First version of static verification tool.
The current state of the tool:
- Reads DAR files.
- Partially evaluates the code.
- Generates constraints for the field and choice to be verified.
- Passes the constraints to an SMT solver.
- Some basic tests.
This moves the code from the Shake testing module to a separate
module. This is required since the Shake test setup is not really
intended to test multiple packages. I’ve added a new test suite that
makes use of this and has an expected failure for the issue with
multiple packages. The issue is not actually fixed in this PR, I just
wanted to keep the refactoring separate since it is quite large and
noisy.
changelog_begin
changelog_end
This should hopefully fix the issues we have been seeing on CI. While
I’m not super keen on including non-upstreamable patches it seems
better than having CI be flaky.
changelog_begin
changelog_end
This PR gets yarn test running the Puppeteer end-to-end tests in the GSG integration test.
The first step of the test is adding the extra dependencies (Jest, Puppeteer and more). The GSG recommends a yarn add command, but this does not work against HEAD. This is because yarn add does not use resolutions in the parent package.json, and then complains about unknown versions 0.0.0 of the daml TS libaries. The solution here is to hack in the extra dependencies into the ui/package.json and then yarn install. This works, but it hard codes version numbers which we would need to maintain. I would like to be able to say version "latest", which is what yarn add would install.
The next step of the test is to copy the index.test.ts file and run yarn test in CI mode.
I've moved the GSG test to a new create-daml-app-tests target. It's marked "exclusive" in Bazel until we figure out how to avoid hardcoding port numbers. This is a bit tricky since the HTTP port is hardcoded in a couple of places in ui/package.json.
Finally, this PR gets the GSG testing docs to reference the code in the new templates/create-daml-app-test-resources folder.
changelog_begin
changelog_end
* Split up repl tests to make them faster
This PR splits up the tests into the tests for TLS and Auth and the
tests for the actual functionality.
The func tests use the repl as a library which allows them to be
significantly faster:
1. We only need to start the service process once.
2. We only need to initialize the package db once.
There is at least one other point that I did not address for now and
that is only loading the packages into the repl service once. While
loading them multiple times is a noop, it still has a performance
implication.
Sadly, this has turned out much more messy than I thought it would be
due to various issues with haskeline/repline/tasty/computers. The
details are in a comment in DA.Test.Repl.FuncTests.
changelog_begin
changelog_end
* Try to nuke cache on window selectively
* Enable cache again
* REPL parse import statements
* Factor parsing out of handleLine
* Pull binding and error handling into handleStmt
This is specific to the case of handling a statemt and does not overlap
with handling imports.
* Track module imports as ImportDecl
* Add new module imports
* Validate module imports
* Add REPL interaction test for import
* REPL document import declarations
CHANGELOG_BEGIN
- [DAML REPL - Experimental] You can now use import declarations at the
REPL prompt to bring additional modules into scope.
CHANGELOG_END
* ReplState strict fields
* Use semigroupoids Alt to simplify parseReplInput
Co-authored-by: Andreas Herrmann <andreas.herrmann@tweag.io>
* Use repline for DAML REPL
CHANGELOG_BEGIN
CHANGELOG_END
* REPL server print error messages
Co-authored-by: Andreas Herrmann <andreas.herrmann@tweag.io>
This PR bumps ghcide, haskell-lsp and haskell-lsp-types. There aren’t
really any important changes in ghcide itself but the haskell-lsp
update includes my fix for crashing completions.
One change in ghcide itself is that NormalizedFilePath got moved to
haskell-lsp. ghcide needs special treatment for empty file paths so we
use `toNormalizedFilePath'` from ghcide instead of
`toNormalizedFilePath` from `haskell-lsp`. I’ve added an hlint rule to
enforce this.
changelog_begin
changelog_end
This used to be a dependency of our zip library that you couldn’t
disable. However, in the meantime it has gotten a flag to disable
it and we don’t actually use it. Given that the upstream URL is dead
atm now seems like a good time to kill it.
changelog_begin
changelog_end
This is a bit ugly but after spending some time digging into the
issues in rules_haskell around data-files, this seems like the most
sensible option especially given that we also want to ship them in the
SDK which woud require additional work even if we do fix it in
rules_haskell.
fixes#4457
changelog_begin
changelog_end
* Switch to HashSet/HashMap for NormalizedUri/NormalizedFilePath
This matches the change to ghcide in
https://github.com/digital-asset/ghcide/pull/420. Now that we have
optimized Hashable instances it makes sense to use them as much as
possible.
changelog_begin
changelog_end
* debug windows crap
* Fix a bug in vr scenario notifications
* Revert "debug windows crap"
This reverts commit f58fdb92c1.
