Dear hackers, I did a self-reviewing and updated a patch. PSA new version. What's new:
1. Fixed a bug which existing name can be specified by ALTER PUBLICATION RENAME. A validation is added in RenamePublication(). Just in case, we want to discuss the case that the renaming publication has thousands of tables. Before the patch, The ALTER PUBLICATION RENAME command just invalidates a syscache, and registered callback would invalidate all of them. With the patch, however, all relsync entries are invalidated within the command so that the execution time can be longer. I feel this can be acceptable because some codes have such a part; E.g., ATTACH/DETACH partition command invalidates relaches of the children when the specified relation also has children - even if it has thousands of leaves. (See d6f1e16). Also note that all other processes like walsender can be faster by avoiding discarding entire caches. how do you feel? Best regards, Hayato Kuroda FUJITSU LIMITED
v8-0001-Introduce-a-new-invalidation-message-to-invalidat.patch
Description: v8-0001-Introduce-a-new-invalidation-message-to-invalidat.patch
v8-0002-Invalidate-Relcaches-while-ALTER-PUBLICATION-RENA.patch
Description: v8-0002-Invalidate-Relcaches-while-ALTER-PUBLICATION-RENA.patch
