfinity requested review from a team as code owners February 3, 2026 13:29

github-actions bot added @defi @idx labels Feb 3, 2026

maciejdfinity added 10 commits February 3, 2026 13:53

add defaults in ledger state

c565e44

remove default

f0a2e19

Merge branch 'master' into maciej-ledger-feecol

53ea05f

Merge branch 'master' into maciej-ledger-feecol

0cf417f

Merge branch 'master' into maciej-ledger-feecol

d45af12

create block from_transaction

c5f7638

use apply_transaction

4de9ca2

small refactor

81262b7

Revert "update canbench"

fa85ece

This reverts commit fe7d54f.

update canbench

4767419

maciejdfinity commented Feb 4, 2026

View reviewed changes

test icrc107_get_fee_collector

ce46106

mbjorkqvist reviewed Feb 5, 2026

View reviewed changes

maciejdfinity and others added 2 commits February 5, 2026 15:29

Update rs/ledger_suite/icrc1/index-ng/tests/tests.rs

fbc089e

Co-authored-by: Mathias Björkqvist <mathias.bjorkqvist@dfinity.org>

use later ledger release for v5

7e3d456

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat(ICRC-Ledger): DEFI-1687: implement ICRC-107 fee collector#8217

feat(ICRC-Ledger): DEFI-1687: implement ICRC-107 fee collector#8217
maciejdfinity wants to merge 103 commits intomasterfrom
maciej-ledger-feecol

maciejdfinity commented Jan 5, 2026 •

edited

Loading

Uh oh!

maciejdfinity Feb 2, 2026

Uh oh!

maciejdfinity Feb 2, 2026

Uh oh!

maciejdfinity Feb 4, 2026

Uh oh!

maciejdfinity Feb 4, 2026

Uh oh!

maciejdfinity Feb 4, 2026 •

edited

Loading

Uh oh!

mbjorkqvist left a comment

Uh oh!

mbjorkqvist Feb 5, 2026

Uh oh!

maciejdfinity Feb 5, 2026

Uh oh!

mbjorkqvist Feb 5, 2026

Uh oh!

maciejdfinity Feb 5, 2026

Uh oh!

Uh oh!

mbjorkqvist Feb 5, 2026

Uh oh!

mbjorkqvist Feb 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

status	name	calls	ins	ins Δ%	HI	HI Δ%	SMI	SMI Δ%
+/-	bench_icrc1_transfers::icrc2_approve	1	19.65B	+6.57%	24	-4.00%	128	0.00%
-	bench_icrc1_transfers::icrc3_get_blocks	1	7.84M	-7.01%	0	0.00%	0	0.00%

status	name	calls	ins	ins Δ%	HI	HI Δ%	SMI	SMI Δ%
+/-	bench_icrc1_transfers::icrc2_approve	1	20.55B	+6.23%	25	-13.79%	128	0.00%
+	bench_icrc1_transfers::icrc2_transfer_from	1	21.03B	-1.86%	4	+33.33%	0	0.00%
+/-	bench_icrc1_transfers::icrc1_transfer	1	12.64B	-2.20%	34	+6.25%	0	0.00%
-	bench_icrc1_transfers::icrc3_get_blocks	1	8.22M	-7.89%	0	0.00%	0	0.00%

Conversation

maciejdfinity commented Jan 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

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!

Choose a reason for hiding this comment

Uh oh!

maciejdfinity Feb 4, 2026 • 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!

mbjorkqvist 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!

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

maciejdfinity commented Jan 5, 2026 •

edited

Loading

maciejdfinity Feb 4, 2026 •

edited

Loading