Commit Graph

124 Commits

Author SHA1 Message Date
Mitchell Hashimoto 25d19ef3d0 Merge pull request #10080 from hashicorp/f-tf-version
terraform: support version requirement in configuration
2016-11-14 11:53:30 -08:00
Mitchell Hashimoto df34fa88ce Merge pull request #10076 from hashicorp/f-depend-module
terraform: depends_on can reference entire modules
2016-11-14 11:53:12 -08:00
Mitchell Hashimoto 85d3439fa0
config: parse and validate terraform.required_version 2016-11-12 16:22:35 -08:00
Mitchell Hashimoto 576b61a21d
config: validate depends_on with module values 2016-11-12 08:21:27 -08:00
Mitchell Hashimoto 10426ba619
config: parse depends_on for outputs 2016-11-11 17:46:34 -08:00
Mitchell Hashimoto 29287937e3 Merge pull request #9818 from hashicorp/b-var-parse
config: manually parse variable blocks for better validation
2016-11-04 08:47:40 -07:00
Mitchell Hashimoto ef3148bbfa
config: validate that outputs have a name 2016-11-02 16:56:20 -07:00
Mitchell Hashimoto f054c5ca2c
config: manually parse variable blocks for better validation
Fixes #7846

This changes from using the HCL decoder to manually decoding the
`variable` blocks within the configuration. This gives us a lot more
power to catch validation errors. This PR retains the same tests and
fixes one additional issue (covered by a test) in the case where a
variable has no named assigned.
2016-11-02 14:59:16 -07:00
Mitchell Hashimoto 694b16de5d
config: ignore_changes cannot have interpolations
This is the limitation of all lifecycle attributes currently. Right now,
interpolations are allowed through and the user ends up thinking it
should work. We should give an error.

In the future it should be possible to support some minimal set of
interpolations (static variables, data sources even perhaps) but for now
let's validate that this doesn't work.
2016-10-24 23:06:33 -07:00
Mitchell Hashimoto 609219fc65 command/meta: validate config immediately
* config: test for validating multi-vars (passes)

* command/plan: test invalid run

* command/meta: validate module on load
2016-09-03 15:26:49 -07:00
James Bardin 94674fe93c Add a test load of a data source with count 2016-09-03 13:08:41 -07:00
Sander van Harmelen 47dd1ad153 Add wildcard (match all) support to ignore_changes (#8599) 2016-09-02 15:44:35 +02:00
Mitchell Hashimoto 706b2e2aea Merge pull request #8482 from hashicorp/b-output-dup
config: variable names and outputs must be unique
2016-08-26 13:57:59 -07:00
Mitchell Hashimoto 0fceeaaeb0
config: test for var uniqueness in overrides 2016-08-26 13:48:21 -07:00
Mitchell Hashimoto 8a3559560d
config: JSON resource keys with only one item load properly GH-5140
When a resource has only a single key set, the HCL parser treats that
key as part of the overall set of object keys. This isn't valid since
we expect resources to have exactly two keys. In this scenario, we have
to "unwrap" the keys back into a set of objects.
2016-08-25 17:18:18 -07:00
Mitchell Hashimoto fbf06e2a59
config: vars must be unique 2016-08-25 14:51:49 -07:00
Mitchell Hashimoto 099293b690
config: outputs must be unique 2016-08-25 14:43:57 -07:00
Mitchell Hashimoto f4faf2274b
config: count can't be a SimpleVariable 2016-08-16 13:48:12 -07:00
Paul Hinze ffa29090ec
core: Better error for dot indexing on user vars
Dot indexing worked in the "regexps and strings" world of 0.6.x, but it
no longer works on the 0.7 series w/ proper List / Map types.

There is plenty of dot-indexed config out in the wild, so we need to do
what we can to point users to the new syntax.

Here is one place we can do it for user variables (`var.somemap`). We'll
also need to address Resource Variables and Module Variables in a
separate PR.

