)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":7233,"name":"Matthew Oliver","email":"matt@oliver.net.au","username":"mattoliverau"},"change_message_id":"d9950ab59dd6f6ab5c9ec96b91b2f5aa4ba3ce3f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"a537dae4_fd8eb052","updated":"2023-03-22 04:38:25.000000000","message":"Maybe something like this Al? This is just one thought that isn\u0027t a special case InternalClient child class.","commit_id":"19a2b3c40e7bfcf401d4db698b6ff131878d7b84"},{"author":{"_account_id":1179,"name":"Clay Gerrard","email":"clay.gerrard@gmail.com","username":"clay-gerrard"},"change_message_id":"7c5b485591a86dfe3d369bb671629c314a27aa3c","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"e84c5db6_4e5ebf29","updated":"2023-03-22 14:15:27.000000000","message":"this seems like an un-nessecary new interface.  It seems *entirely* reasonable to me that we will not allow the gatekeeper in InternalClient pipelines (or at least not the ones we depend on).\n\nI think instead of trying to find a complicated way to \"make it work\" when someone MIS-configures their internal_client.conf to include the gatekeeper we should error LOUDLY.\n\nBesides, don\u0027t we have probetests and other consumers of InternalClient that already send x-backend or sysmeta headers directly?!","commit_id":"19a2b3c40e7bfcf401d4db698b6ff131878d7b84"},{"author":{"_account_id":7233,"name":"Matthew Oliver","email":"matt@oliver.net.au","username":"mattoliverau"},"change_message_id":"6c9545240e36c490c419f309ae9c30d36ceb4bc4","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"4dbd7279_7c614a7d","in_reply_to":"e84c5db6_4e5ebf29","updated":"2023-03-22 22:32:06.000000000","message":"yeah, the sharder relies on it too.. so there is probably already an edgecase where sharding could break today.\n\nWe default to not modify pipeline, so the gatekeeper by default isn\u0027t in the pipeline.","commit_id":"19a2b3c40e7bfcf401d4db698b6ff131878d7b84"}]}