Yahor Yuzefovich
Yahor Yuzefovich
@ajwerner do you have an idea for why this diff ``` diff --git a/pkg/sql/catalog/typedesc/type_desc.go b/pkg/sql/catalog/typedesc/type_desc.go index 02e77e5a40..08820572d9 100644 --- a/pkg/sql/catalog/typedesc/type_desc.go +++ b/pkg/sql/catalog/typedesc/type_desc.go @@ -796,6 +796,9 @@ func EnsureTypeIsHydrated( return err...
Why do you say so? The patch seems like a noop to me.
I ran [microbenchmarks](https://gist.github.com/yuzefovich/b5a6e68c45e42a79345ed5962830676b) of `colexec` package, and they mostly show minor speedup. The only concerningly looking result is for `OrderedSynchronizer`, but when running it in isolation, I see no difference,...
@michae2 will you have time to take a look at this this week - that's ok if not? I'd like to merge it before stability, and I'm off on Th-F.
> I feel like the 'right' way to solve this in the long term might be to make the various Set, Copy and Append methods deterministic in terms of their...
> I think only bytes columns would require any changes in Set, right? Value types like int would have no change, while pointer types like decimal or datum would just...
bors r- I think this needs a rebase.
Curious what people think about such a linter in general @cockroachdb/sql-queries.
Add an opt-out to disable the linter. TFTRs! bors r+