This fixes the panic in #7103 - a proper error message is now returned.
2016-06-12 10:45:48 -05:00
James Nugent 01cd596c60 core: Fix detection of empty list/map defaults
This commit changes config parsing from weak decoding lists and maps
into []string and map[string]string respectively to decode into
[]interface{} and map[string]interface{} respectively. This is in order
to take advantage of the work integrated in #7082 to defeat the backward
compatibility features of the mapstructure library.

Test coverage of loading empty variables and validating their default
types against expectation.
2016-06-12 11:19:03 +02:00
Martin Atkins 860140074f config: Data source loading
This allows the config loader to read "data" blocks from the config and
turn them into DataSource objects.

This just reads the data from the config file. It doesn't validate the
data nor do anything useful with it.
2016-05-14 08:26:35 -07:00
James Nugent f49583d25a core: support native list variables in config
This commit adds support for native list variables and outputs, building
up on the previous change to state. Interpolation functions now return
native lists in preference to StringList.

List variables are defined like this:

variable "test" {
    # This can also be inferred
    type = "list"
    default = ["Hello", "World"]
}

output "test_out" {
    value = "${var.a_list}"
}
This results in the following state:

```
...
            "outputs": {
                "test_out": [
                    "hello",
                    "world"
                ]
            },
...
```

And the result of terraform output is as follows:

```
$ terraform output
test_out = [
  hello
  world
]
```

Using the output name, an xargs-friendly representation is output:

```
$ terraform output test_out
hello
world
```

The output command also supports indexing into the list (with
appropriate range checking and no wrapping):

```
$ terraform output test_out 1
world
```

Along with maps, list outputs from one module may be passed as variables
into another, removing the need for the `join(",", var.list_as_string)`
and `split(",", var.list_as_string)` which was previously necessary in
Terraform configuration.

This commit also updates the tests and implementations of built-in
interpolation functions to take and return native lists where
appropriate.

A backwards compatibility note: previously the concat interpolation
function was capable of concatenating either strings or lists. The
strings use case was deprectated a long time ago but still remained.
Because we cannot return `ast.TypeAny` from an interpolation function,
this use case is no longer supported for strings - `concat` is only
capable of concatenating lists. This should not be a huge issue - the
type checker picks up incorrect parameters, and the native HIL string
concatenation - or the `join` function - can be used to replicate the
missing behaviour.
2016-05-10 14:49:14 -04:00
Paul Hinze 567a9b9e06 config: remove missing equals test to fix build
This is behavior that's covered in the parser now - and the error
message is nicer to boot!
2016-03-21 10:39:20 -05:00
Paul Hinze 3f72837f4b core: Make copies when creating destroy nodes
Fixes an interpolation race that was occurring when a tainted destroy
node and a primary destroy node both tried to interpolate a computed
count in their config. Since they were sharing a pointer to the _same_
config, depending on how the race played out one of them could catch the
config uninterpolated and would then throw a syntax error.

The `Copy()` tree implemented for this fix can probably be used
elsewhere - basically we should copy the config whenever we drop nodes
into the graph - but for now I'm just applying it to the place that
fixes this bug.

Fixes #4982 - Includes a test covering that race condition.
2016-02-09 09:25:16 -06:00
James Nugent cb6cb8b96a core: Support explicit variable type declaration
This commit adds support for declaring variable types in Terraform
configuration. Historically, the type has been inferred from the default
value, defaulting to string if no default was supplied. This has caused
users to devise workarounds if they wanted to declare a map but provide
values from a .tfvars file (for example).

The new syntax adds the "type" key to variable blocks:

```
variable "i_am_a_string" {
    type = "string"
}

variable "i_am_a_map" {
    type = "map"
}
```

This commit does _not_ extend the type system to include bools, integers
or floats - the only two types available are maps and strings.

Validation is performed if a default value is provided in order to
ensure that the default value type matches the declared type.

