* Clean up how we check DAML-LF feature support in damlc
Currently, we have a predicate that tells us whether a DAML-LF version
supports a specific feature. We replace this by a data type which captures
the name of a feature and the minimal DAML-LF version which supports it.
* Rename featureDisjunctionChoices into featureFlexibleControllers
* Error only once
Co-Authored-By: hurryabit <martin.huschenbett@posteo.me>
There were two issues in `daml studio`:
1. We need to use `shell` instead of `proc` to launch VSCode.
2. Creating symlinks requires admin privileges on Windows so instead
we just copy the directory.
Note that `daml studio` will not install or upgrade the extension if
it is already installed (this fact is unchanged by this PR). We should
change this to upgrade to newer versions by default (and add an option
to not do this) but I’ll leave that for a separate PR.
The ci/release.sh fails if the BUILD_SOURCEBRANCHNAME environment
variable is not set. Although this variable is normally set by the
CI system, it is sometimes useful to run the script manually and
simply adding an 'invalid' default to the check of the env variable
means that the script still works if the variable is unbound.
* language: change default output dir of `package` command
This changes the default output path of the `package` and `build`
command from the project root to the `dist` directory.
* fixed daml-assistant integration tests
* Improve comment on `SubmitRequest`
Point out that `commands` corresponds to top-level actions and that order is preserved.
* Correct correspondence between commands and events
* Fix markup from MD to RST
* Update command_submission_service.proto
Since Windows builds aren’t sandboxed properly (see
https://github.com/bazelbuild/bazel/issues/5136), concurrent
compilations interact with the same tmp_db directory which leads to
flaky build failures.
* Improve bazel to generate Haskell for more protos. Get the generated Haskell to compile.
* adapt to @haskell_proto3__suite//:compile-proto-file
* haskell_proto3__suite now works for proto files contains map<key,value>
* remove unnecessary bash loop
When echo is enabled (which is the default), the IDE can get really
confused since it tries to interpret the commands as LSP messages
which obviously fails.
* Move to using proto3-wire from upstream
* Move to upstream proto3-suite, with some custom patches in my fork
* Delete the BUILD.bazel for hte proto3 stuff, not used and the test was failing
* Delete the old proto3-wire and proto3-suite forks
* Delete proto3-wire
* Prettify BUILD.bazel files, sort the deps
* Remove some special cases from the license checker
* Delete unused Nix files from grpc-haskell
* Switch to upstream proto3-suite
* Make old-time work on Windows
* Formatting
* Patch rules_haskell to use a response file for -optP to avoid overflowing argument size limits on Windows
* Update 3rdparty/haskell/BUILD.old-time
Co-Authored-By: neil-da <35463327+neil-da@users.noreply.github.com>
* Update the comments in old-time
* Use the revised location of proto3-suite
* GenTransaction.WithTxValue alias
* undo parts of parent dealing explicitly with serialization
* use GenNode.WithTxValue more often
* use Node*.WithTxValue more often
* introduced a SteppingMode abstraction for Dispatcher
* changing InMemoryLedger to use range queries
* SQL Ledger uses range queries
* documenting + fixing off by one semantics of Dispatcher
* adapting DispatcherIT
* added some docs
* reverting off-by-one offset semantics required on the ledger-api
* formatting
* introducing a case class ParsedEntry for sql result parsing
* testing Dispatcher with slow range queries as well
* adding missing ORDER BY ASC
* removing commented out code
Currently, the structural record returned by DAML-LF's `fetchByKey` is
unpacked in the compiler. This is not very nice as it requires too much
code for my taste.
This PR shifts the unpacking into DAML land by means of the recently
introduced `unpackPair` function.
Add a check to the Haskell implementation of the DAML-LF type checker to
make sure that the current DAML-LF version supports contract keys when they
are encountered.
If we don't inline let bindings of the form `x = this.f`, the rewriting of
the maintainer expression in terms of the key can fail when it shouldn't.
Usually, these bindings do not exist in the surface language but are produced
by the compiler. I think the performance penalty we might pay for this is
negligible.
This fixes#384.
* Fix termination of scenario service on Windows
The lack of a proper Windows IO manager resulted in us being unable to
kill the conduits reading the output of the scenario service so `damlc
test` and `damlc ide` blocked forever. This PR fixes the problem by
shutting down the scenario service (by closing its stdin) before
killing the conduits .
* Use fail instead of error
* Add debugging output
* Remove debug output
* Bump timeout of perf test
* Remove a comment regarding the consuming method of the Choice class
The comment has been out of date since the new "dictionary sanitization" was
implemented.
* Change default implementation of consuming method of type class Choice
* daml-on-x: Drop StateInit and use AtomicReference
- Drop the StateInit update message and introduce "getLedgerInitialConditions".
- Switch to AtomicReference from SyncVar as this is non-blocking and hence safer to use from Akka.
* ledger/participant-state: Rename recordTimeEpoch to initialRecordTime
and use the time from timeProvider at the ledger object initialization
time as the initial record time.
* fmt
* Fix use of repo as external workspace. Fix compilation issue in IndexState.scala
* Refix the use of this repo as external bazel workspace
* Fix compilation after rebase
* refmt
Apart from the fact that START was missing the windows title argument
it launches a new terminal window and then exits immediately (the
terminal windows is closed immediately as well) so it seems like the
wrong thing to use. Just calling the executable directly seems to work
fine both in cmd.exe and in powershell on my Windows 10 VM so
hopefully this is reasonably robust.
* seal engine.Error
* don't reinvent flatMap in CommandExecutorImpl
* add DivulgenceVisibility sandbox test case
- suggested by @oggy-; thanks
* add global implicit disclosure step to sandbox ACS transaction acceptance
* reorder ActiveContractsManager and include an implicitlyDisclose action
* add submitter argument to SubmissionHandle#lookupActiveContract
- the receiver must filter its response based on witnesses thereof
* propagate implicit disclosure in more places
* it's a bunch of functions
* proper error from the no-divulgence test
* rename ActiveContracts and ActiveContractsSteps to ...InMemory and ActiveContracts
* we need another table
* rename ActiveContract.witnesses to disclosedTo
* don't exclude other disclosures in ActiveContractsManager
* add local implicit disclosure when adding transaction to sandbox acs
* extra argument in ledger-api-server-example
* too late to blind?
* adapt to 10fcab8#339
* scalafmt
* non-total map get
* disable postgres test in SandboxSemanticTestsLfRunner until implicit disclosure added
* un-flatten witnesses and divulgences from "disclosedTo"; fixes DivulgenceIT
* comment on filtering of lookupActiveContract
- suggested by @francesco-da; thanks
* semantic errors from parent
* scenario to test divulgence success
* submit line 6 as owner; disable divulgence test under SQL
* comment what else to reenable
* Extract codegen-common module, #166
* Scala Codegen Main using the same option parser as Java Codegen, #166
There is one important difference, Scala Codegen does not allow mapping
dars to different package names, all dars have to be mapped to the same
package name.
Replace Scala Codegen println's with scala logging, respecting the
configured codegen verbosity
* Fix bazel formatting
* Update the release dry run script
* Releasing codegen-common
* Improving Scala Codegen error reporting (code review)
* Addressing codereview comments
* Make it explicit that we skip not supported option
