terraform/backend/unparsed_value.go

163 lines
6.7 KiB
Go
Raw Normal View History

terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
package backend
import (
"fmt"
"github.com/hashicorp/hcl/v2"
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
"github.com/hashicorp/terraform/configs"
"github.com/hashicorp/terraform/terraform"
"github.com/hashicorp/terraform/tfdiags"
"github.com/zclconf/go-cty/cty"
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
)
// UnparsedVariableValue represents a variable value provided by the caller
// whose parsing must be deferred until configuration is available.
//
// This exists to allow processing of variable-setting arguments (e.g. in the
// command package) to be separated from parsing (in the backend package).
type UnparsedVariableValue interface {
// ParseVariableValue information in the provided variable configuration
// to parse (if necessary) and return the variable value encapsulated in
// the receiver.
//
// If error diagnostics are returned, the resulting value may be invalid
// or incomplete.
ParseVariableValue(mode configs.VariableParsingMode) (*terraform.InputValue, tfdiags.Diagnostics)
}
// ParseVariableValues processes a map of unparsed variable values by
// correlating each one with the given variable declarations which should
// be from a root module.
//
// The map of unparsed variable values should include variables from all
// possible root module declarations sources such that it is as complete as
// it can possibly be for the current operation. If any declared variables
// are not included in the map, ParseVariableValues will either substitute
// a configured default value or produce an error.
//
// If this function returns without any errors in the diagnostics, the
// resulting input values map is guaranteed to be valid and ready to pass
// to terraform.NewContext. If the diagnostics contains errors, the returned
// InputValues may be incomplete but will include the subset of variables
// that were successfully processed, allowing for careful analysis of the
// partial result.
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
func ParseVariableValues(vv map[string]UnparsedVariableValue, decls map[string]*configs.Variable) (terraform.InputValues, tfdiags.Diagnostics) {
var diags tfdiags.Diagnostics
ret := make(terraform.InputValues, len(vv))
// Currently we're generating only warnings for undeclared variables
// defined in files (see below) but we only want to generate a few warnings
// at a time because existing deployments may have lots of these and
// the result can therefore be overwhelming.
seenUndeclaredInFile := 0
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
for name, rv := range vv {
var mode configs.VariableParsingMode
config, declared := decls[name]
if declared {
mode = config.ParsingMode
} else {
mode = configs.VariableParseLiteral
}
val, valDiags := rv.ParseVariableValue(mode)
diags = diags.Append(valDiags)
if valDiags.HasErrors() {
continue
}
if !declared {
switch val.SourceType {
case terraform.ValueFromConfig, terraform.ValueFromAutoFile, terraform.ValueFromNamedFile:
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
// These source types have source ranges, so we can produce
// a nice error message with good context.
backend: Undeclared variables in -var-file is a warning, not an error In Terraform 0.11 and earlier we just silently ignored undeclared variables in -var-file and the automatically-loaded .tfvars files. This was a bad user experience for anyone who made a typo in a variable name and got no feedback about it, so we made this an error for 0.12. However, several users are now relying on the silent-ignore behavior for automation scenarios where they pass the same .tfvars file to all configurations in their organization and expect Terraform to ignore any settings that are not relevant to a specific configuration. We never intentionally supported that, but we don't want to immediately break that workflow during 0.12 upgrade. As a compromise, then, we'll make this a warning for v0.12.0 that contains a deprecation notice suggesting to move to using environment variables for this "cross-configuration variables" use-case. We don't produce errors for undeclared variables in environment variables, even though that potentially causes the same UX annoyance as ignoring them in vars files, because environment variables are assumed to live in the user's session and this it would be very inconvenient to have to unset such variables when moving between directories. Their "ambientness" makes them a better fit for these automatically-assigned general variable values that may or may not be used by a particular configuration. This can revert to being an error in a future major release, after users have had the opportunity to migrate their automation solutions over to use environment variables. We don't seem to have any tests covering this specific situation right now. That isn't ideal, but this change is so straightforward that it would be relatively expensive to build new targeted test cases for it and so I instead just hand-tested that it is indeed now producing a warning where we were previously producing an error. Hopefully if there is any more substantial work done on this codepath in future that will be our prompt to add some unit tests for this.
2019-01-19 01:37:38 +01:00
//
// This one is a warning for now because there is an existing
// pattern of providing a file containing the superset of
// variables across all configurations in an organization. This
// is deprecated in v0.12.0 because it's more important to give
// feedback to users who make typos. Those using this approach
// should migrate to using environment variables instead before
// this becomes an error in a future major release.
if seenUndeclaredInFile < 3 {
diags = diags.Append(tfdiags.Sourceless(
tfdiags.Warning,
"Value for undeclared variable",
fmt.Sprintf("The root module does not declare a variable named %q but a value was found in file %q. To use this value, add a \"variable\" block to the configuration.\n\nUsing a variables file to set an undeclared variable is deprecated and will become an error in a future release. If you wish to provide certain \"global\" settings to all configurations in your organization, use TF_VAR_... environment variables to set these instead.", name, val.SourceRange.Filename),
))
}
seenUndeclaredInFile++
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
case terraform.ValueFromEnvVar:
// We allow and ignore undeclared names for environment
// variables, because users will often set these globally
// when they are used across many (but not necessarily all)
// configurations.
case terraform.ValueFromCLIArg:
diags = diags.Append(tfdiags.Sourceless(
tfdiags.Error,
"Value for undeclared variable",
fmt.Sprintf("A variable named %q was assigned on the command line, but the root module does not declare a variable of that name. To use this value, add a \"variable\" block to the configuration.", name),
))
default:
// For all other source types we are more vague, but other situations
// don't generally crop up at this layer in practice.
diags = diags.Append(tfdiags.Sourceless(
tfdiags.Error,
"Value for undeclared variable",
fmt.Sprintf("A variable named %q was assigned a value, but the root module does not declare a variable of that name. To use this value, add a \"variable\" block to the configuration.", name),
))
}
continue
}
ret[name] = val
}
if seenUndeclaredInFile >= 3 {
extras := seenUndeclaredInFile - 2
diags = diags.Append(&hcl.Diagnostic{
Severity: hcl.DiagWarning,
Summary: "Values for undeclared variables",
Detail: fmt.Sprintf("In addition to the other similar warnings shown, %d other variable(s) defined without being declared.", extras),
})
}
// By this point we should've gathered all of the required root module
// variables from one of the many possible sources. We'll now populate
// any we haven't gathered as their defaults and fail if any of the
// missing ones are required.
for name, vc := range decls {
if _, defined := ret[name]; defined {
continue
}
if vc.Required() {
diags = diags.Append(&hcl.Diagnostic{
Severity: hcl.DiagError,
Summary: "No value for required variable",
Detail: fmt.Sprintf("The root module input variable %q is not set, and has no default value. Use a -var or -var-file command line argument to provide a value for this variable.", name),
Subject: vc.DeclRange.Ptr(),
})
// We'll include a placeholder value anyway, just so that our
// result is complete for any calling code that wants to cautiously
// analyze it for diagnostic purposes. Since our diagnostics now
// includes an error, normal processing will ignore this result.
ret[name] = &terraform.InputValue{
Value: cty.DynamicVal,
SourceType: terraform.ValueFromConfig,
SourceRange: tfdiags.SourceRangeFromHCL(vc.DeclRange),
}
} else {
ret[name] = &terraform.InputValue{
Value: vc.Default,
SourceType: terraform.ValueFromConfig,
SourceRange: tfdiags.SourceRangeFromHCL(vc.DeclRange),
}
}
}
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 19:33:53 +02:00
return ret, diags
}