32532 Commits (eecbfbf2cccedfb581469ee63599c4b4abc5c60f)

 

All Branches   

Search

Author	SHA1	Message	Date
Daniel Schmidt	a8c1229720	Merge pull request #34677 from hashicorp/improve-plan-for-lists	2 years ago
Daniel Schmidt	adf008d489	adjust tests to new implementation	2 years ago
Daniel Schmidt	19f9bdac14	handle multi-line string diffs correctly	2 years ago
Daniel Schmidt	a64df71152	fix comment	2 years ago
Daniel Schmidt	1531fee472	clarify comment	2 years ago
Daniel Schmidt	2a625b75a4	handle reordering explicitly	2 years ago
Daniel Schmidt	31bd2b5be0	plan: compare individual elements if list is same length ... Closes #33779	2 years ago
Liam Cervante	64b311d53a	stacks: apply sensitive marks to outputs from components (#34684) ... * stacks: apply sensitive marks to outputs from components * gen headers * ensure consistent test ordering	2 years ago
Liam Cervante	f0b61090a2	Move the testing MockProvider into a dedicated shared package (#34681) ... * move the testing MockProvider into a dedicated shared package * address comments * fix new changes	2 years ago
Alisdair McDiarmid	06723d3d4b	stackeval: Add support for replace actions to hook ... The terraform.Hook implementation in stackeval is used to track the operations performed during stack runtime operations, for later reporting to the caller. This hook did not correctly support the replace actions (create-then-destroy, destroy-then-create), resulting in a loss of data between plan and apply. This exhibited as a plan with a replace reporting 1 add and 1 remove operation, which when applied would report only 1 add operation. Previously, we stored the action performed for each apply operation on a given resource instance in the PreApply hook, then allowed access to it via the ResourceInstanceObjectAppliedAction method. Here we extend the PostApply hook to look for a previous action performed on this instance, and use that to reconstruct the planned replace action. This method is called in the stackruntime package, where replace operations are counted for both add and remove. No changes are needed at the call site to fix the bug.	2 years ago
Martin Atkins	fde6277565	stackruntime: Pass ExperimentsAllowed into stackeval for apply ... This was originally part of 7dad938fdb but unfortunately seems to have got lost during some rebasing, or some other similar sort of annoying reason. This now allows the "experiments allowed" flag to propagate into the stackeval package when we're running the apply phase, for consistency with all of the other phases. Without this, it's possible to plan a configuration that's participating in experiments, but then it fails in a strange way during the apply step due to Terraform suddenly thinking it's a stable release where experiments are disabled.	2 years ago
James Bardin	084bce65dc	Merge pull request #34683 from hashicorp/jbardin/detect-unknown-provider-functions ... Better diagnostics for unknown provider functions	2 years ago
James Bardin	336145c384	update diagnostic message	2 years ago
James Bardin	bac2f892c6	more helpful provider function diagnostics ... Use the new FunctionCallUnknownDiagExtra feature from hcl to help guide users through problems with provider function calls. Now that we can detect diagnostics from unknown function calls, we can correlate the namespace and names with certain problems specific to Terraform.	2 years ago
James Bardin	bc3c185f2e	update hcl for function diags	2 years ago
Radek Simko	82eca73994	gitignore: Add go.work{.sum} (#34682)	2 years ago
Liam Cervante	07f6621091	stacks: include resources in state when calculating required providers (#34645) ... * stacks: include resources in state when calculating required providers * also support apply time * add copywrite headers	2 years ago
Martin Atkins	67d5d8f79f	Update CHANGELOG.md ... Cleanup after today's v1.8.0-alpha release	2 years ago
CJ Horton	7a37c393a0	Merge pull request #34678 from hashicorp/radditude/standard-panics ... standardize panic output	2 years ago
CJ Horton	6e1e00ae3f	standardize panic output ... Programs that monitor Terraform's output to report panics might make the reasonable assumption that the string "panic:" is always included and is therefore safe to monitor for. Our custom panic output unfortunately breaks these assumptions at the moment. Instead of asking consumers to add their own handling to deal with this problem, let's add that greppable string to our custom panic output.	2 years ago
James Bardin	0d712f99df	Prepare for 1.8.0-alpha20240131 release	2 years ago
Radek Simko	0c509183d1	ci: Add test for license headers (#34624) ... * ci: Add test for copyright headers * replace GHA workflow with make target (copyright) This makes it reusable for outside of CI, such that anyone cloning the repo locally can still easily verify compliance before raising a PR. * remove go:generate for copywrite This should no longer be needed as we leverage copywrite via a shell script now. * introduce make copyrightfix	2 years ago
