Austin Drenski
Austin Drenski
> I think this issue is more a matter of detecting URIs and then internally constructing instances of `NpgsqlConnectionStringBuilder` from them, it shouldn't be too hard... Do we need to...
Would be great to see this finally added. Several options, including the recent PR #1472 from @lucaskatayama, and an older PR #1380 from @LynnL4.
@roji I think hardcoding discreteness makes sense for the built in range types. But to clarify, I wasn't suggesting a database lookup for the canonical functions. Initially, I was playing...
@roji wrote: > Considering that a concrete range implementation would typically be a struct, having mutators on the interface would make the struct mutable, which is a pretty bad idea...
I've opened #1974 to update `NpgsqlRange` to use `ReadOnlySpan` per some pre-`4.0.0` [discussion](https://github.com/npgsql/npgsql/pull/1939#pullrequestreview-121308396). @Brar I've included your unit test as `Roundtrip_canonicalization_causes_issues`. The condition is flipped such that passing means the...
@Brar Can you post a test?
@Brar I'll keep on it. This month's a bit stretched for me, but I should be able to make some headway by the end of July.
> One thing (apart from general too-much-to-do-ness!) that seemed potentially difficult in carrying on with this is that it looks as if it might be quite a chore to set...
> would I code it as a pull request against the `dev` branch? Yep
@roji I'll grab this one too. Do you want to milestone this for `2.1.1` or hold it for `2.2.0`?