1120 Commits (paddy_module_attribution_tmp)

Author	SHA1	Message	Date
Paddy Carver	cf3f03bde6	Fill in provider_meta where it was missing in schema.	6 years ago
Paddy Carver	7d14463660	Update in response to review. ... So thanks to review, we have a few changes all rolled up in one here, to account for a design change. First, the codebase has marched on while we wrote this, meaning we need to refer to providers as addrs.Provider, not as strings. This is for the best, but is a lot of replacing through the codebase. Second, some methods were missing from the provider-side GRPC implementations, so we added those. Third, it turns out that attaching the meta configs to resources in transform_provider wouldn't actually work, because resources aren't guaranteed to be in the same module as the provider config powering them. So we need to attach them in transform_attach_config_resource, not in transform_provider, so they'll be on the right side of the module boundary. Unfortunately, this means we're going to have to stop passing around the provider's provider_meta specifically, and are going to have to pass the entire block, because we don't have the provider data we need when we attach the config to filter out just what we need. We now separate them out in the Eval methods. This, unfortunately, also causes us issues, because we don't have that info already when we call EvalValidate, which means we're either going to need to pipe new info through to it, or we're not going to be able to validate the schema pre-plan-time. There's still a recommendation for an end-to-end test to do, but I'm hopeful.	6 years ago
Paddy Carver	300b054f12	OK but run the tests this time. ... Fix some panics because I didn't run _all_ the tess, like a dolt.	6 years ago
Paddy Carver	f3a950de5b	Make ProviderMeta available in the SDK.	6 years ago
Paddy Carver	546b58befc	First implementation of module attribution. ... A pass at implementing module attribution ("provider_meta blocks in the terraform block", whatever, it's for module attribution). The general idea is that we're adding a new ProviderMeta struct to our configs package, which is then added to Files, Modules, and parsed as part of the terraform block. The configs.Config type then makes that available to the terraform.ProviderTransformer type, which is already scoped by module. It picks out the ProviderMeta for the provider we need in its Transform method, and checks to see if the node it's working on implements the terraform.GraphNodeAttachProviderMetaConfigs interface, which is a new interface we're adding. It has only one method: AttachProviderMetaConfigs, which is passed a *configs.ProviderMeta. We implement this on the terraform.NodeAbstractResource type, storing the passed *configs.ProviderMeta in a new property on that struct. That struct is eventually made available to the terraform.NodeApplyableResourceInstance type, and its method evalTreeManagedResources, which creates an EvalApply type. The EvalApply type gets a ProviderMeta type, which is set to the *configs.ProviderMeta obtained from the terraform.NodeAbstractResource available to it. This means that EvalApply now has the raw config values the user entered for the provider_meta block, but we also need a schema; providers can define their own schema for the provider_meta block, and we need that schema to know how to translate what is in the config into a cty.Value, which is what we eventually want here. Fortunately, EvalApply already has access to the provider's schema; we can augment the GetSchema RPC call to pass back the ProviderMeta schema (added to helper/schema.Provider) in addition to resource, data source, and provider schemas. This makes it available, eventually, to EvalApply. Now that EvalApply has both the config values and the schema, we can parse them into a cty.Value. Finally, we augment the ApplyResourceChange RPC call with a ProviderMeta field, which can be populated by EvalApply with the parsed cty.Value, which successfully makes the config values available to the provider. This still needs tests written for it to ensure it works, but no current tests are broken on it and it compiles, which is reassuring.	6 years ago
Kristin Laemmert	e3416124cc	addrs: replace "Type string" with "Type Provider" in ProviderConfig ... * huge change to weave new addrs.Provider into addrs.ProviderConfig * terraform: do not include an empty string in the returned Providers / Provisioners - Fixed a minor bug where results included an extra empty string	6 years ago
Martin Atkins	97d64581c6	helper/logging: Loosen the filtering heuristic even more ... Now we'll consider any message that doesn't start with a digit as a continuation. This is _definitely_ in loose, best-effort territory now since a continuation line could very well start with a digit, but ultimately this is still better than the totally broken behavior that preceded it. Just to make sure that none of these heuristics interfere with producing complete, accurate logs for TRACE, we'll skip the LevelFilter altogether in that case.	6 years ago
