When the install fails, we end up printing the entire contents of
yarn's stdout to stdout. This output can often be quite long and will
cause Travis to fail in some cases.
The regular error output should be sufficent for us to diagnose any
issues we'll face.
Add a new property to ProgramTestOptions, `Overrides` that allows a
test to request a different version of a package is used instead of
what would be listed in the package.json file.
This will be used by our nightly automation to run everything "at head"
While the lifecycle tests wrote a `.yarnrc` file to ensure that copies
of `yarn` did not race with one another, the more barebones testing
framework did not.
This should address some of the yarn issues we've been seeing in CI
recently
The lockfile is not super interesting when running the lifecycle
tests, since the test project is not a long lived artifact. There was
no real harm in writing it, exepct that in some cases `pipenv` gets
very confused when you have dependencies on pre-release
versions (which is the case right now as we are brining up Python 3).
The packages are still installed correctly, even when we don't write a
lock file.
* Fix Python support in integration test framework
Update the integratino test framework to use pipenv to bootstrap new
virtual environments for tests and use those virtual environments to run
pulumi update and pulumi preview.
Fixespulumi/pulumi#2138
* Install packages via 'Dependencies' field
* Remove code for installing packages from Dependencies
In preparation for some workspace restructuring, I decided to scratch a
few itches of my own in the code:
* Change project's RuntimeInfo field to just Runtime, to match the
serialized name in JSON/YAML.
* Eliminate the no-longer-used Context and NoDefaultIgnores fields on
project, and all of the associated legacy PPC-related code.
* Eliminate the no-longer-used IgnoreFile constant.
* Remove a bunch of "// nolint: lll" annotations, and simply format
the structures with comments on dedicated lines, to avoid overly
lengthy lines and lint suppressions.
* Mark Dependencies and InitErrors as `omitempty` in the JSON
serialization directives for CheckpointV2 files. This was done for
the YAML directives, but (presumably accidentally) omitted for JSON.
* Remove TODO for issue since fixed in PPCs.
* Update issue reference to source
* Update comment wording
* Remove --ppc arg of stack init
* Remove PPC references in int. testing fx
* Remove vestigial PPC API types
This change adds GitLab CI support, by sniffing out the right
variables (equivalent to what we already do for Travis).
I've also restructured the code to share more logic with our
existing CI detection code, now moved to the pkg/util/ciutil
package, and will be fleshing this out more in the days to come.
Add two command-line options to the test framework, "-dirs" and
"list-dirs". The former accepts a regex that is used to select which
integration tests to run. The latter lists available integration tests
without running them.
The pattern we are using here is generally prone to error, so I hope we find a way to move away from this more generally - but for now we need to be able to configure more of these in places we are using `With`.
Also remove some `+""` that are tripping up the linter for me locally.
Often when the test fails, you'll want to use the web console to
investigate the stack, looking at detailed log messages, for
example. But since we always delete the stack today, you often can't.
Stop running `pulumi stack rm` when the test fails, however do run a
`pulumi destroy` so we at least clean up any resources the stack may
have allocated during testing.
Fixes#1722
Replace the Source-based implementation of refresh with a phase that
runs as the first part of plan execution and rewrites the snapshot in-memory.
In order to fit neatly within the existing framework for resource operations,
these changes introduce a new kind of step, RefreshStep, to represent
refreshes. RefreshSteps operate similar to ReadSteps but do not imply that
the resource being read is not managed by Pulumi.
In addition to the refresh reimplementation, these changes incorporate those
from #1394 to run refresh in the integration test framework.
Fixes#1598.
Fixespulumi/pulumi-terraform#165.
Contributes to #1449.
We retain a few tests on the RunBuild plan, with `typescript` set to
false in the runtime options, but for the general case, we remove the
build steps and custom entry points for our programs.
This change lets us set runtime specific options in Pulumi.yaml, which
will flow as arguments to the language hosts. We then teach the nodejs
host that when the `typescript` is set to `true` that it should load
ts-node before calling into user code. This allows using typescript
natively without an explicit compile step outside of Pulumi.
This works even when a tsconfig.json file is not present in the
application and should provide a nicer inner loop for folks writing
typescript (I'm pretty sure everyone has run into the "but I fixed
that bug! Why isn't it getting picked up? Oh, I forgot to run tsc"
problem.