* attempt to fix windows
who are we kidding, this is not going to work
* Upgrade to haskell-lsp 0.20
This includes another ghcide bump but the only change there is the
haskell-lsp ugprade. haskell-lsp 0.20 includes some performance
improvements.
changelog_begin
changelog_end
* Bump parser-combinators
* Update rules_haskell
* Includes Bazel 2.0.0 support
* Some patches have been upstreamed
CHANGELOG_BEGIN
CHANGELOG_END
* External haskell_cabal_binary|library verbose = False
`haskell_cabal_binary|library` now supports setting `verbose = False`
which also avoids warnings coming from `Setup.hs`. As these are external
dependencies we are not going to address these warnings anyway. So, they
are just noise. This makes it unnecessary to pass `-w` or `-optF=-w`.
In case of build failure all errors and warnings will be displayed.
Co-authored-by: Andreas Herrmann <andreash87@gmx.ch>
The GitHub API is paginated (30 items by default). This creates two
problems:
1. At the moment, older versions silently drop from the docs website,
without us having made any explicit decision about it.
2. When we prepare a new version, it gets created as a pre-release
version. Our script filters that out, but that happens on our end so
we end up with 29 published versions and the list is different form the
existing one. If the prerelease then gets dropped, the oldest version
comes back.
It is possible that we will sometime decide we do not want to keep old
documentation around forever, but that should be an explicit decision.
This patch changes the logic to fetch the list of versions from GitHub
so that we always get all the published versions (barring race
conditions inherent to that kind of paginated API).
CHANGELOG_BEGIN
CHANGELOG_END
This fixes the ZIP modification times in all DARs to a specific
value (1980-01-01) whereas they used the current time before. This
both gives us the nice property that not only our DALF builds but also
DAR builds should be deterministic (and there is a test for this). I
have a suspicion that this could help significantly with build times
and avoid rerunning half of the Scala tests on a change to damlc that
should not change the DALFs.
changelog_begin
- [DAML Compiler] The modification times in a DAR are now fixed to a
given value which makes the output of ``daml build`` deterministic
in single-threaded mode (which is the default).
changelog_end
* Bump ghcide to include bugfix
This includes a fix for a bug in completions which we introduced in
the latest ghcide bump which in turn broke the DAML IDE completely.
changelog_begin
- [DAML Studio] Fix a bug in completions that caused DAML Studio to
stop responding after the first completion was requested.
changelog_end
* Use ghcide from master branch and simplify test
* Remove manual stack update
* Update rules_haskell
* rules_haskell_worker_dependencies after bazel-haskell-deps
* Update rules_haskell Windows patch
* make cabal haddock optional
* Don't generate Haddocks on stack_snapshot
Fails with ghc-lib and takes more time to build.
Co-authored-by: Andreas Herrmann <andreash87@gmx.ch>
* Get grpc from nix on unix
The one from Bazel seems to cause linking issues when trying to run
things in GHCi. I’ve spent some time trying to use rules_foreign_cc to
build gRPC using CMake but decided that for my own sanity it’s better
to not pursue that further.
* Address review comments
* Add missing module load
* Cleanup GHCI_SCRIPT
* use the correct file ending on macos
* Import is_linux
* Switch back to grpc-1.23
The newer version seems to cause issues in combination with the java libraries.
* Try to fix package_app on macos
* more debugging
* Maybe this is not necessary, we will never know
* linkers are the worst
* Remove debugging output again
* readd rpaths
* treat libdispatch specially
* remove hack
* more fooling around
* lalala
rules_haskell looks for stack in PATH. On Windows it is provided by
dadew (i.e. scoop). rules_haskell then symlinks (copies on Windows) the
stack binary. Unfortunately, this breaks with scoop as the shim file is
then not found.
