Sean Smith

Another failure: https://github.com/GlareDB/glaredb/actions/runs/14434060685/job/40472353293?pr=3647

https://github.com/GlareDB/glaredb/blob/main/crates/glaredb_http/src/handle.rs#L102-L118

Not fixed ``` test [slt_http/multi] slt/http/read_csv_http.slt ... ok ---- PANIC ---- Info: panicked at /home/runner/work/glaredb/glaredb/crates/glaredb_core/src/execution/partition_pipeline.rs:122:14: poll_execute to be called on pipeline that hasn't completed Backtrace: 0: glaredb_slt::run::{{closure}} at /home/runner/work/glaredb/glaredb/crates/glaredb_slt/src/lib.rs:99:25 1:...

Would be other way around. Having dependencies would allow us to enforce not being able to drop connections that other tables depend on. Even then, I would say this is...

This could be an issue in dd in `fetch_parquet_metadata`: https://github.com/apache/arrow-datafusion/blob/58fc80eddaf93d4c6399e4ac4fc1234649b26e43/datafusion/core/src/datasource/file_format/parquet.rs#L408-L470 Or an issue with how we construct object meta for http sources: https://github.com/glaredb/glaredb/blob/9e0e4b3dd8b196d151c5515cb58fe91273fa4f91/crates/datasources/src/object_store/http.rs#L64-L84

Were you able to reproduce the panic? I tried on cloud, and only get OOM. Running locally gets me: ``` Error: External error: Parquet error: Parquet error: Could not parse...

Content length (through wget): 14779976446 Size on disk: 14779976446

Working on getting us some stuff to allow for adding context to errors: https://github.com/sqlparser-rs/sqlparser-rs/pull/1046.

Those functions should still be returning tables, TODO here: https://github.com/glaredb/glaredb/blob/0fab58b30045282051dabb6ff5d5f7cbc2b5cbbb/crates/glaredb_core/src/logical/resolver/resolved_function.rs#L3-L25

We did have this previously, but I toned it back to only running on Mondays instead of every day in https://github.com/GlareDB/glaredb/pull/1881. If we want to, we can bump that back...