Fixes#958
This allows us to delete the one off export/import test, which is nice
because it failed to run when PULUMI_ACCESS_TOKEN was not set in the
environment (due to an interaction between `pulumi login` and the
bare-bones integration test framework)
Because we run our golang integration tests "in tree" (due to
the need to be under $GOPATH and have a vendor folder around), the
"command-output" folder was getting left behind, dirtying the worktree
after building.
This change does two things:
1. On a succecssful run, remove the folder.
2. Ignore the folder via .gitignore (this way if a test fails and you
do `git add .` you don't end up commiting this folder).
Due to the way GOPATH and vendoring works, copying Go tests out to a
random temp directory simply will not work. This is largely a holdover
to the days of when .pulumi/ directories would pollute the directory,
but is also done for languages like Node.js whose preparatory and build
steps also pollute the working directory (with bin/ and node_modules/
directories). Go does not have this problem, so we can safely skip.
Today we defaulted our tests to create stacks in the `pulumi`
organization. We did this because our tests run with `pulumi-bot` and
we'd rather create the stacks in our shared organization, so any
Pulumi developer can see them.
Of course, as we prepare to have folks outside of the Pulumi
organization write and run tests, this has now become a bad default.
Remove the ability to explicitly set an owner in
ProgramTestOptions (since that would more or less only lead to pain
going forward) and default to just creating the stacks in whatever
account is currently logged in. In CI, we'll set a new environment
variable "PULUMI_TEST_OWNER" which controls the owner of the stacks,
which we'll set to `pulumi`.
Impact to day to day developers is during test runs locally you'll see
stacks in your list of stacks. If any of the tests fail to clean up,
you'll see these lingering stacks (but you can go clean them up).
math/rand uses a fixed seed, meaning that across runs the Kth call to
`rand.Int63()` will always return the same value.
Given that we want to provide a unique suffix across multiple
concurrent runs, this isn't the behavior we want.
I saw an instance fail in CI where all three legs ran the test
concurrently and they raced on creating the test stack, since they all
generated the same name.
This lets us set these values globaly, in our Travis and AppVeyor
configurations instead of forcing every test to opt-in. It also means
that by default, local builds will not report any of this data (and
will not need access to these endpoints).
These changes add support for pre- and post-`pulumi` callbacks to the
integration test framework. These callbacks will be invoked immediately
before and after (respectively) running a `pulumi` command.
This changes the CLI interface in a few ways:
* `pulumi preview` is back! The alternative of saying
`pulumi update --preview` just felt awkward, and it's a common
operation to want to perform. Let's just make it work.
* There are two flags consistent across all update commands,
`update`, `refresh`, and `destroy`:
- `--skip-preview` will skip the preview step. Note that this
does *not* skip the prompt to confirm that you'd like to proceed.
Indeed, it will still prompt, with a little warning text about
the fact that the preview has been skipped.
* `--yes` will auto-approve the updates.
This lands us in a simpler and more intuitive spot for common scenarios.
I found the flag --force to be a strange name for skipping a preview,
since that name is usually reserved for operations that might be harmful
and yet you're coercing a tool to do it anyway, knowing there's a chance
you're going to shoot yourself in the foot.
I also found that what I almost always want in the situation where
--force was being used is to actually just run a preview and have the
confirmation auto-accepted. Going straight to --force isn't the right
thing in a CI scenario, where you actually want to run a preview first,
just to ensure there aren't any issues, before doing the update.
In a sense, there are four options here:
1. Run a preview, ask for confirmation, then do an update (the default).
2. Run a preview, auto-accept, and then do an update (the CI scenario).
3. Just run a preview with neither a confirmation nor an update (dry run).
4. Just do an update, without performing a preview beforehand (rare).
This change enables all four workflows in our CLI.
Rather than have an explosion of flags, we have a single flag,
--preview, which can specify the mode that we're operating in. The
following are the values which correlate to the above four modes:
1. "": default (no --preview specified)
2. "auto": auto-accept preview confirmation
3. "only": only run a preview, don't confirm or update
4. "skip": skip the preview altogether
As part of this change, I redid a bit of how the preview modes
were specified. Rather than booleans, which had some illegal
combinations, this change introduces a new enum type. Furthermore,
because the engine is wholly ignorant of these flags -- and only the
backend understands them -- it was confusing to me that
engine.UpdateOptions stored this flag, especially given that all
interesting engine options _also_ accepted a dryRun boolean. As of
this change, the backend.PreviewBehavior controls the preview options.
This change introduces support for using the cloud backend when
`pulumi init` has not been run. When this is the case, we use the new
identity model, where a stack is referenced by an owner and a stack
name only.
