this commit propagates the usage of `ValidatedConfig` throughout the
`output-targets/` directory. This is being done following making `sys` a
required field on `ValidatedConfig` in 991843a
(#3491), where driving down the number of `sys`-related violations of the compile
time check is the primary driver.  however, this commit does also lay groundwork
for future fields being made required on the validated config, and require less
proliferation of the type in the future.

all output target types were updated in this commit. that includes the legacy angular
output target, and the custom elements bundle output target. both of the aforementioned
targets are intended to be deprecated/removed in stencil v3. however, the level of effort
required to proliferate the change provides us with a more accurate count of
`strictNullChecks` violations, and does not require waiting for a major version bump of
stencil to drive the violation count further

…3497)

this commit propagates the usage of `ValidatedConfig` throughout the
portion of the code that validates an output target defintion from a
configuration. This is being done following making sys a
required field on ValidatedConfig in 991843a
(#3491), where driving down the number of sys-related violations of
the compile time check is the primary driver. however, this commit
does also lay groundwork for future fields being made required on the
validated config, and require less proliferation of the type in the
future.

all output target validations were updated in this commit. the
validations for the legacy angular output target, and the custom
elements bundle output target. both of the aforementioned targets are
intended to be deprecated/removed in stencil v3. however, the level of
effort required to proliferate the change provides us with a more
accurate count of strictNullChecks violations, and does not require
waiting for a major version bump of stencil to drive the violation
count further

this commit propagates the usage of `ValidatedConfig` throughout
various parts of the code. the changes are distributed and small in size,
which made it feel unnecessary to split into multiple pull requests, given
the nature of the changes being proposed.

this is being done following making sys a required field on ValidatedConfig
in 991843a (#3491), where driving down the number of sys-related violations of
the compile time check is the primary driver.

this commit updates `ValidatedConfig` in order to make the `testing`
field required. this is being done in pursuit of dropping the number of
`strictNullChecks` violations in the code, and was a side task of
stencil-501 (the author of this commit was actively in the testing code,
making it a good opportunity to knock this out).

this commit makes `testing` default to an empty object literal, rather
than `null` or pre-populating existing fields. choosing the former would
require additional optional checks throughout the code (which assumes
`testing` is a defined field, and choosing the latter would expand the
scope of this commit vastly. an empty object literal should incur
minimal runtime overhead, while keeping the size/scope of this commit to
a minimal

this commit updates the generation of a `ValidatedConfig` entity to take
a `sys` property on a user provided `Config` in the context of
`runTask`.

when generating the validated configuration, the `sys` entity takes
precedence over all other configurations, with the belief that in the
event argument was provided, it is the intent of the user to use it. in
the event `sys` is not expressly provided, use the one on the provided
configuration entity. this is useful in the case that a user used
stencil's external facing apis in a flow that:
1. loads a configuration
2. validates that configuration
3. uses that configuration when calling `runTask` directly

this commit fixes a bug introduced making `sys` a required field on
`ValidatedConfig`, where `config.sys` was not properly accounted for.
the bug was initially introduced in
a6a9171 (#3491)

This fixes an issue (#3461) where CSS is being inappropriately mangled
when using the hot-reload dev server. The issue has details about
exactly what the problem is, but basically in short if you are using the
`dist-hydrate-script` output target and running the dev server then any
change to a CSS file will cause all styling to be wiped from the related
component in the browser, making for a pretty inadequate developer
experience.

This commit fixes the issue by changing the conditions under which
original CSS selectors are commented out

this commit removes a handful of strict null check violations that can
be found in stencil's jest preprocessor. most changes are concerned with
making types agree - e.g. if a variable is declared to be of type `T`,
but is assigned by a function that returns `T | U`, the type of the
variable is expanded for agreement.

the two changes that are worth calling out in particular is introducing
a `throw` statement in the event that the `transformOptions` parameter
is falsy both when we process a file and attempt to retrieve a cache
key. from jest 23 thru 27, the function signatures of these two
functions changed a few times (documented in the code today). however,
a typing issue that was introduced in
e0640de (#3171) was `transformOptions`
could be considered `undefined` by typescript. we know that for jest
[23,26], this value will never be falsy (as it is a part of the call
signature). for jest 27, we re-assign a value to this variable for
backwards compatibility reasons.

there is a high level of confidence that this `throw` is safe to
introduce for the following reasons:
1. if `transformOptions` is undefined/null, stencil would already be
   erroring out when running the test task
2. the function signatures changes of these two functions have been
   handled in e0640de
3. in jest [23,26], `transformOptions` _always_ has a `rootDir` field of
   type `string` (which is the field we're accessing here causing
   errors)

this commit adds additional types for the function that finds a user's
configuration file on disk in the hope of adding clarity and removing a
type assertion in the function

this commit updates the "Git status check" step
of the browserstack workflow to print the names
of the files that are uncommitted