Commit Graph

27754 Commits

Author SHA1 Message Date
Pam Selle b25d0251ca Add name to untaint missing warning 2021-01-13 15:14:19 -05:00
James Bardin 685022fae7 stop passing errors into the instance apply method
Trying to track these error values as they wint into and out of the
instance apply methods was quite difficult. They were mis-assigned, and
in many cases lost diagnostic information.
2021-01-13 15:08:53 -05:00
Pam Selle 67853d8fa8 Show resource name in taint -allow-missing warning
Show the resource name in the warning when allow-missing
is used and no resource matches
2021-01-13 15:00:35 -05:00
James Bardin 1707685350 remove the applyError argument to apply
This was never assigned, and the empty value made the usage very
difficult to trace.
2021-01-13 14:09:20 -05:00
James Bardin 842c5b4136 ensure status, private, warnings, etc are retained
Various pieces of the state and/or warnings were dropped when the
provider returns an error. Do a little cleanup of `.apply` to make the
logic easier to follow.
2021-01-13 13:37:20 -05:00
Pam Selle 61f9d2177b
Merge pull request #27486 from hashicorp/pselle/version-doc
Add documentation page for version
2021-01-13 12:44:11 -05:00
Pam Selle 80ad3924de
Update CHANGELOG.md 2021-01-13 12:43:06 -05:00
Pam Selle d34e1d926a
Merge pull request #27484 from hashicorp/pselle/revision
Remove revision from version command
2021-01-13 12:42:09 -05:00
Alisdair McDiarmid ddac282553 cli: Fix double check of remote workspace version
After verifying the remote backend workspace version is compatible with
the local Terraform version, we need to record that this check was
successful. Otherwise the fallback error handling path will run a strict
version check, even if the versions are compatible, which will cause an
unexpected failure.
2021-01-13 10:43:35 -05:00
James Bardin dcf49dc7dd
Merge pull request #27491 from hashicorp/jbardin/private-data-error
preserve resource private data on error
2021-01-13 09:00:27 -05:00
James Bardin e0d3b97c8e preserve resource private data on error
If the provider returns an empty response during apply, we restore the
prior state for the resource, but the private data was not being
restored.
2021-01-13 08:31:32 -05:00
James Bardin 7e880299c0
Merge pull request #27485 from hashicorp/jbardin/mock-provider
Update the MockProvider to use correct types and allow default behaviors
2021-01-13 08:25:07 -05:00
Rémy Léone edb4c00a86
website: add links to the registry provider documentation (#27416)
* docs: add links to the registry provider documentation

* Adjust wording/link URL

Co-authored-by: Nick Fagerlund <nick@hashicorp.com>
2021-01-12 16:06:50 -08:00
Nick Fagerlund 303b326008
Merge pull request #27279 from kowshikRoy/add-var-multiple-example
website: add example of multiple var cli
2021-01-12 15:58:31 -08:00
James Bardin 3bc7d77230 update MockProvider usage 2021-01-12 17:47:55 -05:00
James Bardin 2e2c9e07e5 add default behavior to the MockProvider
Add reasonable default behavior to the mock provider, so that may do not
need to depend on the idiosyncrasies of the old (though updated) test
testDiffFn and to a lesser extend the testApplyFn. This behavior is
based directly upon the documented resource lifecycle, rather
than be an ad-hoc collection of behaviors from old tests.

