The omitUnknowns and unknownAsBool functions were previously trying hard
to preserve the same collection types in the output as they had in the
input, by attempting to keep everything matched up so that the results
would be valid.
Unfortunately, this turns out to be a harder problem than we originally
thought: it was possible for a collection value going in to produce
inconsistent element types out (and thus a panic) in the following
situations:
- when a collection with mixed known and unknown values was passed in
to omitUnknowns.
- when a collection of collections where the inner collections are a
mixture of empty and not empty in unknownAsNull.
The results of these functions are only used to marshal to JSON anyway,
and JSON serialization can't distinguish between the three sequence types
or the two mapping types, so in practice we can just standardize on
converting all sequences to tuple and all mappings to object here and not
change the resulting output at all, and then we don't have to worry about
making sure all of the inner types get preserved exactly.
A nice consequence of that relaxation is that we can now do what we
originally wanted to do with unknownAsBool, and omit map keys and
object attributes altogether if their values would've been false,
producing a much more compact result. This is easiest to do now when
there's only one known user of this JSON plan output, and we know that
user will treat both false and omitted as the same here.
This includes a small fix to ensure the parser doesn't produce an invalid
body for block parsing syntax errors, and instead produces an incomplete
result that calling applications like Terraform can still analyze.
The problem here was affecting our version-constraint-sniffing code, which
intentionally tried to find a core version constraint even if there's a
syntax error so that it can report that a new version of Terraform is a
likely cause of the syntax error. It was working in most cases, unless
it was the "terraform" block itself that contained the error, because then
we'd try to analyze a broken hcl.Block with a nil body.
This includes a new test for "terraform init" that exercises this
recovery codepath.
There are a number of use cases that can require a user to select a workspace after initializing Terraform.
To make sure we cover all these use cases, we will always call the selectWorkspace method to verify a valid workspace is already selected or (if needed) offer to select one before moving on.
cty now guarantees that sets of primitive values will iterate in a
reasonable order. Previously it was the caller's responsibility to deal
with that, but we invariably neglected to do so, causing inconsistent
ordering. Since cty prioritizes consistent behavior over performance, it
now imposes its own sort on set elements as part of iterating over them so
that calling applications don't have to worry so much about it.
This change also causes cty to consistently push unknown and null values
in sets to the end of iteration, where before that was undefined. This
means that our diff output will now consistently list additions before
removals when showing sets, rather than the ordering being undefined as
before.
The ordering of known, non-null, non-primitive values is still not
contractually fixed but remains consistent for a particular version of
cty.
* internal/initwd: Allow deprecated relative module paths
In Terraform 0.11 we deprecated this form but didn't have any explicit
warning for it. Now we'll still accept it but generate a warning. In a
future major release we will drop this form altogether, since it is
ambiguous with registry module source addresses.
This codepath is covered by the command/e2etest suite.
* e2e: Skip copying .exists file, if present
We use this only in the "empty" test fixture in order to let git know that
the directory exists. We need to skip copying it so that we can test
"terraform init -from-module=...", which expects to find an empty
directory.
* command/e2etests: Re-enable and fix up the e2etest "acctests"
We disabled all of the tests that accessed remote services like the
Terraform Registry while they were being updated to support the new
protocols we now expect. With those services now in place, we can
re-enable these tests.
Some details of exactly what output we print, etc, have intentionally
changed since these tests were last updated.
* e2e: refactor for modern states and plans
* command/e2etest: re-enable e2etests and update for tf 0.12 compatibility
plugin/discovery: mkdirAll instead of mkdir when creating cache dir
Once you start reading from stdin, that is a blocking call that will
never finish. So when a context is canceled causing the input method to
return, the read will remain blocking in the running goroutine.
There isn't a real solution for it (e.g. its not possible to unblock the
read) so the only solution is to make the reader reusable.
When rendering the diff, the NoOp changes should come from the LCS
sequence, rather than the new sequence. The two indexes will not align
in many cases, adding the wrong new object or indexing out of bounds.
* command/state_list.go: fix bug loading user-defined state
If the user supplied a state path via the `-state` flag and terraform
was running in a workspace other than `default`, the state was not being
loaded properly. Fixes#19920
In study of existing providers we've found a pattern we werent previously
accounting for of using a nested block type to represent a group of
arguments that relate to a particular feature that is always enabled but
where it improves configuration readability to group all of its settings
together in a nested block.
The existing NestingSingle was not a good fit for this because it is
designed under the assumption that the presence or absence of the block
has some significance in enabling or disabling the relevant feature, and
so for these always-active cases we'd generate a misleading plan where
the settings for the feature appear totally absent, rather than showing
the default values that will be selected.
NestingGroup is, therefore, a slight variation of NestingSingle where
presence vs. absence of the block is not distinguishable (it's never null)
and instead its contents are treated as unset when the block is absent.
This then in turn causes any default values associated with the nested
arguments to be honored and displayed in the plan whenever the block is
not explicitly configured.
The current SDK cannot activate this mode, but that's okay because its
"legacy type system" opt-out flag allows it to force a block to be
processed in this way anyway. We're adding this now so that we can
introduce the feature in a future SDK without causing a breaking change
to the protocol, since the set of possible block nesting modes is not
extensible.
If the registry is unresponsive, you will now get an error
specific to this, rather than a misleading "provider unavailable" type
error. Also adds debug logging for when errors like this may occur
Due to these tests happening in the wrong order, removing an object from
the end of a sequence of objects would previously cause a bounds-check
panic.
