VishalPrasanna

Observed a bug in the cache invalidation callback `repset_relcache_invalidate_callback()` setting `row_filter` to `NULL` during the initial `COPY` command execution at Publisher node. As a result, all data from the publisher...

I'm not sure a deterministic test case can be created for this, since relcache invalidation can happen at any point during execution, it difficult to reproduce cache invalidation with the...

Currently, `debug_discard_caches=0` is used in regression tests. Even with `debug_discard_caches=1`, it is not possible to reproduce the exact invalidation scenario observed in [#496](https://github.com/2ndQuadrant/pglogical/issues/496) during regression test. #### Suggesting Test Case:...

> If you add those queries to the test suite in sql/, do they show wrong output without the rest of the change and correct output with the rest of...

Great, please proceed with your suggested changes and do the merge. Thanks