Michael J. Sullivan

Inlining the definition of `new_price` into the `set` seems to fix it here, as a workaround. Definitely a bug here. I think it has *something* to do with the tuple...

Another workaround is to put something in the schema (a property, a function arg/return, whatever) of type `tuple`, like ``` function _dummy_hack_bug_8917( x1: tuple, ) -> int64 using (0); ```...

The issue is that we are inferring the return type of the `??` as having type `tuple`, which we then need to cast from. 1. That is probably wrong on...

A full repro: ``` _localdev:main> create global foo := (select 'loooooooooooooooool' filter {true, false}); OK: CREATE GLOBAL _localdev:main> select global foo; warning: QueryError ┌─ :1:19 │ 1 │ select global...

Though, two caveats I need to think about: * Sometimes the source isn't getting sent right anyway. I've seen cases where the source was listing as `` until I restarted...

The bug that causes this is fixed, but unfortunately that means that if you have 7.10.0 installed, you'll need to reinstall the CLI using the install script or another mechanism.

> I was able to reinstall the CLI and use gel project init. Thank you!! I'm not sure what was causing the "Invalid access to memory location" error. It seems...

I think that probably the right way to do this is to have edgedb-rust not actually decode the WKB, but provide it to clients as bytes. This is so that...

It should work if you don't put quotes around the uuids. I think we need to document this stuff better, and probably also always support a string cast style

See also - [ ] https://github.com/edgedb/edgedb/issues/5183 - [ ] https://github.com/edgedb/edgedb/issues/7699