Panic in Row::try_get

[conradludgate]

I don't have a good explanation for what happened, but I see a panic in our production environment from the prepared statement code.

index out of bounds: the len is 1 but the index is 1

This panic comes from

rust-postgres/tokio-postgres/src/row.rs

Line 187 in c5ff8cf

let range = self.ranges[idx].to_owned()?;

which in turn came from

rust-postgres/tokio-postgres/src/prepare.rs

Line 149 in c5ff8cf

let type_: i8 = row.try_get(1)?;

For some reason, the typeinfo query returned a row data entry with only a single item in it. I haven't yet searched any further to understand how this could possibly happen, but it did so I am reporting it. I plan to fix this in our fork by replacing ranges[idx] with ranges.get(idx)? but maybe there's a better solution (erroring on Row::new)?

[sfackler]

Very weird! Yeah, I think I'd prefer checking up front in Row::new.