24949 Commits (f-workspaces2-prototype)

 

All Branches   

Search

Author	SHA1	Message	Date
Martin Atkins	07f6975ff2	command: Minimal "terraform init" for workspaces2 prototype ... This is just enough to get a functional working directory for a workspaces2 project so that "terraform plan" and "terraform apply" can work.	6 years ago
Martin Atkins	4817a6a9c1	command: Make "terraform workspace select" work under workspaces2 prototype	6 years ago
Martin Atkins	52c1b4e8fc	backend: Minimal support for workspaces2-style operations ... This allows "terraform plan" and "terraform apply" to integrate with the project concepts in a minimal way, showing a rough-and-ready end-to-end of applying multiple inter-dependent workspaces.	6 years ago
Martin Atkins	23c21531f2	command: Make room for project.Workspace objects ... In our existing architecture, workspaces belong to backends and so the command package only deals with workspace _addresses_. With workspaces becoming a project concept, we need to make room for preparing and passing around project.Workspace objects, which requires some renaming to make it clear when we're dealing with actual workspace objects vs. workspace addresses.	6 years ago
Martin Atkins	9a692d45d8	projects: Force local state storage for now ... For the sake of prototyping real operations, for now we'll just force all workspaces to use local state files in a prototype-specific directory. Eventually we should make this customizable using either state_storage or remote configuration, but we don't have the necessary remote client APIs for that yet (existing packages are built around the old backend model) and this temporary solution allows for some forward progress without significant refactoring of the state storage mechanisms.	6 years ago
Martin Atkins	294b9ae045	projects: Load individual workspaces, evaluating their configuration	6 years ago
Martin Atkins	36919a4867	projects: Determine dependencies for a workspace	6 years ago
Martin Atkins	9e340d941f	command: -workspace command line argument ... This is similar to the TF_WORKSPACE environment variable in that it allows overriding the workspace selection that would normally come from a file on disk. The command line argument takes priority over the environment variable, allowing it to be used to override just for a single operation.	6 years ago
Martin Atkins	120c84b1bf	Messy initial plumbing of workspaces into the command package ... This is the crossover point from greenfield development into structural renovation, so things will get messy and awkward from here onwards so that I can patch things out enough to get Terraform to compile without having to complete all of the adaptation work in a single step.	6 years ago
Martin Atkins	6ecdbfaeac	addrs: Compact addresses for local project workspaces	6 years ago
Martin Atkins	b431efea3b	backend/init: Disable all remote state backends ... Under the workspaces2 prototype, "remote state backend" is no longer a concept, so we'll remove these here and then in a later commit adapt these to be remote state storage implementations in providers instead.	6 years ago
Martin Atkins	cc64f5ebe2	backend/atlas: Remove ... This backend is obsolete, because the API it interacts with is EOL.	6 years ago
Martin Atkins	7ba328cb85	project/projectconfigs: Don't panic on dynamic for_each exprs	6 years ago
Martin Atkins	3852814797	projects: Find all the workspaces for a project	6 years ago
Martin Atkins	7b888b00f4	projects: Start of sketching out project config evaluation graph ... This doesn't feel quite right yet, since we don't know the things we need to know here, so this is likely to get revised again in a later commit but this is a checkpoint of the investigation so far.	6 years ago
Martin Atkins	a41d5c6e7f	projects/projectconfigs: Initial static validation of configuration ... This isn't fully comprehensive, but catches some things early on so that we can depend on them being true in any situation where we have a project configuration object.	6 years ago
Martin Atkins	f03ea8b847	addrs: New address types for use in project-level configuration	6 years ago
Martin Atkins	a9ab01e1a7	projects/projectconfigs: Use stable HCL v2.0.0 ... Previously we were using the experimental version from the separate hcl2 repository, but HCL is now stabilized in the main HCL repository as release v2.0.0.	6 years ago
Martin Atkins	332fc6ef31	projects/projectconfigs: decode "upstream" block contents	6 years ago
Martin Atkins	edd27f4ec2	projects/projectconfigs: Decode workspace block contents	6 years ago
Martin Atkins	283c0127e7	projects/projectconfigs: Test parsing of locals	6 years ago
Martin Atkins	aeccfcd1dc	projects/projectconfigs: Decode "context" blocks	6 years ago
Martin Atkins	b9b600a236	projects/projectconfigs: Beginnings of config parsing/decoding ... Finds all of the top-level objects, but still more work to do in order to decode the block bodies.	6 years ago
