Fix RestESDocumentsTest.iCanPerformESQLPageProviderOnRepository systematic failure

Context

With MongoDB / ES / Kafka

java.lang.AssertionError: expected:<20> but was:<19>
	at org.junit.Assert.fail(Assert.java:89)
	at org.junit.Assert.failNotEquals(Assert.java:835)
	at org.junit.Assert.assertEquals(Assert.java:647)
	at org.junit.Assert.assertEquals(Assert.java:633)
	at org.nuxeo.elasticsearch.test.rest.RestESDocumentsTest.iCanPerformESQLPageProviderOnRepository(RestESDocumentsTest.java:136)

JSON response debug logw attached when OK/KO.
We can reproduce locally against Mongi/ES/Kafka but only when running the whole nuxeo-elasticsearch-core module tests.

Problem

The issue was a remaining document from TestAutomaticIndexing#shouldIndexAfterPublishThenRestore which changed the PP result.
After investigation, we've noticed that it was impossible to delete the working copy with MongoDB in the scenario described below:

• publish a document under version 0.1
• make a change to this document and version it to 0.2
• restore version 0.1 to the document
• delete the working the copy

This scenario ended with this exception:

org.nuxeo.ecm.core.api.DocumentExistsException: Cannot remove WORKING_COPY_ID, subdocument WORKING_COPY_ID is the target of proxy 0.1_PROXY_ID

The issue was in the DBSSession#restoreVersion code which was restoring ecm:proxyIds whereas it shouldn't.