Martin Atkins	a77d351ea2	plugin+plugin6: Don't panic if ProviderMeta not populated ... "ProviderMeta" is a niche feature that's used by only two providers, but this code was written under the assumption that the caller would always provide it for any provider that has a schema for it, and unfortunately the Terraform SDK seems to always provide a meta schema even for providers that don't use it, and thus it's empty. The current phase of the "unknown_instances" language experiment is not fully wired in to the main logic as a way to reduce the risk of it impacting behavior for those not participating in the experiment, but that means that right now it doesn't actually have a resolved ProviderMeta value to present, and so was just omitting it on the assumption that it is always optional to provide anyway. As a pragmatic stopgap to resolve that conflict, this makes the gRPC provider dispatching logic tolerate an absent ProviderMeta value and synthesize a null value to use in that case. This makes that logic slightly more robust and also means we can defer doing all of the work to weave ProviderMeta into this experimental codepath. The new TODO comment in that codepath is intended to remind us to consider this again should we decide to stablize the experiment later; when we do that, we'll hopefully do it by refactoring to share more code between the two codepaths, at which point ProviderMeta will be easier to support.	2 years ago
Craig Wright	c02c855423	Merge pull request #34635 from hashicorp/crw-patch-1 ... Update pull_request_template.md	2 years ago
Bruno Schaatsbergen	c6ee2652a0	chore: bump golang.org/xoauth2 to v0.17.0	2 years ago
Liam Cervante	58289db4a9	terraform test: use platform independent path functions (#34666)	2 years ago
Martin Feineis	e5db180534	double quote typo ... Fix missing double quote	2 years ago
Martin Atkins	e6665c4f43	stackeval: Basic support for deferred-change propagation ... This is the bare minimum functionality to ensure that we defer all actions in any component that depends on a component that already had deferred actions. We will also eventually need to propagate out a signal to the caller for whether the stack plan as a whole is complete or incomplete, but we'll save that for later commits, since the stack orchestration in Terraform Cloud will do the right thing regardless, aside from the cosmetic concern that it won't yet know to show a message to the user saying that there are deferred changes.	2 years ago
Martin Atkins	2e6cf0a518	terraform: Callers can force all actions to be deferred ... When the modules runtime is being used inside the stacks runtime, it's possible that a component could refer to another component that has deferred changes in its plan. In that case, we do still want to plan the downstream component (to give earlier feedback if there are obvious problems with it) but we need to force all planned actions to be treated as deferred so that we preserve the correct dependency ordering across all objects described in a stack. This commit only deals with the modules runtime handling that case. We'll make the stacks runtime use it in a subsequent commit.	2 years ago
Martin Atkins	823aa8e4f3	terraform: Handle partial-expanded resources during planning ... When at least one module in the configuration is opted in to the unknown_instances language experiment, we'll now divert into an experimental adaptation of nodeExpandPlannableResource.DynamicExpand which knows how to deal with partial-expanded modules and resources. Whenever a partial-expanded resource prefix is detected, we'll make a special graph node for it which goes through some of the motions of planning but does so only really for the side-effect of checking the configuration for dynamic errors. We do also save the result as a placeholder value for the object for completeness, although it's not really used for much yet because the partial-expansion tracker saves this primarily for reporting information about the deferred changes as extra context in the plan, and we're not implementing that yet. This isn't yet quite right because we're not preserving enough information from plan phase to apply phase to correctly handle references to the resources that had deferred actions. However, since this is currently just an experiment anyway, this is a useful interim milestone that shows at least that things are functional enough to converge on the correct result in the end. We'll correct the evaluation misses (as documented in the TestContext2Apply_deferredActionsResourceForEach test) in subsequent commits, which will require tracking a little more information in the saved plan so that we can re-populate all of the deferral metadata when we start the apply phase.	2 years ago
