Skip to content

chore: sync main with upstream splice main - #86

Merged
sadiq1971 merged 59 commits into
mainfrom
chore/sync-main-upstream-2026-10-21
Oct 22, 2026
Merged

sadiq1971 merged 59 commits into
mainfrom
chore/sync-main-upstream-2026-10-21

Conversation

@sadiq1971

@sadiq1971 sadiq1971 commented Oct 21, 2026 •

Copy link
Copy Markdown
Collaborator

Refs ChainSafe/canton-extending-mainnet#114

Summary: Merges the latest upstream canton-network/splice main (f2be6819a, 55 commits since our last sync) into main, keeping the dedicated-synchronizer work intact. Upstream has no Daml changes this time, so no package bump; it brings Canton 3.7 and LATEST_RELEASE 0.9.3. The sync operator follows the SV's move to protocol version 37, and LocalNet's Grafana moves to 13.2.2 for upstream's V2 dashboards.

stephencompall-DA and others added 30 commits October 13, 2026 09:43
…-network#7731)

Add an IN-clause batch version of listTransactionViews that fails if any
verdict exceeds the per-verdict view limit, and use it in /v0/events
instead of one query per verdict.

Assisted-by: Claude:claude-5.5-opus
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
Signed-off-by: Itai Segall <itai.segall@digitalasset.com>
…HTTP layer (canton-network#7716)

* [ci] BFT calls return typed outcomes, and only the HTTP layer maps a failed BFT call to a status code

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Tests: RetryProvider retries every failed BFT call except never available, and the HTTP layer answers each with its status code

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] A BFT call that is never available answers 502 at the HTTP layer instead of 404; scaladoc for the BftCallFailed cases

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

---------

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>
…es (canton-network#7726)

* [ci] Bulk storage updates segments are plain record times; the segment dump finds its start migration with getLowestMigrationForRecordTime

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Bulk storage ACS snapshot cursors are plain record times; the DB writer finds the snapshot migration with getLowestMigrationForRecordTime

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Bulk storage KV keys get a _v2 suffix so the record-time cursors start from genesis instead of failing to decode

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] debugForceStartFromGenesis also deletes the bulk storage KV keys from before the _v2 rename

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] ACS snapshot writer walks the snapshot rows with their migration, and looks the migration up only when resuming from a persisted record time

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] TODO(canton-network#7745) to remove the cleanup of the pre-_v2 bulk storage keys once deployed on DevNet

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

---------

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>
Taken from DACH-NY/canton#36481

[ci]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
…anton-network#7646)

* create command priority util

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* use command priority util in 4 functions

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* use command priority util in AutoAcceptTransferOffersTrigger

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* add Party check helper; add command priority to createTransferPreapprovalProposal

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* change command priority util name

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* change command priority util name again

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* comment delete

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* adding integration tests for acceptTransferOffer and acceptTokenStandardTransfer

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* adding integration tests to one file only; adding control low priority test

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* dummy commit to add [ci]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* add AutoAcceptTransferOffersTrigger test [ci]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* change test to normal instead of time based [ci]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* add WalletNoDevNetIntegrationTest to the test class list [ci]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* pause validator top-up and faucet triggers in WalletNoDevNetIntegrationTest [ci]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* pause top-up trigger for non sv validator and faucet trigger for all validators in WalletNoDevNetIntegrationTest [ci]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

---------

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>
…te segments (canton-network#7725)

* [ci] ACS snapshot trigger starts when the history readiness allows it, the same condition as before until a snapshot seed exists

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Bulk storage writers wait on a BulkStorageWritersGate: a Scan whose history starts after genesis writes and commits its own update segments from its first own segment start

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Bulk storage progress per kind: commit pipelines and listings ask peers for update or snapshot progress, so a Scan waiting for its snapshot seed still counts for update segments

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] New-network integration test: sv2 commits its own update segments equal to sv1's and writes none across its first own segment start

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Bulk storage and the ACS snapshot trigger share one history readiness; the snapshot trigger's wait reason matches mayStart

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] BulkStorageWritersGate reads as a match on the history start with named outcomes, and uses if on booleans

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [ci] Bulk storage builds the writers gate after the backfilling progress it reads, so the progress is not created with uninitialized cursors

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

---------

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>
* add my user to cluster-lock-users

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

* Trigger Build [static]

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>

---------

Signed-off-by: Madalena Lourenco <madalena.lourenco@digitalasset.com>
Signed-off-by: Paweł Perek <pawel.perek@digitalasset.com>
…canton-network#7777)

fixes DACH-NY/cn-test-failures#10345

[ci]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
…ton-network#7774)

[ci]

Signed-off-by: Robert Autenrieth <robert.autenrieth@digitalasset.com>
…-network#7733)

getChildBatchHashes is capped at 10000 children, far above any
reasonable batchSize, and getLeafAllowances at the size of the batch's
party range. Both use HardLimit, so exceeding them logs a warning.

