Jeremy Taylor

For the curious: https://github.com/refset/crux/tree/upstream/jt/109-datafy-nav

Hey @zirota thanks for contributing the initial work here and for the updates :) I'm a bit out of context to understand all these changes currently, but it would be...

Just noting that the CI build test (for `xtdb.kafka-connect-test/test-sink-task`) failed: https://app.circleci.com/pipelines/github/xtdb/xtdb/2542/workflows/ee8fcc64-0eeb-4b91-aff1-74414cbbbe86/jobs/9864?invite=true#step-103-1169 - I'll convert this PR to 'draft' in the meantime. For anyone interested, there's some discussion also happening here:...

Noting the workarounds I suggested on the Slack: ```clojure ;; it should be acceptable in this specific instance to drop the [e :something something-val] clause and use (pull e [:something])...

Hi @brdloush thanks for the update on your findings. I can't remember if we discussed / tried it, but what happens if you change `(or ...` to `(or-join [e id]...

I'm not advocating to bring `xodarap` in as a dependency (and `core.async` along with it), but it's been good to me before and might provide some inspiration: https://github.com/divs1210/xodarap (see also...

This has now been reported by a different user running `21.06-1.17.1-beta` (which includes the fixes mentioned above), along with the other related WARN from `crux-s3`: > There are 34 occurrences...

Noting that in the event of a ~catastrophic document store failure that these merged Part One changes could be useful for some forensic investigation/recovery assuming valid index-stores/checkpoints exist.

Hey @blshao84 - thanks for the report! I'll dig in over the coming days and see if there's some easy fix for this.

I have no firm update to offer on this specifically yet, but I just wanted to briefly note that it's _possible_ that https://github.com/xtdb/xtdb/pull/1759 also resolved this...but it needs confirming either...