Martin Atkins	a0b0996ad1	terraform: Evaluator returns DynamicVal for partial-expanded resource ... For now this is just an up-top weird exception to minimize the impact on the code that's runs when not participating in the unknown_instances experiment. In future it would be nice to rework this GetResource method to rely on the ResourceInstanceKeys result even in the known case, since that will simplify this function and remove a some duplicate code, but our priority for now is minimizing the possibility that the code for this experiment can impact users who are not participating in the experiment.	2 years ago
Martin Atkins	64f7b6992f	terraform: BuiltinEvalContext function eval in partial-expanded module ... An earlier commit changed the EvalContext model to support evaluation in both fully-expanded and partial-expanded modules, but the provider-contributed functions feature landed semi-concurrently with it and so inadvertently introduced a codepath that only worked in the fully-expanded case. This will now handle both situations, since all we really need is the addrs.Module, which we can obtain in both modes.	2 years ago
Martin Atkins	d574082caf	terraform: nodeExpandModule must check experiments on moduleCtx ... Previously we were incorrectly checking for the experiment being enabled on the global EvalContext, but that doesn't make sense because experiments are explicitly module-specific. Now we'll ask the question separately for each module so we can use the module-scoped EvalContext.	2 years ago
Martin Atkins	3e2b6ffec8	core: Resource nodes aware when unknown_instances experiment is on ... Usually language experiments have single-module scope, but this particular one can potentially force modules other than the one that opt in to do some behavior differently for correct results. To minimize the risk of the new experiment code impacting those who are not participating in the experiment, we'll want to branch into a new DynamicExpand codepath only when at least one module is participating in the experiment, but graph nodes alone have a more tightly-scoped view that prevents them from answering that question directly, and so we'll temporarily ask the ConfigTransformer to help in detecting that up top and propagating the result into all of the resource nodes. This commit doesn't yet make use of the new temporary field; that'll follow in a subsequent commit that introduces the new partial-expansion-aware alternative DynamicExpand codepath.	2 years ago
Martin Atkins	47fd5e3e35	terraform: Expose a shared deferring.Deferred object ... As with the various other context-tracking containers we use during graph walks, we'll need a deferring.Deferred object shared across all graph nodes' EvalContexts and with the expression evaluator so that we can report deferrals and then make decisions based on deferrals previously registered upstream. As of this commit we're not yet doing anything with this new object, but in future commits we'll use it to report what we've deferred and to "artifically" defer anything that has a side-effect and depends on something else that's already been deferred.	2 years ago
Martin Atkins	db07b905e9	deferring: A new package for tracking deferred actions ... When we defer planning actions that have direct side-effects -- that is, resource instance actions -- we need to remember that we did so because we then need to force deferring planning for any other resources downstream to preserve the required order of operations. That means we need to keep track of what we've already deferred so we can check whether a newly-visited resource instance must also be deferred. We also need to track deferrals for the purpose of reporting them as part of the plan presented to the end-user. This change _starts_ to implement that, but for now the public API is limited only to generating a boolean signal for whether there are any deferrals at all, because that's the minimum possible reporting that allows for correct behavior (so the caller can treat the resulting plan slight differently in the workflow). In future we'll also want to be able to return placeholder values for not-yet-expanded resource instances, but that's not fully built out here yet since for an initial implementation we can cheat and use cty.DynamicVal as a universal "we know nothing at all" placeholder. This API is likely to grow and change in forthcoming commits as we continue wiring these new behaviors into the Terraform modules runtime.	2 years ago
Martin Atkins	53f00667d5	instances: Some additional commentary ... The internal API here was a little unclear, so this is just some retroactive additional documentation resulting from me having "relearned" how this API works after some time away from it, in the hope that future maintainers will not need to repeat that exercise.	2 years ago