Assisted-by: Claude:claude-5.5-opus
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
…canton-network#7776)

* [static] Bump the splice-info nginx image to 1.31.6-alpine-slim, which ships pcre2 10.49 (CVE-2026-103111)

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

* [static] Bump the splice-info curl-jq updater image to 5.0.3, which ships pcre2 10.48

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>

---------

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>
…nton-network#7772)

---------

Signed-off-by: Raymond Roestenburg <raymond.roestenburg@digitalasset.com>
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
Co-authored-by: Stephen Compall <stephen.compall@digitalasset.com>
Signed-off-by: Itai Segall <itai.segall@digitalasset.com>
…twork#7787)

Signed-off-by: Paweł Perek <pawel.perek@digitalasset.com>
Signed-off-by: Mateusz Błażejewski <mateusz.blazejewski@digitalasset.com>
…ed (canton-network#7735)

Always creates the legacy Dataset in pulumi if BigQuery is enabled at
all. Turning off enableLegacyDatastream will still tear down the
Datastream, publication, &c.

Assisted-by: Claude:claude-5.5-opus
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
…anton-network#7757)

Where Canton fork only widened the access of a def or val so Splice
could call it, we restore the upstream access and instead provide an
implicit class `X access` (in object XAccess) in the member's package
under new sbt projects.

Many cleanups of CANTON_CODE_CHANGES.md are also included.

Assisted-by: Claude:claude-5.5-opus
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
canton-network#7758)

Independent add-on methods move from the canton fork to canton-fork (the
hard fork).

Assisted-by: Claude:claude-5.5-opus
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
Signed-off-by: Itai Segall <itai.segall@digitalasset.com>
Currently fails with when bumping splice in the internal repo
https://github.com/DACH-NY/canton-network-internal/actions/runs/38859975496/job/116850229104?pr=7539.

Tested locally that it fixes the internal repo.

Perhaps a bit debatable whether it makes sense to rerun this test on
internal at all but I'll leave that for another day.

[static]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
---------

Signed-off-by: Julien Tinguely <julien.tinguely@digitalasset.com>
Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
… SvFrontendIntegrationTest (canton-network#7810)

Signed-off-by: Paweł Perek <pawel.perek@digitalasset.com>
I finally got tired of commenting on yet another PR that this is the
wrong repo to modify docs.

I kept enough so that the autogenerated docs are still
tested. Maybe eventually we also want to kill that but for now keeping
it seems safer. Probably could strip it down further but I don't think
i care enough.

[static]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
…VALIDATOR_SIZE env var (canton-network#7802)

Signed-off-by: Mateusz Błażejewski <mateusz.blazejewski@digitalasset.com>
Replace the fork's TraceContextGrpc with upstream's, plus only the
additions upstreamed in DACH-NY/canton#36652:
addTraceContextToCallOptions, inferCallerTraceContext,
inferClientRequestTraceContext, and a client interceptor that generates
a trace id when the caller supplies none.

This drops TraceContextCallOptionKey, the TraceContextThreadLocalKey
rename and the unused inferServerRequestTraceContext.
ApiClientRequestLogger now uses TraceContextOptionsKey.

Assisted-by: Claude:claude-5.5-opus
Signed-off-by: Stephen Compall <stephen.compall@digitalasset.com>
fixes canton-network#7340

---------

Signed-off-by: Vladislav Lapin <51929896+loglapa@users.noreply.github.com>
Co-authored-by: moritzkiefer-da <45630097+moritzkiefer-da@users.noreply.github.com>
Main part of canton-network#7775

Signed-off-by: Martin Florian <martin.florian@digitalasset.com>
pawelperek-da and others added 28 commits October 19, 2026 11:56
Signed-off-by: Paweł Perek <pawel.perek@digitalasset.com>
…etwork#7786)

Signed-off-by: DA Automation <splice-maintainers@digitalasset.com>
Co-authored-by: DA Automation <splice-maintainers@digitalasset.com>
[static]

Signed-off-by: Mateusz Błażejewski <mateusz.blazejewski@digitalasset.com>
canton-network#7823)

* Support PartyToParticipant with embedded keys in ValidatorRightTrigger

fixes canton-network#7805

Added a config flag to restore the prior behavior. While this
_shouldn't_ be necessary as the new behavior is definitely correct,
I'm still a bit worried it may result in some wallets blowing past the
200 parties limit.
Part of DACH-NY/canton-network-internal#7506

The internal part: DACH-NY/canton-network-internal#7515

---------

Signed-off-by: krzysztofczyz-da <krzysztof.czyz@digitalasset.com>
Co-authored-by: Martin Florian <martin.florian@digitalasset.com>
…suite of traffic (canton-network#7834)

Signed-off-by: Paweł Perek <pawel.perek@digitalasset.com>
…canton-network#7840)

---------

Signed-off-by: Oriol Muñoz <oriol.munoz@digitalasset.com>
…anton-network#7844)

Signed-off-by: Julien Tinguely <julien.tinguely@digitalasset.com>
Signed-off-by: Julien Tinguely <julien.tinguely@digitalasset.com>
Part of DACH-NY/canton-network-internal#4570

Needed to make sure that we can push metrics like we used to.