As a proof of concept, remove all uses of testDiffFn from the plan
context tests that don't cause the tests to fail.
2021-01-12 17:47:55 -05:00
Nick Fagerlund 66f2c82d33 website: Add version command to sidebar nav and to plugins overview page 2021-01-12 14:40:45 -08:00
Nick Fagerlund ccb205896d website: Fix alphabetical ordering of commands 2021-01-12 14:34:45 -08:00
Pam Selle cf0c54f834 Add documentation page for version
The version command, although not the mightiest of our commands,
did not have a page among our docs.
2021-01-12 17:06:53 -05:00
Pam Selle 83e6703bf7 Remove revision from version command
The revision field is only populated on dev builds so this means
most releases of Terraform have an empty "terraform_revision" field
in the JSON output. Since we recommend developers use go tooling
to `go build` this tool when developing, the revision is not useful
data and so it is removed.
2021-01-12 16:35:30 -05:00
James Bardin e01d37d0dc Block.AttribuuteByPath
There are a few places where we want to perform some transformation on a
cty.Value, but require information from the schema. Rather than create
bespoke functions to walk the cty.Value and schema in concert, we can
provide Attribute information from a cty.Path allowing the use of
Value.Transform in these cases.
2021-01-12 16:31:18 -05:00
James Bardin 7dd570ef6c change mock provider to use GetSchemaResponse
This ensures that test providers are using the same types as actual
providers.
2021-01-12 16:31:18 -05:00
James Bardin 5fe848b642 change mock return values to pointers
This allows up to detect an unset value from the zero value so that
defaults can be implemented, while still allowing tests to return
specific values of their choosing.
2021-01-12 15:16:48 -05:00
Pam Selle 9371b60d71 More verbose error for whole resource move failure
When running state mv with a resource source, but the destination
fails, provide a hint that the source is a resource (not an instance)
in case the user means to address it this way
2021-01-12 13:21:25 -05:00
Pam Selle 52c77ba33e Print addresses in state mv errors
Using the addrTo after it has failed its check means <invalid>/no
address will be printed. Change this throughout, but particularly
add a test for the origin issue for this.
2021-01-12 13:05:01 -05:00
Pam Selle ae52190c01
Merge pull request #27415 from techsolx/patch-1
Chage backend.tmpl to config.tmpl in Maps example,
2021-01-12 10:43:08 -05:00
Alisdair McDiarmid f5785b43c2
Merge pull request #27440 from hashicorp/alisdair/fix-rendering-of-long-integers
cli: Fix rendering of long integers
2021-01-12 10:06:40 -05:00
Pam Selle 516ad2c8e4
Merge pull request #27432 from bulletinmybeard/patch-1
Update gcs.html.md
2021-01-11 16:19:13 -05:00
Pam Selle c806162ddd
Update CHANGELOG.md 2021-01-11 15:03:22 -05:00
Pam Selle 0d87c66211
Merge pull request #27463 from hashicorp/pselle/get-plugins
Remove get-plugins flag
2021-01-11 15:02:22 -05:00
Pam Selle d232da4077
Update CHANGELOG.md 2021-01-11 14:58:41 -05:00
Pam Selle 801f8aa0b7
Merge pull request #27461 from hashicorp/pselle/verify-plugins
Remove verify-plugins flag
2021-01-11 14:57:45 -05:00
Pam Selle e22ccd424c Add note about removal to web docs 2021-01-11 14:55:49 -05:00
Pam Selle 462caaf538
Merge pull request #27464 from hashicorp/pselle/init-lock
Remove state lock flags from init
2021-01-11 14:54:42 -05:00
James Bardin d26c149174
Merge pull request #27408 from hashicorp/jbardin/destroy-plan-refresh
Refresh during destroy
2021-01-11 14:24:28 -05:00
James Bardin 62815debbf
Merge pull request #27450 from hashicorp/jbardin/internal-provisioner-panics
provisioners: ensure context cancel is never nil in builtin provisioners
2021-01-11 12:42:53 -05:00
Kristin Laemmert 7467805cc6
Update CHANGELOG.md 2021-01-11 12:34:43 -05:00
Jacob Martin b49655724d
command: Fix terraform show not outputting child_modules properly in certain circumstances (#27352)
* Add test for module nesting without resources.

* Add test

* Fix showing resources when a module has no resources, only submodules.
2021-01-11 12:31:20 -05:00
Pam Selle bbcf6ae20b Remove state lock flags from init 2021-01-11 12:02:03 -05:00
Pam Selle 5ad6100ff3 Remove get-plugins
Remove the no-op get-plugins flag
2021-01-11 11:47:23 -05:00
Pam Selle aab0dd102d
Merge pull request #27455 from cladius/patch-1
Corrected minor typo
2021-01-11 11:40:05 -05:00
Pam Selle da5a28e6ae Remove verify-plugins flag
This flag does not do anything, and so this removes the mention in
the CLI docs and the parsing of the flag variable
2021-01-11 11:38:17 -05:00
Cladius Fernando 25003a0caa
Corrected minor typo
On Windows, the file must be named named terraform.rc => On Windows, the file must be named terraform.rc
2021-01-09 10:52:59 +05:30
James Bardin fb2208a6d9 refactor context plan
get rid of the mutation of the plans.Plan and return values.
2021-01-08 13:29:54 -05:00
James Bardin e614fb9aed refresh is expected for destroy
These tests were not previously running a refresh, and hence did not
expect the resources to be read.
2021-01-08 13:29:54 -05:00
James Bardin 0b3b84acc1 refresh state during a destroy plan
Because the destroy plan only creates the necessary changes for apply to
remove all the resources, it does no reading of resources or data
sources, leading to stale data in the state. In most cases this is not a
problem, but when a provider configuration is using resource values, the
provider may not be able to run correctly during apply. In prior
versions of terraform, the implicit refresh that happened during
`terraform destroy` would update the data sources and remove missing
resources from state as required.

The destroy plan graph has a minimal amount of information, so it is not
feasible to work the reading of resources into the operation without
completely replicating the normal plan graph, and updating the plan
graph and all destroy node implementation is also a considerable amount
of refactoring. Instead, we can run a normal plan which is used to
refresh the state before creating the destroy plan. This brings back
similar behavior to core versions prior to 0.14, and the refresh can
still be skipped using the `-refresh=false` cli flag.
2021-01-08 13:29:54 -05:00
James Bardin 3e224df379 destroy plan with provider config from data 2021-01-08 13:29:54 -05:00
James Bardin 026ba5b013 ensure context cancel is never nil
The context used for Stop is more appropriately tied to the lifetime of
the provisioner rather than a call to the ProvisionResource method. In
some cases Stop can be called before ProvisionResource, causing a panic
the provisioner.  Rather than adding nil checks to the CancelFunc call
for Stop, create a base context to use for cancellation with both Stop
and Close methods.
2021-01-08 12:36:45 -05:00
Pam Selle d175e67bc9
Merge pull request #27429 from hashicorp/pselle/staticcheck
Staticcheck fixes in terraform package
2021-01-08 11:10:40 -05:00
Alisdair McDiarmid d763e4f73e cli: Fix rendering of long integers
Recent changes to the human-readable rendering of outputs and console
values led to long integer values being presented in scientific
notation (e.g. 1.2345678e+07). While technically correct, this is an
unusual way to present integer values.

This commit changes the formatting mode to 'f', which never uses
scientific notation and displays integer values as a sequence of digits
instead (e.g. 12345678).
2021-01-08 09:38:18 -05:00