Martin Atkins	8e02f5dd21	addrs: PartialExpandedResource module address accessors ... This is kinda awkward because this address type represents both resources whose own instances are not expanded yet and resources belonging to whole modules whose instance keys haven't been expanded yet, and those two cases have different address types. However, in return for this awkward API only for the rare case where we need to isolate the module instance address, the rest of the system gets to not worry very much about this distinction: it can share most code between the two cases, since they both need similar evaluation treatment anyway.	2 years ago
Martin Atkins	67d8a5137b	rpcapi+stacks: Stacks runtime can see whether experiments are allowed ... We allow experiments only in alpha builds, and so we propagate the flag for whether that's allowed in from "package main". We previously had that plumbed in only as far as the rpcapi startup. This plumbs the flag all the way into package stackeval so that we can in turn propagate it to Terraform's module config loader, which is ultimately the one responsible for ensuring that language experiments can be enabled only when the flag is set. Therefore it will now be possible to opt in to language experiments in modules that are used in stack components.	2 years ago
Martin Atkins	73df33fe6e	instances: Fix data race in Expander.ExpandAbsModuleCall ... This method must hold a read lock while it does its work, or else it will race with calls to register the expansion mode for other module calls.	2 years ago
Martin Atkins	e6ce1edf22	logging: Send captured panics to real stderr ... Our goal with this panic-interception was to largely mimic how the Go runtime would normally report panics except for two intentional exceptions: an extra prompt explaining to the user that Terraform crashed, and exiting with status code 11 instead of 2. Unfortunately we accidentally deviated in a different way: we're reporting to whatever os.Stderr happens to refer to, instead of to the real process stderr. It seems like that shouldn't really matter, but unfortunately go-plugin intentionally changes os.Stderr to refer to a totally separate stream that it manages, causing the captured panic messages to be routed over a grpc-based channel to the plugin client. This deviation makes the panic messages not visible to usual strategies for trying to heuristically detect that a Go program has panicked. Without a special interception like Terraform is doing here, the Go runtime writes directly to the stderr file descriptor without going through the os.Stderr abstraction, and so to achieve consistent behavior we need to do a little hoop-jumping to approximate that result. In particular, this makes the behavior now consistent with what happens when a provider plugin running as a child of Terraform Core panics, and so a system which tries to sniff stderr for content that seems like a panic message will be able to handle both situations equally and avoid making a special case for Terraform Core/CLI's own panics.	2 years ago
Ansgar Mertens	cd1306430b	fix(docs): Include product in relative link (#34510) ... CDKTF uses these definitions to generate docs which end up on developer.hashicorp.com There are content checks that fail for this link as the product (terraform) is required to be explicitly stated now	2 years ago
Eloi Perdereau	30505dd1bf	website docs: fix "Block Representation" for `provider schema -json` (#34653) ... * website docs: fix "Block Representation" for `provider schema -json` * Update website/docs/cli/commands/providers/schema.mdx --------- Co-authored-by: Radek Simko <radeksimko@users.noreply.github.com>	2 years ago
Liam Cervante	3475d22741	Fix nil pointer exception when planning stacks with undeclared variables (#34638) ... * fix nil pointer exception when planning with undeclared variables * another function does the same	2 years ago
Alisdair McDiarmid	9255db67b3	Merge pull request #34648 from hashicorp/alisdair/unmark-stack-provider-config ... stackeval: Unmark provider config for plugin RPC	2 years ago
Daniel Schmidt	3211068f10	Merge pull request #34644 from hashicorp/document-provider-lock-using-cache ... document `-enable-plugin-cache` flag	2 years ago
Daniel Schmidt	18796d0dea	chore: clarify wording ... Co-authored-by: Liam Cervante <liam.cervante@hashicorp.com>	2 years ago
Daniel Schmidt	cb46c9937a	Merge pull request #34658 from hashicorp/fix-is-sensitive-mdx ... chore: remove trailing whitespace from file path	2 years ago
Daniel Schmidt	7d88f79aaf	chore: add navigation entry	2 years ago
Daniel Schmidt	ae56655afe	chore: remove trailing whitespace from file path	2 years ago