Move index creation from baseline script to a 5.1.1 migration, so it …

…will be applied in order (cannot create an older, out-of-order version as this would break migrations for everyone). #7256

src/main/resources/db/migration/V1__flyway_schema_baseline.sql

@@ -1,11 +1,2 @@
 -- TODO: we still should add the real base line here, too. That would avoid conflicts between EclipseLink
 --       trying to create new tables on existing databases. See https://github.com/IQSS/dataverse/issues/5871
-
--- This is unsupported by JPA, as it is PostgreSQL specific. Has to be done here, cannot be done in code.
--- (Only other option would be a lowercase copy of the data as a separate column, automatically filled py JPA)
-CREATE UNIQUE INDEX IF NOT EXISTS dataverse_alias_unique_idx on dataverse (LOWER(alias));
-CREATE UNIQUE INDEX IF NOT EXISTS index_authenticateduser_lower_email ON authenticateduser (lower(email));
-
--- Edit Dataset: Investigate and correct multiple draft issue: https://github.com/IQSS/dataverse/issues/2132
--- This unique index will prevent the multiple draft issue, yet it cannot be done in JPA code.
-CREATE UNIQUE INDEX IF NOT EXISTS one_draft_version_per_dataset ON datasetversion (dataset_id) WHERE versionstate='DRAFT';

src/main/resources/db/migration/V5.1.1.3__7256-purge-referencedata.sql

@@ -0,0 +1,17 @@
+-- #5361 and #7256 is about faster deployments, especially during development, sitting on an empty database.
+--
+-- This script has been part of  scripts/database/reference_data.sql that had to be executed manually on every new
+-- deployment (manually in the sense of Flyway didn't, the outside installer or an admin took care of it).
+--
+-- While this is pretty old stuff and should have been done earlier (baseline...), it will be a nice migration
+-- and behave like nothing happened if this is an existing installation. All new installation have an empty database
+-- on first app boot and benefit from this Flyway-based management.
+
+-- This is unsupported by JPA, as it is PostgreSQL specific. Has to be done here, cannot be done in code.
+-- (Only other option would be a lowercase copy of the data as a separate column, automatically filled py JPA)
+CREATE UNIQUE INDEX IF NOT EXISTS dataverse_alias_unique_idx on dataverse (LOWER(alias));
+CREATE UNIQUE INDEX IF NOT EXISTS index_authenticateduser_lower_email ON authenticateduser (lower(email));
+
+-- Edit Dataset: Investigate and correct multiple draft issue: https://github.com/IQSS/dataverse/issues/2132
+-- This unique index will prevent the multiple draft issue, yet it cannot be done in JPA code.
+CREATE UNIQUE INDEX IF NOT EXISTS one_draft_version_per_dataset ON datasetversion (dataset_id) WHERE versionstate='DRAFT';