fixes #1521.

based on PR feedback from @eladb

rather than the DomainName value

Fixes #1483.

Small change to enable tagging KMS keys.

The `cdk diff` command intentionally returns non-zero when a difference is found,
similar to how the POSIX `diff` command behaves (returning 0 when no difference is
found, 1 when some difference is found, and >1 when an error occurred). This
behavior was however undocumented, possibly leading to user confusion.

This change adds a note in `cdk --help` that mentions the exit code 1 will be used
when a difference is found.

Fixes #1440

Since #973, the renames of CFN properties from "name" to "xxxName" were
removed. But the ELBv2 library still used `loadBalancerName`. The reason
this was not discovered was because we are splatting `additionalProps` of
type `any`, and this caused the type checker to stop verifying property
names.

Fixes #1481

Use a unified diff format to render differences in arbitrary values,
making it easier to understand what is changing in possibly large JSON
structures, for example.

The number of context lines used when rendering the JSON differences
can be customized using the `--context-lines` option of `cdk diff`, 
which has a default value of `3`.

Adds support for CNAME records via the `CnameRecord` construct.

Misc:

- `TXTRecord` was renamed to `TxtRecord`.
- `hostedZoneNameServers` attribute added to IHostedZone
- Made `HostedZone` a concrete (non-abstract) class so it will be
  compatible with how CloudFormation represents this resource,
  but left PublicHostedZone and PrivateHostedZone to allow
  a more strongly-typed experience if you like.

Credits for original PR (closes #1420): @MikeBild

BREAKING CHANGE: The `route53.TXTRecord` class was renamed to `route53.TxtRecord`.

Now that we can represent attributes as native strings and string lists,
this covers the majority of resource attributes in CloudFormation.

This change:

* String attributes are represented as `string` (like before).
* String list attribute are now represented as `string[]`.
* Any other attribute types are represented as `cdk.Token`.

Attributes that are represented as Tokens as of this change:

* amazonmq has a bunch of `Integer` attributes (will be solved by #1455)
* iot1click has a bunch of `Boolean` attributes
* cloudformation has a JSON attribute
* That's all.

A few improvements to cfn2ts:

* Auto-detect cfn2ts scope from package.json so it is more self-contained
  and doesn't rely on cdk-build-tools to run.
* Added a "cfn2ts" npm script to all modules so it is now possible
  to regenerate all L1 via "lerna run cfn2ts".
* Removed the premature optimization for avoiding code regeneration
  (it saved about 0.5ms).

Fixes #1406

BREAKING CHANGE: any `CfnXxx` resource attributes that represented a list of strings are now typed as `string[]`s (via #1144). Attributes that represent strings, are still typed as `string` (#712) and all other attribute types are represented as `cdk.Token`.

BREAKING CHANGE: the deprecated `cloudformation.XxxResource` classes have
been removed. Use the `CfnXxx` classes instead.

Change `conditionXx` methods to accept an interface `IConditionExpression`
instead of concrete class and implement this interface by the `Condition`
construct.

This enables chaining conditions like so:

    const cond1, cond2, cond, cond4 = new cdk.Condition(...)
    Fn.conditionOr(cond1, cond2, Fn.conditionAnd(cond3, cond4))

Fixes #1457

AWS Step Functions released support for 'Parameters' option in the input definition.

Closes #1480

It is now no longer necessary to use `export()` and `import()` when
sharing constructs between two `Stacks` inside the same CDK app.
Instead, objects defined in one stack can be used directly in another
stack.

The CDK will detect when an attribute (such as an ARN, ID or URL) of
such an object is used in a different stack, and will automatically
create the required `Export` in the producing stack and insert the
corresponding `Fn::ImportValue` in the consuming stack.

BREAKING CHANGE: if you are using `export()` and `import()` to share
constructs between stacks, you can stop doing that, instead of
`FooImportProps` accept an `IFoo` directly on the consuming stack,
and use that object as usual. `ArnUtils.fromComponents()` and
`ArnUtils.parse()` have been moved onto `Stack`. All CloudFormation
pseudo-parameter (such as `AWS::AccountId` etc) are now also
accessible via `Stack`, as `stack.accountId` etc. `resolve()` has
been moved to `this.node.resolve()`. `CloudFormationJSON.stringify()`
has been moved to `this.node.stringifyJson()`. `validate()` now
should be `protected`.

Fixes #1324.

In order to allow passing in an imported role to a Lambda/Alias,
we must use `IRole` instead of `Role`.

The "toCloudFormation" method of generated CFN resources invoke
"resolve" so that any lazy tokens are evaluated. This escaped the
global settings set by `findTokens` which collect tokens so they
can be reported as references by `StackElement.prepare`.

To solve this, findTokens now accepts a function instead of an object
and basically just wraps it's invocation with settings that will collect
all tokens resolved within that scope. Not an ideal solution but
simple enough.

This was not discovered because we didn't have any tests that validated
the behavior of the generated CFN resources (they are not accessible
from the @aws-cdk/cdk library). We add a unit test in the IAM library to cover this use case.

Two improvements to stack dependency handling in the toolkit:

* Destroy stacks in reverse order. This is necessary to properly dispose of stacks that
  use exports and `Fn::ImportValue`. Fixes #1508.
* Automatically include stacks that have a dependency relationship
  with the requested stacks, unless `--exclusively` (`-e`) is
  passed on the command line. Fixes #1505.

* v0.22.0

See CHANGELOG

add packages.txt to github pages

Jetbrains IDEs will index the files in a project to provide intelligent suggestions to users. However, this project uses lerna and has cyclical symlinks inside node_module directories. The IDE will index until it runs out of memory following these links. Given the size of this project, doing the exclusions by hand is infeasable every time one does a git clean. This script will modify the .iml file directly to add the appropriate exclusions.

* The JetBrains IDEs exclusion list assumed your iml file was named aws-cdk. It now uses the root directory's name instead.
* IntelliJ stores the iml file at root. The script now accounts for that difference.
* For uncommon setups, the script can take an optional path arguement to opt-out of conventional file discovery, and be explicit.