Martin Atkins	faffb3a784	helper/logging: Warn about levels other than TRACE ... The filtering for other log levels is unreliable and glitchy because it's trying to infer information from the log stream that isn't reliably represented. Although the previous commit has improved the situation somewhat, it is still a tricky and unreliable heuristic, so worth a warning to anyone who is reading such a log that if they see something confusing it could be a result of the heuristic not working fully.	6 years ago
Martin Atkins	30bf83cdeb	helper/logging: Bring the LevelFilter into our own codebase ... In order to make this work reasonably we can't avoid using some funny heuristics, which are somewhat reasonable to apply within the context of Terraform itself but would not be good to add to the general "logutils". Specifically, this is adding the additional heuristic that lines starting with spaces are continuation lines and so should inherit the log level of the most recent non-continuation line.	6 years ago
Kristin Laemmert	9891d0354a	providers: use addrs.Provider as map keys for provider.Factory (#23548) ... * terraform/context: use new addrs.Provider as map key in provider factories * added NewLegacyProviderType and LegacyString funcs to make it explicit that these are temporary placeholders This PR introduces a new concept, provider fully-qualified name (FQN), encapsulated by the `addrs.Provider` struct.	6 years ago
James Bardin	bee703360c	Merge pull request #23252 from hashicorp/jbardin/abs-state-dependencies ... store absolute addresses for resource dependencies in the state	7 years ago
James Bardin	16b368e292	only need to handle depends_on to shim old state	7 years ago
James Bardin	49439d02d1	sanitize provisioner output strings ... The grpc protocol requires strings to be valid utf8, but because provisioners often don't have control over the command output, invalid utf8 sequences can make it into the response causing grpc transport errors. Replace all invalid utf sequences with the standard utf replacement character in the provisioner output. The code is a direct copy from the go1.13 std library, and can be replaced with strings.ToValidUTF8 once it's available.	7 years ago
Radek Simko	7860f55e4f	Version tools per Go convention under tools.go	7 years ago
Martin Atkins	39e609d5fd	vendor: switch to HCL 2.0 in the HCL repository ... Previously we were using the experimental HCL 2 repository, but now we'll shift over to the v2 import path within the main HCL repository as part of actually releasing HCL 2.0 as stable. This is a mechanical search/replace to the new import paths. It also switches to the v2.0.0 release of HCL, which includes some new code that Terraform didn't previously have but should not change any behavior that matters for Terraform's purposes. For the moment the experimental HCL2 repository is still an indirect dependency via terraform-config-inspect, so it remains in our go.sum and vendor directories for the moment. Because terraform-config-inspect uses a much smaller subset of the HCL2 functionality, this does still manage to prune the vendor directory a little. A subsequent release of terraform-config-inspect should allow us to completely remove that old repository in a future commit.	7 years ago
James Bardin	a3c374b227	don't lose track of private data in the ACC tests ... If the resource schema version is > 0, the metadata containing the timeout values was overwritten during acceptance tests.	7 years ago
appilon	d3dc1263bf	Merge pull request #22382 from hashicorp/appilon/delete-files ... [CLEANUP] Prune internal usage of config package (with 1 exception)	7 years ago
Radek Simko	ecbfc365a0	Merge pull request #20157 from shinmog/conflictswith ... helper/schema: Add "part" to ConflictsWith validation error	7 years ago
Radek Simko	db7d1826e4	Merge pull request #22312 from ewbankkit/add-rand-ip-address ... helper/acctest: Add function to return random IP address	7 years ago
Radek Simko	7861cc6141	Merge pull request #20309 from chanzuckerberg/ryanking/remove_lifecycle_comment ... helper/schema: Remove comment about nested schema.Resources having their own lifecycle	7 years ago
Radek Simko	350395d804	Add better comment	7 years ago
Radek Simko	a08ad08169	Merge pull request #18277 from ndmckinley/patch-1 ... helper/resource: Minor docs tweak on ImportStateVerify	7 years ago
Alex Pilon	77757d9f5b	prune references to config/module ... delete config/module prune references to config except in terraform/resource.go move, cleanup, and delete inert code	7 years ago
Alex Pilon	4bf43efcfd	move hcl2shim package to configs	7 years ago
Alex Pilon	ec0402a238	add case to decoder to assert to slice, then each item to a map	7 years ago
Alex Pilon	6ae281b34f	set Err to true for test ... This was likely meant to be set to Err true but the old parser mistakenly didn't raise an error	7 years ago
Alex Pilon	6de3c32652	comment out failing test	7 years ago
Alex Pilon	83aa07f907	prune NewResourceConfig and update tests	7 years ago
Kit Ewbank	949a6ae695	Add 'RandIpAddress'.	7 years ago
The Terraform Team	176f790323	Merge pull request #22262 from hashicorp/mildwonkey/b-deprecation-warnings ... helper/schema: don't skip deprecation check during validation	7 years ago