Rather than a more severe rework of the logic here, for now we'll just
introduce an extra precondition to prevent the panic. The code that
follows already handles the case where there _is_ no new object (i.e. the
"old" object is being deleted) as long as we're able to pass through this
type-checking logic.
The new "JSON list of objects - removing item" test covers this problem
by rendering a diff for an object being removed from the end of a list
of objects within a JSON value.
Terraform Registry (and other registry implementations) can now return
an array of warnings with the versions response. These warnings are now
displayed to the user during a `terraform init`.
In earlier refactoring we updated these commands to support the new
address and state types, but attempted to partially retain the old-style
"StateFilter" abstraction that originally lived in the Terraform package,
even though that was no longer being used for any other functionality.
Unfortunately the adaptation of the existing filtering to the new types
wasn't exact and so these commands ended up having a few bugs that were
not covered by the existing tests.
Since the old StateFilter behavior was the source of various misbehavior
anyway, here it's removed altogether and replaced with some simpler
functions in the state_meta.go file that are tailored to the use-cases of
these sub-commands.
As well as just generally behaving more consistently with the other
parts of Terraform that use the new resource address types, this commit
fixes the following bugs:
- A resource address of aws_instance.foo would previously match an
resource of that type and name in any module, which disagreed with the
expected interpretation elsewhere of meaning a single resource in the
root module.
- The "terraform state mv" command was not supporting moves from a single
resource address to an indexed address and vice-versa, because the old
logic didn't need to make that distinction while they are two separate
address types in the new logic. Now we allow resources that do not have
count/for_each to be treated as if they are instances for the purposes
of this command, which is a better match for likely user intent and for
the old behavior.
Finally, we also clean up a little some of the usage output from these
commands, which hasn't been updated for some time and so had both some
stale information and some inaccurate terminology.
* command/providers schemas: return empty json object if config parses successfully but no providers found
* command/show (state): return an empty object if state is nil
* configs/configupgrade: detect possible relative module sources
If a module source appears to be a relative local path but does not have
a preceding ./, print a #TODO message for the user.
* internal/initwd: limit go-getter detectors to those supported by terraform
* internal/initwd: move isMaybeRelativeLocalPath check into getWithGoGetter
To avoid making two calls to getter.Detect, which potentially makes
non-trivial API calls, the "isMaybeRelativeLocalPath" check was moved to
a later step and a custom error type was added so user-friendly
diagnostics could be displayed in the event that a possible relative local
path was detected.
Our initial prototype of new-style diff rendering excluded this because
the old SDK has no support for this construct. However, we want to be able
to introduce this construct in the new SDK without breaking compatibility
with existing versions of Terraform Core, so we need to implement it now
so it's ready to be used once the SDK implements it.
The key associated with each block allows us to properly correlate the
items to recognize the difference between an in-place update of an
existing block and the addition/deletion of a block.
Our null-to-empty normalization was previously assuming these would always
be collection types, but that isn't true when a block contains something
dynamic since we must then use tuple or object types instead to properly
represent all of the individual element types.
We use cty a little differently when a nested list block contains a
dynamically-typed attribute: it appears as a tuple value instead of a
list value so that we can retain the individual types of each element.
Here we introduce a test for that case, but doing so required also making
the runTestCases function handle types in a stricter way so that it will
produce planned values that match how Terraform Core would do it,
including the necessary late-bound type information for the
dynamically-typed attribute.
Previously, these commands were not checking if the user specified a
`-plugin-dir` flag during `terraform init` and would therefor fail if
providers were not in one of the standard directories.
Fixes#20547
When the user aborts input, it may end up as an unknown value, which
needs to be converted to null for PrepareConfig.
Allow PrepareConfig to accept null config values in order to fill in
missing defaults.
When a planfile is supplied to the `terraform show -json` command, the
context that loads only included schemas for resources in the plan. We
found an edge case where removing a data source from the configuration
(though only if there are no managed resources from the same provider)
would cause jsonstate.Marshal to fail because the provider schema wasn't
in the plan context.
jsonplan.Marshal now takes two schemas, one for plan and one for state.
If the state schema is nil it will simply use the plan schemas.
* command/show: fixing bugs in modulecalls
jsonconfig and jsonplan both had subtle bugs with the logic for
marshaling module calls that only showed up when multiple modules were
referenced. This PR fixes those bugs and extends the existing tests to
include multiple modules.
* sort all the things, mostly for tests
* docs: update plan command documentation. Fixes#19235
* docs: added a missing reserved variable name. Fixes#19159.
* website: add note that resource names cannot start with a number
* website: add some notes to the 0.12 upgrade guide
We are now allowing the legacy SDK to opt out of the safety checks we try
to do after plan and apply, and so in such cases the before/after values
in planned changes may be inconsistent with our usual rules.
To avoid adding lots of extra complexity to the diff renderer to deal with
these situations, instead we'll normalize the handling of nested blocks
prior to using these values.
In the long run it'd be better to do this normalization at the source,
immediately after we receive an object from a provider using the opt-out,
but we're doing this at the outermost layer for now to avoid risking
unintended impacts on other Terraform Core components when we're just
about to enter the beta phase of the v0.12.0 release cycle.
This mirrors the change made for providers, so that default values can
be inserted into the config by the backend implementation. This is only
the interface and method name changes, it does not yet add any default
values.
Martin Atkins