Non-HTTP errors previously resulted in a panic due to dereferencing the
resp pointer while it was nil, as part of rendering the error message.
This commit changes the error message formatting to cope with a nil
response, and extends test coverage.
Fixes#24384
* terraform: large refactor to use Provider from configs.Resource
configs.Resource.ImpliedProvider() now returns a string; it is the
callers' responsibility to turn that into an addrs.Provider if needed.
GraphNodeProviderConsumer ProvidedBy() no longer returns nil (reverting
to earlier, pre-provider-fqn behavior): it will return either the
provider set in config, provider set in state, or the default provider.
It seems that the checksum for v3.3.10+incompatible has changed at some
point, causing "go mod vendor" to fail now.
We can see by the fact that no files within "vendor" have changed that the
change in checksum is not the result of any material change in the module
code, and therefore presumably resulted from some change in metadata or
a change in the Go module hashing algorithm since Go 1.12.
Make the interface name reflect the new return type of the method.
Remove the confusingly named and unused ResourceAddress method from the
resource nodes as well.
The ModuleInstance is known while building the state resource, but it's
not recorded. Since a resource may be retrieved via a ConfigResource
address, we need to know from which module instance it was loaded.
Core needs a way to address resources through unexpanded modules, as
they are present in the configuration. There are already some cases of
paring `addrs.Module` with `addrs.Resource` for this purpose, but it is
going to be helpful to have a single type to describe that pair, as
well as have the ability to use TargetContains.
Remove unused variables, sync.Once, and init in BuiltinEvalContext.
Replace some shim calls.
GraphNodeAttachProvider doesn't need to be a NodeModuleInstance.
This function can already produce suitable diagnostic messages which we'd
like to preserve, but it cannot produce source location information, and
so we'll amend the diagnostics to include that on the way out while
retaining all of the other values in the diagnostics.
Most of the time we're converting from HCL diagnostics to tfdiags as we
expose diagnostics directly from HCL, but occasionally we need to to the
reverse.
For example, our configs package uses hcl.Diagnostics by convention
because it's primarily working with HCL, but sometimes it interacts with
functions elsewhere (like in the "addrs" package) that return
tfdiags.Diagnostics, where they need to be adapted to return in an HCL
shape.
This should be used with some care because, similar to Diagnostics.ForRPC,
it forces immediate flattening of all of the diagnostics to a single
type and so can potentially lose internal tracking information that
appears in other tfdiags.Diagnostic information, such as the additional
metadata tracked in the ConsolidateWarnings result to allow later
appending to existing groups.
* configs: parse provider source string during module merge
This was the smallest unit of work needed to start writing provider
source tests!
* Update configs/parser_test.go
Co-Authored-By: Alisdair McDiarmid <alisdair@users.noreply.github.com>
Change ModuleInstance to Module in AbsProviderConfig, because providers
need to be handled before module expansion, and should not be used
defined inside an expanded module at all.
Renaming of the addrs type can happen later, when there's less work
in-flight around provider configuration.
Remove the requirement for most *Resource types to be a
GraphNodeModuleInstance, ensuring that they never call ctx.Path while
being evaluated. This gets rid of most of the direct need for the Path
method currently implemented by NodeResourceAbstract, leaving the
provider and schema related calls for a subsequent PR.
This adds more shimming into that node itself, but allows us to pull it
out of the config transformer, and ensure we can create the resources
correctly from the config. The shimmed address usage can then be raised
out of the abstract resource, into the expanded node types.
Using this in the same manner as NodePlannableOutput, which expands the
local values within modules. All thee output and local types are used in
both plan and apply, we may rename these to better reflect their usage
in expanding. That wait until we are certain that apply won't need any
extra machinery for handling values that aren't stored in the plan.
Unexpanded nodes can't implement GraphNodeModuleInstance (nee
GraphNodeSubPath), because they aren't aware how they have been
expanded, and may be in multiple distinct paths.
Since that means the EvalContext won't be in the correct path during the
walk, we just have to ensure that we don't use `ctx.Path()` inside Eval.
Alisdair McDiarmid