Currently, has some content has corrupted data (for e.g. RichText field with incorrect XML), the search indexer is not able to finish index creation/re-indexing process. It stops on error.
The idea is to add --continue-on-errors option to the command, which allows indexer to skip potentially corrupted data and finish index at least once. The second important improvement is to make possible to display corrupted content IDs during executing create index/reindex command.