Refine ValidKeysConstraintValidator's messages: use #3401883's new Mapping infrastructure

The updated test coverage should fail like this:

Testing /Users/wim.leers/core/core/tests/Drupal/KernelTests/Core/TypedData
.F..                                                                4 / 4 (100%)

Time: 00:04.529, Memory: 10.00 MB

There was 1 failure:

1) Drupal\KernelTests\Core\TypedData\ValidKeysConstraintValidatorTest::testUnknownKeys
Failed asserting that two arrays are identical.
--- Expected
+++ Actual
@@ @@
 Array &0 (
-    0 => ''use_site_logo' is an unknown key because plugin is system_powered_by_block (see config schema type block.settings.*).'
-    1 => ''use_site_name' is an unknown key because plugin is system_powered_by_block (see config schema type block.settings.*).'
-    2 => ''use_site_slogan' is an unknown key because plugin is system_powered_by_block (see config schema type block.settings.*).'
+    0 => ''use_site_logo' is not a supported key.'
+    1 => ''use_site_name' is not a supported key.'
+    2 => ''use_site_slogan' is not a supported key.'
 )

So:

Before

use_site_logo' is not a supported key.

After

use_site_slogan' is an unknown key because plugin is system_powered_by_block (see config schema type block.settings.*).

In other words: a much more detailed message.

ℹ️ This MR contains 311 insertions(+), 60 deletions(-) of the 1125 insertions(+), 151 deletions(-) that the MR for #3364108: Configuration schema & required keys contains.

IOW: this shrinks the other MR by ~30%. The other MR can be layered cleanly on top of this one; not a single line added there would need to change.

This looks good and makes sense. I have a fair number of small comments (as I usually do), in the hope of improving clarity, but overall I find this close to RTBC-able..

I would hope so, since you've already reviewed exactly this code multiple times in #3364108: Configuration schema & required keys 😄

On it :)

Some good questions about code clarity, I hope I made you proud, @phenaproxima 😄 (This ended up taking 1.5 hour 🤯 — naming & explaining can be time-consuming :D)

It looks like @phenaproxima and @Wim Leers have already taken the fun out of the review, since it is already improved over the orignal with much more/better comments and better naming.

🤣

You've both already heavily influenced this MR, by reviewing #3364108: Configuration schema & required keys as much as you have 😊 Thank you!

Applied all of @phenaproxima's remaining suggestions because #3406487: Consistently use "dynamic type name" and "expression" instead of "variable value" in TypedConfigManager's terminology landed 🥳

Restored ValidKeysConstraint::getAllowedKeys() method per @alexpott's review. Since this is just trivially moving logic from one spot to another, there's 0 functional changes and tests are still passing: back to RTBC.

Addressed all of @larowlan's feedback 😊

(And merged in 11.x, which allowed reverting 2 lines 👍)

I merged in upstream in #16. This caused a failure in \Drupal\Tests\block\Kernel\Migrate\d6\MigrateBlockTest due to #3371219: Remove forum from block migration tests (landed 2 days ago), whose expected messages become more precise thanks to this issue. 👍 Trivial fix!

Based on @larowlan's feedback, I reverted a few lines from the test coverage additions (because they belong in #3364108: Configuration schema & required keys, which is blocked by this issue, plus Lee marked the MR thread as resolved himself), added a typehint and tweaked an assert().

The actual changes are so minor that I feel a re-self-RTBC is reasonable.

Yay! Next up: #3364108-90: Configuration schema & required keys 😊