In the case that a type is not declared, the old logic is used for
determining the type. This allows backwards compatiblity with previous
Terraform configuration.
2016-01-24 11:40:02 -06:00
Paul Hinze 87a9701f91 config: validation error when output is missing value field
Also lists out invalid keys in errmsg when they are present

Closes #4398
2016-01-20 14:00:36 -06:00
Mitchell Hashimoto 99fbb91ba2 config: validate lifecycle keys [GH-4413] 2016-01-19 11:28:45 -08:00
Paul Hinze b6626eed57 config: friendlier error message on resource arity mismatch
closes #2072
2015-12-09 18:05:49 -06:00
James Nugent 5ea25363a1 Add regression test for #4069
This may be brittle as it makes use of .gitattributes to override the
autocrlf setting in order to have an input file with Windows line
endings across multiple platforms.
2015-12-01 13:37:18 -05:00
Paul Hinze d90eb2d88e config: test replicating #4079
Should help cover terraform against regression once
https://github.com/hashicorp/hcl/pull/70 lands.
2015-12-01 10:31:05 -06:00
James Nugent 7f5f8d300d Add failing test replicating #4065 2015-11-26 15:08:48 +02:00
Paul Hinze afb5136ac2 Merge pull request #3986 from hashicorp/phinze/hcl-escaped-quotes
config: test covering escaped quotes syntax error
2015-11-19 12:32:30 -06:00
Paul Hinze 15e7927009 config: test covering escaped quotes syntax error
This was never intended to be valid syntax, but it worked in the old HCL
parser, and we've found a decent number of examples of it in the wild.

Fixed in https://github.com/hashicorp/hcl/pull/62 and we'll keep this
test in Terraform to cover the behavior.
2015-11-19 12:11:42 -06:00
James Nugent 6ae3218f8a Add failing tests for JSON configuration parsing
Reproduces the issue reported by @svanharmelen in #3964.
2015-11-19 16:06:30 +02:00
James Nugent f4164b5322 Add resource with heredoc to config load tests
This test reproduces the issue which is likely the root cause of #3840.
Test is currently failing with an "illegal character" message
corresponding with the location of the heredoc, which is also seen in
various acceptance tests for providers.
2015-11-10 18:12:21 -05:00
Rob Zienert a1939e70f7 Adding ignore_changes lifecycle meta property 2015-10-14 16:34:27 -05:00
Mitchell Hashimoto aa1e66c16c config: parse lifecycle block with mapstructure for weak decode 2015-06-07 22:04:23 -07:00
Mitchell Hashimoto cca4964552 config: error in validation if module has self variable 2015-04-22 10:39:07 +02:00
Mitchell Hashimoto d0a6d78b97 config: test that resources reference good providers 2015-04-20 14:47:31 -07:00
Mitchell Hashimoto 8ee18e2e14 config: happy case test for multiple providers 2015-04-20 14:27:44 -07:00
Mitchell Hashimoto a599d5f224 config: validate that a multi provider is only configured once each 2015-04-20 14:25:33 -07:00
Paul Hinze 975a96f271 core: protect against count.index in modules
Modules should get a validation error just like outputs do.

refs #1528
2015-04-15 10:41:56 -05:00
Paul Hinze 347690a73e core: don't crash when count.index is used in the wrong context
It's bad manners! :)

Also adds a validation error up at the configuration layer so the user
sees the case from #1528 as an error message.

fixes #1528
2015-04-15 10:23:53 -05:00
Mitchell Hashimoto fdded8ca14 config: allow atlas block 2015-03-24 13:30:22 -07:00
Radek Simko f794f30b7d Ignore hidden files per Unix conventions 2015-03-18 07:50:33 +00:00
Mitchell Hashimoto 965fe45b9e config: self var validation 2015-02-23 14:43:14 -08:00
Mitchell Hashimoto a31f2a276b Merge pull request #1015 from hashicorp/b-depends-on-var
config: depends on cannot contain interpolations [GH-985]
2015-02-23 13:49:34 -08:00
Mitchell Hashimoto c14e84a657 config: validate provisioner splats can only reference others 2015-02-20 09:21:29 -08:00
Mitchell Hashimoto f156d0d1bd config: test we can ref splat of other resources 2015-02-20 09:19:13 -08:00
Mitchell Hashimoto 90a6a627ed config: validate configuration doens't contain splats to ourselves 2015-02-20 09:18:08 -08:00