Martin Atkins	d93192b993	projects/projectconfigs: Find a project root directory given a cwd ... To allow running terraform commands anywhere inside a project, we'll go hunting for a project configuration file in the current working directory and in any parent directories. This is intended to make a Terraform project feel similar to a VCS work tree, where you can generally do VCS operations anywhere within the work tree. Indeed, we expect that in many cases the VCS root directory will _be_ the project root, although that isn't required.	6 years ago
Chris Griggs	74249a8c23	Merge pull request #23235 from hashicorp/cgriggs01-cherryservers ... [Website] CherryServer doc links	6 years ago
cgriggs01	e3b18cd0d9	[Website] CherryServer doc links	6 years ago
James Bardin	cd7c3e4231	update CHANGELOG.md	6 years ago
James Bardin	7829a4f6ce	Merge pull request #23215 from hashicorp/jbardin/for-each-self ... Resource value may be an object in self evaluation	6 years ago
James Bardin	3839405d3d	correctly evaluate self in for_each resources ... The fallback type for GetResource from an EachMap is a cty.Object, because resource schemas may contain dynamically typed attributes. Check for an Object type in the evaluation of self, to use the proper GetAttr method when extracting the value.	6 years ago
James Bardin	f11c836181	failing test with for_each self reference	6 years ago
Martin Atkins	ce68b4d27c	config: Remove legacy interpolation function implementations ... These are often confusing for new contributors, since this looks suspiciously like the right place to add new functions or change the behavior of existing ones. To reduce that confusion, here we remove them entirely from this package (which is now dead code in Terraform 0.12 anyway) and include in the documentation comments a pointer to the current function implementations.	6 years ago
Martin Atkins	f8a32f0b83	website: Consistently recommend the required_providers block ... Previously we were inconsistent in whether we were recommending the new required_providers block or the "version" setting inside a "provider" block.	6 years ago
Martin Atkins	7c110f9cf8	website: provider version constraints in modules	6 years ago
Thomas Alton	ddd0d2a442	website: Provider proxy config only supports alias	6 years ago
Thomas Alton	24386bcfcb	website: Link to `required_providers` block from provider documentation	6 years ago
Colby Rome	d58d91a6b2	website: Fix a missing word in the Module Composition page	6 years ago
Pam Selle	e44800a696	Merge pull request #23209 from pselle/docs-address ... Quote differently so render works on taint docs	6 years ago
Chris Griggs	334c8e9dbb	Merge pull request #23212 from hashicorp/cgriggs01-dome9-links ... [Website] Dome9 provider links	6 years ago
Justin Campbell	07624eeb31	website: Add Terraform Registry Provider Docs page (#23139) ... Co-Authored-By: Kim Ngo <kngo@hashicorp.com> Co-Authored-By: Nick Fagerlund <nick@hashicorp.com>	6 years ago
cgriggs01	7e53919912	[Website] Dome9 provider links	6 years ago
Pam Selle	22ae5cf52d	Quote differently so render works	6 years ago
Chris Griggs	8a34c33ad3	Merge pull request #23183 from hashicorp/cgriggs01-vultr-links ... [Website] Vultr provider links	6 years ago
Pam Selle	5b453f6ba9	Merge pull request #23186 from pselle/pselle/docs-syntax-link ... Docs around splat and for_each	6 years ago
Pam Selle	566f22a34e	Docs around splat	6 years ago
Chris Griggs	7dac6eed82	Update fix	6 years ago
cgriggs01	67105ceeae	[Website] Vultr provider links	6 years ago
James Bardin	744b835e17	Merge pull request #22937 from hashicorp/jbardin/cbd-modules ... create_before_destroy across modules	6 years ago
James Bardin	f766bb8380	prune unused resources from apply ... If a resource is only destroying instances, there is no reason to prepare the state and we can remove the Resource (prepare state) nodes. They normally have pose no issue, but if the instances are being destroyed along with their dependencies, the resource node may fail to evaluate due to the missing dependencies (since destroy happens in the reverse order). These failures were previously blocked by there being a cycle when the destroy nodes were directly attached to the resource nodes.	6 years ago
James Bardin	7108560228	failing test with destroy cycle ... after the previous commit the cycle is broken, but we can't evaluate resource counts that depends on data sources being destroyed.	6 years ago
James Bardin	eb6e7bba2e	do not connect destroy and resource nodes ... Destroy nodes do not need to be connected to the resource (prepare state) node when adding them to the graph. Destroy nodes already have a complete state in the graph (which is being destroyed), any references will be added in the ReferenceTransformer, and the proper connection to the create node will be added in the DestroyEdgeTransformer. Under normal circumstances this makes no difference, as create and destroy nodes always have an dependency, so having the prepare state handled before both only linearizes the operation slightly in the normal destroy-then-create scenario. However if there is a dependency on a resource being replaced in another module, there will be a dependency between the destroy nodes in each module (to complete the destroy ordering), while the resource node will depend on the variable->output->resource chain. If both the destroy and create nodes depend on the resource node, there will be a cycle.	6 years ago