Kristin Laemmert	bfd66083de	helper/schema: don't skip deprecation check during validation ... If an attribute was not wholly known, helper/schema was skipping the `validateType` function which (among other things) returned deprecation messages. This PR checks for deprecation before returning when skipping validateType.	7 years ago
Paul Tyng	e400c25358	Minor typo	7 years ago
Paul Tyng	366f70a71e	Update copy for Exists ... You can signal the same information in `Read` with an empty ID if the object does not exist, Implementing `Exists` is not the only way to do so and in some providers is also not the preferred way.	7 years ago
James Bardin	7648f99bca	Merge pull request #22236 from hashicorp/jbardin/nil-in-config ... don't reflect nil in schema validation	7 years ago
James Bardin	016c4f782d	don't reflect nil in schema validation ... Nil values were not previously expected during validation, but they can appear in some situations with the new protocol. Add checks to prevent using zero reflect.Values.	7 years ago
James Bardin	4bed030d40	don't validate MinItems with unknowns in a block ... If there are unknowns, the block may have come from a dynamic declaration, and we can't validate MinItems. Once the blocks are expanded, we will get the full config for validation without any unknown values.	7 years ago
Alex Pilon	7f8f198719	remove UnknownVariabeValue from config and update references to shim	7 years ago
James Bardin	75602df5ef	Revert "Remove removed attribute from applied state" ... This reverts commit 2e2a363052.	7 years ago
James Bardin	8e8eb6d6d3	Merge pull request #21884 from hashicorp/jbardin/validate-null-diags ... don't append nil elements to a diagnostic cty.Path	7 years ago
James Bardin	d4b0788854	don't append nil elements to a diagnostic cty.Path ... If a set element is nil in validateConfigNulls, we don't want to append that element to the diagnostic path, since it doesn't offer any useful info to the user.	7 years ago
James Bardin	0d9f84414a	Merge pull request #21879 from hashicorp/jbardin/sdk-panics ... prevent sdk panics in 2 specific cases	7 years ago
James Bardin	cd3ac50ddb	prevent sdk panics in 2 specific cases ... Fix 2 specific panics in the sdk when reading nil or computed maps from various configurations. The legacy sdk code is too dependent on undefined behavior to attempt to find and fix the root cause at this point. Since the code is essentially frozen for future development, these changes are specifically targeted to only prevent panics from within providers. Because any code effected by these changes would have panicked, there cannot be anything depending on the behavior, and these should be safe to fix.	7 years ago
James Bardin	f362198467	Merge pull request #21806 from hashicorp/jbardin/lost-empty-strings ... prevent an empty string from being lost	7 years ago
James Bardin	9365a2d97d	private and timeout handling in grpc_provider ... Load private data for read, so the resource can get it's configured timeouts if they exist. Ensure PlanResourceChange returns the saved private data when there is an empty diff. Handle the timeout decoding into Meta in the PlanResourceChange, so that it's always there for later operations.	7 years ago
James Bardin	2448d1d38b	move timeout handling to grpc_provider ... simpleDiff is only called from the grpc_provider, and we always need to make sure the timeouts are encoded in the private data.	7 years ago
James Bardin	4dcda04d96	prevent an empty string from being lost ... The helper/schema diff process loses empty strings, causing them to show up as unset (null) during apply. Besides failing to show as set by GetOk, the absence of the value also triggers the schema to insert a default value again during apply. It would also be be preferable if the defaults weren't re-evaluated again during ApplyResourceChange, but that would require a more invasive patch to the field readers, and ensuring the empty string is stored in the plan should block the default.	7 years ago
James Bardin	8ae31aa2db	normalize empty blocks during import ... Like Upgrade, Import is another case where we don't have the context of the configuration and need to ensure missing blocks are normalized.	7 years ago
James Bardin	e6ee78555a	Merge pull request #21721 from hashicorp/jbardin/remove-new-removed ... Remove removed attribute from applied state	7 years ago
James Bardin	814f068dfe	Merge pull request #21711 from hashicorp/jbardin/upgrade-blocks ... Ensure all object attrs & empty blocks in upgrade	7 years ago
James Bardin	dbe22181ae	Ensure all object attrs & empty blocks in upgrade ... When upgrading from a flatmap state, unset blocks would not exist in the state, while they will represented as empty in the new cty.Value. This will cause an unexpected diff in the first plan after upgrade. This situation may normally be applied with no impact, but some providers may have unexpected behavior, and if the attributes force replacement it may require manual alteration of the state to complete the upgrade.	7 years ago