Add deprecation layers

[alcaeus]

Q	A
Type	improvement
BC Break	no

Summary

Note: PR is best reviewed commit-by-commit.

This PR brings in a first round of deprecations:

• [2.0] Remove YAML mapping support #1734: Deprecates the YAML metadata drivers
• [2.0] Separate ID mapping from fields in XML driver #1807: Add the id element to the mapping schema and deprecate passing id="true" to a field element
• [2.0] Drop namespace property from ClassMetadata #1812: Deprecate the namespace property and its accessor and trigger a deprecation warning when passing non-FQCN
• [2.0] Merge ClassMetadataInfo into ClassMetadata #1733: Move logic from ClassMetadataInfo to ClassMetadata, deprecate ClassMetadataInfo and provide an alias
• [2.0] Validate mapping files against schema #1814: Validate XML mapping files against a schema and trigger a deprecation warning when passing an invalid mapping file
• [2.0] Consistently use kebab-case in XML mappings #1834: Deprecate writeConcern and fieldName attributes in schema in favour of write-concern and field-name

I might have to revisit the BC layer for ClassMetadata once again. @greg0ire has tried the same for ORM in doctrine/orm#6886. What I'd like to do: trigger a deprecation notice when someone uses ClassMetadataInfo but allow using ClassMetadata in place of it. Using class_alias in ClassMetadata.php means that we can't trigger a deprecation notice. We might have to split this out of this PR.

[greg0ire]

See also: https://dev.to/greg0ire/how-to-deprecate-a-type-in-php-48cf

Using class_alias in ClassMetadata.php means that we can't trigger a deprecation notice.

You actually can, but conditionally.

[malarzm]

I'll take your word for ClassMetadata stuff, rest looks ok to me except for invalid schema that fails some tests :)

[alcaeus]

@greg0ire updated code with your suggestions and changed ClassMetadataInfo deprecation to follow the advice in your blog post. Could you take another look and see if you spot anything that seems wrong?

[alcaeus]

@greg0ire updated the comment to add both reasons.