jeff-davis

Don't worry about the license, I just thought I would mention it but it is not a real problem. For postgres, the error levels basically mean: * ``ERROR`` - something...

> ...but there's an issue when using libraries that include unsafe Rust code or call other C libraries... If the extension calls back into PostgreSQL code that touches shared memory,...

This is caused by using a `regclass` in `columnar.options` instead of the storage ID.

I still don't see why `pg_try` should be a no-op in plrust. It certainly surprises me as a user. It can't be made safe in trusted plrust, so I would...

> I don't really intend to focus on addressing higher-level niceties at this time. By that you mean that `pg_try` will remain a no-op in plrust? I can't say that...

Likely, but I wasn't able to confirm so I filed a separate issue.

I didn't repro the failure, but I added a couple lines to the SQL file and ran it locally, and I see: ``` SELECT 'col_table_1'::regclass::oid AS col_table_1; col_table_1 --------------------------------------------------------------------- 17180...

@m-ou-se Thank you for working on this. What you describe satisfies my use case almost perfectly. One other thing to consider is what happens with ``cargo test``. A plugin-style shared...

In Cargo.toml, I specified "PostgreSQL OR MIT" as the license. Is it expected that rust crates also have a LICENSE file?

Ultimately, what I am hoping to achieve is something like: ``longjmp_panic!(my_ffi_function(x,y,z))`` Which would call ``my_ffi_function()`` and return normally, or panic if ``my_ffi_function()`` longjmps. It would also be nice to have:...