core: Fixed variables not being in scope for destroy -target on modules #9021
Conversation
| } | ||
|
|
||
| h := new(HookRecordApplyOrder) | ||
| h.Active = true |
There was a problem hiding this comment.
Looks like this Hook is not used in this test.
| Targets: []string{"module.child"}, | ||
| }) | ||
|
|
||
| // First plan and apply a create operation |
There was a problem hiding this comment.
This comment isn't applicable here anymore, since this is the destroy.
| Providers: map[string]ResourceProviderFactory{ | ||
| "aws": testProviderFuncFixed(p), | ||
| }, | ||
| }) |
There was a problem hiding this comment.
The above WritePlan / ReadPlan dance is meant to emulate writing out to a tfplan file. It's unnecessary here, since the bug reproduces w/o it.
| @@ -86,6 +86,14 @@ func (t *TargetsTransformer) selectTargetedNodes( | |||
| var err error | |||
| if t.Destroy { | |||
| deps, err = g.Descendents(v) | |||
|
|
|||
| // Select any variables that we depend on in case we need them later for interpolating in count | |||
There was a problem hiding this comment.
Style nit: wrap comment at 80 columns.
kyhavlov
force-pushed
the
destroy_target_fix
branch
from
4c4178e
to
3a2819d
Compare
|
@phinze addressed the comments and cleaned up the test. I was going add a test for the nested module case, but I hit the issue with Terraform not including child modules of a module when using -target, so that will have to be another PR. |
|
LGTM |
|
Got the error this was supposed to fix on 0.7.4. Upgraded to 0.7.6. Now I get the following when doing |
|
I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues. If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further. |
ghost
locked and limited conversation to collaborators
Addresses the bug in #8146. The call to
selectTargetedNodes()was only returning the resource and not the variable that it depended on to interpolate its count, so the variable got removed from the graph and wasn't in scope for later interpolations, causing theunknown variable accessederror.@phinze