31795 Commits (2cc94b4e89ccd47475c2b894e7cb9bd758ae73fc)

 

All Branches   

Search

Author	SHA1	Message	Date
Jared Baker	4f26451cd3	Merge pull request #34017 from hashicorp/s3/b-relax-endpoint-validation ... backend/s3: Relax endpoint validation rule	3 years ago
Craig Wright	83a8003482	Merge pull request #34028 from togakangaroo/patch-1	3 years ago
Graham Davison	3de6dbc7a8	Keeps validation error for IMDS endpoint and replaces error with warning for other services	3 years ago
Nick Fagerlund	57c57e4031	Merge pull request #34012 from hashicorp/nf/oct23-tf_workspace-problems ... Fix cloud integration's handling of TF_WORKSPACE environment variable	3 years ago
Jared Baker	9d360ffbed	Merge pull request #34002 from knoppiks/no-sts-s3-backend ... Add skip_requesting_account_id Parameter to S3 Backend	3 years ago
George Mauer	821b7ce750	Fix typos using wrong words	3 years ago
Nick Fagerlund	438e47659b	Remove some redundant test cases for resolveCloudConfig	3 years ago
Nick Fagerlund	26dac26fd1	fix configWithEnvVars tests ... Actually I don't think these _ever_ worked.	3 years ago
Nick Fagerlund	1f9c3eb405	Fix TestCloud_config ... That error happens after PrepareConfig, now.	3 years ago
Nick Fagerlund	d7dadca40f	Re-enable disabled cloud configure tests ... These got turned off in d7e07e66fc by accident	3 years ago
Nick Fagerlund	59c99b1231	Steal most test cases from setConfigurationFields and PrepareConfig ... These tests were rather chaotic, because the basic job of validating the finalized config was split across a bunch of different places. Now they're in (mostly) one spot, testing `resolveCloudConfig`. They might be somewhat excessive or redundant, and can be revised later. In a few cases, I revised test cases that had incorrect expectations, to match the fixed implementation. Also, note that the new implementation is less loosey-goosey around cty values, and expects something that definitely matches the schema (though any amount of it might be null). This commit also splits the TF_FORCE_LOCAL_BACKEND test case into a separate test function.	3 years ago
Nick Fagerlund	be71ab06a8	Add a first test for resolveCloudConfig	3 years ago
Nick Fagerlund	1ae8643201	Refactor `(*Cloud).Configure`, and fix TF_WORKSPACE bug ... This fixes issue #33976, introduced in https://github.com/hashicorp/terraform/pull/33489, which broke the intended behavior of specifying the active workspace via the TF_WORKSPACE variable when using a tag-based workspace mapping. Now that all the default and fallback value behaviors are cleanly isolated in a function, this whole flow can be a bit simpler. - Remove `setConfigurationFields`. Instead, `Configure` can just trade its Cty `obj` for a dumb struct and set a couple fields from it. The `TF_FORCE_LOCAL_BACKEND` handling can just join the relevant section of Configure directly. - Radically chop down PrepareConfig. It turns out we were violating the interface contract, which says PrepareConfig shouldn't trouble itself with the shell environment and fallback values... So, don't do that.	3 years ago
Nick Fagerlund	8e1b65ebc3	Add resolveCloudConfig function ... This will be the central location for everything involving combining environment variables with a `cloud` config block to obtain a final cloud config. It returns a plain Go value (so that nothing downstream of it ever needs to mess with Cty types), and doesn't mutate any fields on the backend, so it has a nice firm boundary of responsibilities. Also, it's quite a bit more pedantic and explicit about HOW the environment variables get consulted, in the hope of reducing future misunderstandings about our UI-level expectations.	3 years ago
James Bardin	c820d448b7	Merge pull request #34004 from hashicorp/jbardin/openpgp-key-expiration ... cli: ignore expired provider signing keys from registry during init	3 years ago
Martin Atkins	dc376bc4c4	go.mod: cty v1.14.1 and HCL v2.18.1 ... These each fix some different bugs related to handling of "refined" unknown values, both of which were causing panics with some input.	3 years ago
James Bardin	0d4a29f7f3	ignore expired provider signing keys ... Provider developers are not currently required to keep the signing keys stored in the registry up to date, and older releases may be signed with a key which has since expired. For our purposes here however, we are validating the key and signature used at the time of publishing, and given that the registry has previously vouched for the validity of the key used, we can continue to trust that key returned by the registry for installation.	3 years ago
James Bardin	86667375fd	remove errant debug print	3 years ago
Jonas Wagner	09df2213f9	Add skip_requesting_account_id Parameter to S3 Backend ... Signed-off-by: Jonas Wagner <jwagner@knoppiks.de>	3 years ago
