changelog_begin
Ledger API Specification: CreateUser and GetUser endpoints of UserManagementService now return the CreateUserResponse or GetUserResponse messages, whereas previously they were returning the User message).
changelog_end
specifying a non-existent logback file is apparently not an error and
just silently produces garbage.
Probably should add a test that checks all logback files for existent
but that seems somewhat convoluted to setup so I’p prefer to skip that
for now.
changelog_begin
changelog_end
* Migrate Haskell tests to Sandbox on X
The diff here looks extremely confusing. sandbox-classic is now
sandbox on x not sandbox classic so this isn’t a typo. It is really
moving to sandbox on x.
changelog_begin
changelog_end
* .
changelog_begin
changelog_end
* Update kvutils Config and SandboxConfig
* Remove `eager-package-loading` and `scenario` from SandboxConfig
* Add `profile-dir` and `stackTraces` to kvutils Config
* Configurable authService in kvutils
changelog_begin
changelog_end
* Addressed Samir's review comments
* Add back Maximum supported deduplication duration enforcement of 1 hour
During refactoring of #9993, the "unsafe" compilation of the seedy command
has been moved out of the scope of `Engine.runSafely`, whose only purpose
was to catch exceptions thrown by the Speedy compiler.
This fixes the issue and renames the function `runSafely` to
`runCompilkerSafely` to make its purpose more obvious.
This is not a critical bug, as the command preprocessing should ensure
that (1) the command are properly typed and (2) all the dependencies
have been loaded.
CHANGELOG_BEGIN
CHANGELOG_END
Enriches #12766 with a chunked, lazy internal approach to
aggregating pages from the participant server, rather than
aggregating all users directly in memory before returning them.
CHANGELOG_BEGIN
- [JSON API] If ``/v1/users`` fails while retrieving users from the participant,
it will be reported as a warning, with a partial set of users, rather than
failing.
See `issue #12777 <https://github.com/digital-asset/daml/pull/12777>`__.
CHANGELOG_END
I suppose whoever fixed this forgot to update these instructions, but as of `2.0.0-snapshot.20220204.9165.0.225c58f4` I observed well-formed "Bob" and "Alice" values in the messaging dropdowns.
from 130 to 125, reason the test is currently flaky (It fails around 20% of the time).
Surprisingly, #12678 seems to have increase the stack usage for the type checker.
CHANGELOG_BEGIN
CHANGELOG_END
* remove scenario from daml-studio Abort example
* remove scenario from daml-studio CreateAuthFailure example
* remove scenario from daml-studio ExerciseAuthFailure example
* remove scenario from daml-studio NotVisibleFailure example
* replace references to 'scenario' with 'script'
* update Iou copyright notice
* update daml-studio docs pictures
changelog_begin
changelog_end
If UserManagementStore retrieved fewer users than requested we can assume this is the last page and thus set the next_page_token to the empty string.
changelog_begin
changelog_end
Somehow left this in accidentally.
Perhaps surprisingly it is harmless since we call this before we
enable `set -e` and the command fails but it is extremely confusing.
changelog_begin
changelog_end
* SandboxNextFixture replaced by Sandbox-on-X based SandboxFixture
changelog_begin
changelog_end
* Some fixed tests
* No direct dependencies on //ledger/sandbox:sandbox and //ledger/sandbox:sandbox-scala-tests-lib
* Fix after rebase
* Rename SandboxFixture and add a missing dep
* Generate valid party names if hint is empty
* Smaller maxInboundMessageSize
* Added test for empty display name
* SandboxServer is a ResourceOwner
* Uses execution context passed as an input for resource management
* Fixes flaky FlywayMigrations issue with null Thread.currentThread.currentClassLoader
* SandboxServer simplification returns Port instead of ApiServer
* Dedicated PMAllocateWithoutDisplayName for non-Canton ledgers
* Created since Canton does not return empty display names
* Return all users from `GET /v1/users/`, paginating 1k at a time
The test runs overall for ~30 seconds, mostly due to the time required to
create 20k users.
changelog_begin
[HTTP-JSON API] Previously, a 10k limit was in place on the number of users returned by
`GET /v1/users/`. This limit has been removed and users are retrieved in chunks from the ledger
changelog_end
As of now, this builds the response in-memory, which is likely not what we want to
support long-term.
* Relax test condition due to interference with other test setups
Mostly to get a version with `info.json`, but I guess this could also be
used to get the ball rolling on that testing we want to do on Monday.
CHANGELOG_BEGIN
CHANGELOG_END
* operate tokenProvidesClaims on user management service output
* ignore everything but the applicationId claim for StandardJWTPayload
* let the client tester work with non-custom tokens
* test claims check on non-custom token
* add changelog
CHANGELOG_BEGIN
- [Auth Middleware] Supports standard auth tokens for participant user management.
See `issue #12563 <https://github.com/digital-asset/daml/pull/12563>`__.
CHANGELOG_END
* VersionClient supports getApiFeatures
This exposes the `features` part of the version service's
`GetLedgerApiVersionResponse` protobuf.
Map the protobuf to user-friendly subclasses of `Feature`,
defined in `c.d.ledger.api.domain`.
VersionClient does not expose experimental features, as they
are meant for internal testing only.
CHANGELOG_BEGIN
CHANGELOG_END
Co-authored-by: Simon Meier <meiersi-da@users.noreply.github.com>
Co-authored-by: Simon Meier <meiersi-da@users.noreply.github.com>
The original solution turned out to reveal a bug in Oracle DB:
the batch insert with IGNORE_ROW_ON_DUPKEY_INDEX with rows containing
significant amount of varying payload can run on Oracle JDBC
driver exceptions.
* Fixes with switching to single insert instead of multi-insert for package upload
for Oracle
* Enhances unit test to uncover the aforementioned issue
changelog_begin
changelog_end
* Use fetchAndPersistBracket instead of fetchAndPersist
* add changelog
CHANGELOG_BEGIN
- [JSON API] Under rare conditions, a POST-style multi-template query backed by
database could have different template IDs appear to be queried at different
ledger times, if updated concurrently. This conditions is now checked and
accounted for. See `issue #12672
<https://github.com/digital-asset/daml/pull/12672>`__.
CHANGELOG_END
* add a searchQuery metric event for each fetch trial in sync search
* revert the contract ID and contract key fetches, and explain why
Co-authored-by: Stephen Compall <stephen.compall@daml.com>
The idempotent insert clauses are racy for Oracle and H2, this is
only used for inserting packages idempotently.
* Fixes H2 by manually retrying maximum 10 times the execution in question
* Fixes Oracle by applying a different approach not requireing retrying:
using IGNORE_ROW_ON_DUPKEY_INDEX
* Adding unit tests which ensures racy idempotent insert is not an issue
(this was failing originally)
* Adding withConnections helper to StorageBackendSpec
changelog_begin
changelog_end
Because there's no reason not to. The only obstacle to automating the
normal release process is that we need an explicit manual validation
step for our audit log when creating a release, but split-releases are
created in the Assemblty repo, so we can have the audit log over there.
The diff is going to be very messy because there's a lot of stuff moving
around. The only thing that is not just moving a job to a separate file
is the `ci/daily-snapshot.yml` file, which is 100% new and is meant as
a new Azure Pipelines entrypoint (which I will create after this gets
approved). I have made a reasonable effort to create individual commits
that simplify reviewing, but I expect it's still going to be kind of a
mess. I'm open to opening separate PRs to ~bump my stats~ move one job
at a time if that makes reviewing (and testing) easier.
CHANGELOG_BEGIN
CHANGELOG_END