There are a few things going on here:
- We add a new `--owner` flag to `pulumi stack init` that lets you
control what account a stack is created in.
- When listing stacks, we show stacks owned by you and any
organizations you are a member of. So, for example, I can do:
* `pulumi stack init my-great-stack`
* `pulumi stack init --owner pulumi my-great-stack`
To create a stack owned by my user and one owned by my
organization. When `pulumi stack ls` is run, you'll see both
stacks (since they are part of the same project).
- When spelling a stack on the CLI, an owner can be optionally
specified by prefixing the stack name with an owner name. For
example `my-great-stack` means the stack `my-great-stack` owned by
the current logged in user, where-as `pulumi/my-great-stack` would
be the stack owned by the `pulumi` organization
- `--all` can be passed to `pulumi stack ls` to see *all* stacks you
have access to, not just stacks tied to the current project.
To prepare for a world where stack names must be unique across an
owner, add some randomness to the names we use for stacks as part of
our integration tests.
This change does three major things:
1. Removes the ability to be logged into multiple clouds at the same
time. Previously, we supported being logged into multiple clouds at
the same time and the CLI would fan out requests and join responses
when needed. In general, this was only useful for Pulumi employees
that wanted run against multiple copies of the service (say production
and staging) but overall was very confusing (for example in the old
world a stack with the same identity could appear twice (since it was
in two backends) which the CLI didn't handle very well).
2. Stops treating the "local" backend as a special thing, from the
point of view of the CLI. Previouly we'd always connect to the local
backend and merge that data with whatever was in clouds we were
connected to. We had gestures like `--local` in `pulumi stack init`
that meant "use the local mode". Instead, to use the local mode now
you run `pulumi login --cloud-url local://` and then you are logged in
the local backend. Since you can only ever be logged into a single
backend, we can remove the `--local` and `--remote` flags from `pulumi
stack init`, it just now requires you to be logged in and creates a
stack in whatever back end you were logged into. When logging into the
local backend, you are not prompted for an access key.
3. Prompt for login in places where you have to log in, if you are not
already logged in.
Tests now target managed stacks instead of local stacks.
The existing logged in user and target backend API are used unless PULUMI_ACCES_TOKEN is defined, in which case tests are run under that access token and against the PULUMI_API backend.
For developer machines, we will now need to be logged in to Pulumi to run tests, and whichever default API backend is logged in (the one listed as current in ~/.pulumi/credentials.json) will be used. If you need to override these, provide PULUMI_ACCESS_TOKEN and possibly PULUMI_API.
For Travis, we currently target the staging service using the Pulumi Bot user.
We have decided to run tests in the pulumi organization. This can be overridden for local testing (or in Travis in the future) by defining PULUMI_API_OWNER_ORGANIZATION and using an access token with access to that organization.
Part of pulumi/home#195.
This change includes a bunch of refactorings I made in prep for
doing refresh (first, the command, see pulumi/pulumi#1081):
* The primary change is to change the way the engine's core update
functionality works with respect to deploy.Source. This is the
way we can plug in new sources of resource information during
planning (and, soon, diffing). The way I intend to model refresh
is by having a new kind of source, deploy.RefreshSource, which
will let us do virtually everything about an update/diff the same
way with refreshes, which avoid otherwise duplicative effort.
This includes changing the planOptions (nee deployOptions) to
take a new SourceFunc callback, which is responsible for creating
a source specific to the kind of plan being requested.
Preview, Update, and Destroy now are primarily differentiated by
the kind of deploy.Source that they return, rather than sprinkling
things like `if Destroying` throughout. This tidies up some logic
and, more importantly, gives us precisely the refresh hook we need.
* Originally, we used the deploy.NullSource for Destroy operations.
This simply returns nothing, which is how Destroy works. For some
reason, we were no longer doing this, and instead had some
`if Destroying` cases sprinkled throughout the deploy.EvalSource.
I think this is a vestige of some old way we did configuration, at
least judging by a comment, which is apparently no longer relevant.
* Move diff and diff-printing logic within the engine into its own
pkg/engine/diff.go file, to prepare for upcoming work.
* I keep noticing benign diffs anytime I regenerate protobufs. I
suspect this is because we're also on different versions. I changed
generate.sh to also dump the version into grpc_version.txt. At
least we can understand where the diffs are coming from, decide
whether to take them (i.e., a newer version), and ensure that as
a team we are monotonically increasing, and not going backwards.
* I also tidied up some tiny things I noticed while in there, like
comments, incorrect types, lint suppressions, and so on.
