Austin Bonander
Austin Bonander
@joeydewaal do you have time to address the merge conflicts?
Another bikeshedding question: should we change the name of `QueryBuilder::push()`, since it allows adding unsanitized data to the query? It's already got a warning on it, but I'm wondering if...
Yeah, let's deprecate `push` in favor of `push_unsanitized` and see what feedback we get on it.
On second thought, I'd rather just merge this. I think `QueryBuilder` already draws enough attention to itself.
> I left out the `Decode` impl for `Rc` because of the `Send` trait bounds in `QueryAs` and `QueryScalar` (which makes it impossible to use `Rc`). @joeydewaal technically it's still...
CI should be fixed if you rebase.
The problem is that we treat all integer widths as compatible with each other which isn't _strictly_ correct. Sure, we can do truncation and sign extension but `TypeInfo::compatible()` should only...
If you're using `0.4.0-beta.1` you can override this behavior by doing ``select forum_posts as `forum_posts: u32` ...`` which will force the type to be `u32`.
Yeah, so what's happening is this: The query macros are made aware of supported types via `impl_type_checking!{}`, which for MySQL is done here: https://github.com/launchbadge/sqlx/blob/e10789d9d76cbff2755f977b7a126bd67a1ec5e5/sqlx-mysql/src/type_checking.rs In the expansion of `impl_type_checking!{}`, it...
I don't think this is a good idea, honestly. If you want to revert every migration that's changed, that's assuming that you're actually _using_ reversible migrations. This also implies that...