Gerd Zellweger
Gerd Zellweger
I don't think these transitions should all be logged on info! especially if they work fine maybe reduce to debug! ``` 2024-09-12 22:46:14 INFO [manager] Transition: running -> failed (desired:...
https://github.com/feldera/feldera/blob/8b3a719e46348b430b9d3c0805a7e7a597a180aa/sql-to-dbsp-compiler/build.sh#L20
the binary that runs the program is removed from the path where it ran from when the pipeline is stopped... It looks like it gets copied there every time. Maybe...
run cargo test in adapters, hit ctrl+c 60min later in your shell randomly this pops up: ``` failures: integrated::delta_table::test::delta_table_file_output_proptest integrated::delta_table::test::delta_table_follow_file_test integrated::delta_table::test::delta_table_snapshot_and_follow_file_test test result: FAILED. 122 passed; 3 failed; 1 ignored;...
do you want to file an issue to refactor that trait? _Originally posted by @mihaibudiu in https://github.com/feldera/feldera/pull/2381#discussion_r1739930668_
seems like we can't serialize TIME into arrow: ``` can serialize stuff: Error: serialize_u64 is not implemented for Time64Builder Backtrace: 0: ::serialize_u64 1: ::serialize_with_context 2: ::serialize_struct_field ```
on macbooks it currently installs/uses the AMD64 images but we should build ARM images for it. This is probably why compilation time on mac is much slower when using docker.
We probably want a specialized strategy for timestamps. It could allow fancy things like bounded lateness. _Originally posted by @ryzhyk in https://github.com/feldera/feldera/issues/2318#issuecomment-2314488714_
I will add a test _Originally posted by @gz in https://github.com/feldera/feldera/pull/2357#discussion_r1735026577_
This allows to share tokio runtimes across different sub-systems inside your application. ## PR Type Feature, Refactor https://github.com/actix/actix-net/issues/580 ## PR Checklist Check your PR fulfills the following: - [x] Tests...