Andrew den Hertog	6088002a2d	fix(s3): allow aliases for kms key	3 years ago
Alisdair McDiarmid	3cb32ec2de	Merge pull request #33964 from hashicorp/tfc-upload-patch ... Change release order of operations	3 years ago
Radek Simko	1d1dcc4b8b	cloud: Fix typo & align 'project' field description (#33946)	3 years ago
kmoe	4b88808201	configs/terraform: refactor expansion code, add diags (#33962) ... * configs: remove unneeded code This code inside addProviderRequirements adds an entry to the provider map for each resource targeted by an import block. In the normal case without config generation, a resource block must exist for each import block, so provider requirements will be collected from the resource block. In the config generation case, the code block directly below this deleted one will add the appropriate provider requirements where these must come from the import block. * terraform: diags in DynamicExpand	3 years ago
Michele Degges	624e9e0e68	Change release order of operations ... This PR moves the important TFC upload workflow higher up the release chain, so that it immediately happens after the binaries are published to the Releases Site, and the GitHub release is published. This will allow the team to perform other post-release steps (announcements, etc) without needing to wait for the more time-consuming linux packaging & docker publishing workflows to complete.	3 years ago
Liam Cervante	01c0480e77	Test framework will produce regular updates in -json mode (#33953) ... * Test framework will produce regular updates in -json mode * add tests * docs * docs tweaking	3 years ago
James Bardin	2664c062c9	Merge pull request #33932 from hashicorp/jbardin/import-for-each ... Import block expansion with for_each	3 years ago
Liam Cervante	8ecda5f708	Increase timeout length after fatal interrupts for test cloud runs (#33956)	3 years ago
Liam Cervante	8360d7566e	Implement double cancel instead of force cancel for test cloud runs (#33951)	3 years ago
Liam Cervante	fba5dad097	Test files should be marked as skipped if all their run blocks were skipped (#33950)	3 years ago
Graham Davison	6066326d93	Merge pull request #33940 from hashicorp/s3/b-put-with-object-locks ... backend/s3: Supports S3 buckets with Object Lock	3 years ago
James Bardin	66637ed228	use new unwrapJSONRefExpr for replace_triggered_by ... Get rid of the duplicated code, and use the new function which also validates the expression result type that was missed in the existing implementation.	3 years ago
James Bardin	a2b189fd71	handle to expression from json files ... Because expression from JSON files are handled differently from HCL, we need to deal with these manually when evaluating the special "to" syntax. We take the same approach as was done for replace_triggered_by, where the raw string is extracted from the json expression, and re-evaluated as an HCL expression.	3 years ago
James Bardin	59d8aa0af1	check for dynamic for_each during validation	3 years ago
James Bardin	2c4608fe9e	update import validate test ... The diagnostics have been made more accurate, pointing to the problematic line rather than just the import block name.	3 years ago
James Bardin	8619450304	remove unused method	3 years ago
James Bardin	1831536df5	fix to expression target error	3 years ago
James Bardin	1f260f0f51	import for_each from data source	3 years ago
James Bardin	e522134fb8	import id ref in module	3 years ago
James Bardin	167f235fa9	partial for_each import	3 years ago
James Bardin	e3cb32c4f4	first import for_each tests	3 years ago
James Bardin	6ee08c7b0c	add new context_plan_import_test.go file ... the plan test file is getting large again, so group import tests together	3 years ago
James Bardin	e1a8270bde	expose ImportTarget.IDString for legacy import ... Let the legacy import path set IDString directly, simplifying the handling slightly. This also uncovered a long broken test which used an invalid import target.	3 years ago
James Bardin	7f6e320b84	handle legacy import in new expansion node ... We use the same graph node to start all resource planning, but must deal with the legacy import path differently because we need to end up in a legacy graphNodeImportState.	3 years ago
James Bardin	ba6f471c74	further refine import validation	3 years ago
James Bardin	41fabcfbe0	don't add identical config-gen resource nodes ... If the -generate-config-out flag was given but there are multiple import blocks for the instance, we can't add multiple nodes for the same resource.	3 years ago
James Bardin	e130c7b9ab	correct check for duplicate imports	3 years ago
James Bardin	5786f824ca	improve import diagnostic handling	3 years ago
James Bardin	5f13eecc38	remove ImportTarget.ID field	3 years ago
James Bardin	0c005ce484	remove legacyImportMode flag	3 years ago
James Bardin	5128b21b55	skip imports for existing resources	3 years ago