Current setup:

the publish_test_result_metrics job in CCI pushes metrics to http://prometheus-pushgateway.observability:9091/metrics/job/<...> (this is internal to the network inside of the kubernetes cluster)
we also do a get on http://prometheus-pushgateway.observability:9091/metrics in the same pipe
So the setup is: CCI runner -> push GW (all inside of k8s)
GHA changes this to:

prometheus-pushgateway.observability:9091 -> pushgateway.${CLUSTER_HOSTNAME}
GHA runner in Azure -> the same push GW
This is why we need the GW exposed temporarily .

This is a temporary setup that will just unblock the migration. Further work will be needed in the future to swap to a different metric sink.

---------

Signed-off-by: krzysztofczyz-da <krzysztof.czyz@digitalasset.com>
…ng issues (canton-network#7842)

[static]

Signed-off-by: Mateusz Błażejewski <mateusz.blazejewski@digitalasset.com>
…lti-validator project (canton-network#7830)

[static]

Signed-off-by: Mateusz Błażejewski <mateusz.blazejewski@digitalasset.com>
See DACH-NY/canton-network-internal#7591

[ci]

Signed-off-by: Martin Florian <martin.florian@digitalasset.com>
Surely mainnet will be on 0.9 anytime and I can ignore that. Seems
like we'll need that more often so I just removed the todo.

fixes canton-network#7248

[static]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
[static]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
…rationTest (canton-network#7856)

Signed-off-by: Paweł Perek <pawel.perek@digitalasset.com>
Signed-off-by: Julien Tinguely <julien.tinguely@digitalasset.com>
To speed up the lsus.

Will upstream to Canton after the cilr lsu works.

[static]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Signed-off-by: krzysztofczyz-da <krzysztof.czyz@digitalasset.com>
…rk#7877)

[static]

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
[static]

We already set it to true on all clusters afaict except for the lsu
test chich is changed in
DACH-NY/canton-network-internal#7576

Should also be good to just fully remove the option but I feel a bit
better to first just default it to true and do that in a separate PR.

Signed-off-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Co-authored-by: moritz.kiefer@digitalasset.com <moritz.kiefer@purelyfunctional.org>
Signed-off-by: Julien Tinguely <julien.tinguely@digitalasset.com>
[static]

Signed-off-by: Jagath Weerasinghe <jagath.weerasinghe@digitalasset.com>
Merges upstream main at f2be681 (0.10.0 dev, latest release 0.9.3, Canton
3.7.0 snapshot) into main.

Conflict resolution:
- Member traffic stores: upstream binds members in SQL through
  StoreMemberId. Our shared MemberTrafficQueries and the sync operator
  store and table take it too, and the SV triggers and our store tests
  wrap the member where upstream's do. The stored value is unchanged.
- Wallet: upstream's highPriorityIncreaseBalanceForTopup wraps our
  hasSufficientFundsForTopupOnGlobalSync, so the new priority call sites
  use the global synchronizer's top-up config like the existing ones.
- Validator: keep our synchronizer config in place of the global alias
  argument and add upstream's embedded party signing keys flag.
- Sync operator: import upstream's StaticSynchronizerParameters access
  class for copy, as LocalSynchronizerNode does.
- build.sbt: drop the CometBFT test group upstream removed; our sync
  operator upgrade group stays ahead of the LSU groups.
- start-canton.sh and splice-localnet-compose.sh: keep both sides'
  additions.
- Docs: take upstream's deletion of the validator API page.
- The sync operator LSU test comment now names protocol version 36,
  which bootstrap-canton.sc uses after this merge.

Signed-off-by: sadiq1971 <sadiqurr8@gmail.com>
…ync operator [ci]

Upstream moved the SV synchronizer node's default protocol version to 37
and applies its configured synchronizer limits from 37 on. The sync
operator's node copies that code, so both move with it. Moving only the
gate would drop configured limits at the default.

Signed-off-by: sadiq1971 <sadiqurr8@gmail.com>
Upstream converted the dashboards to the V2 resource format, which
Grafana 12.4.3 rejects, so LocalNet -G loaded none of them. 13.2.2 is
the version upstream's cluster runs; it loads all 65 dashboards and the
9 alert rules.

Signed-off-by: sadiq1971 <sadiqurr8@gmail.com>
…anton 3.7 [ci]

Canton 3.7 added an identityTransactions parameter to
bootstrap.synchronizer. The mediatorsToSequencers overload this script
uses has no defaults, so the script no longer compiled and the
synchronizer never bootstrapped. None is the default of the other
overload: Canton generates each node's onboarding transactions.

Signed-off-by: sadiq1971 <sadiqurr8@gmail.com>

@isegall-da isegall-da left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Thanks

@sadiq1971
sadiq1971 merged commit f683244 into main Oct 22, 2026
66 of 67 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.