Paul Backus

@ljmf00 The existing overloads of `formattedRead` already accept a tuple as an argument; see the example here: https://github.com/dlang/phobos/blob/v2.101.1/std/format/read.d#L355-L361

What's the use-case for `No.exhaustive`? If the user wants to handle partial reads, it seems to me that they should use the overload that actually reports how much was read.

> The main usage is to comply with the current `slurp` implementation. IMO the way `slurp` handles partial reads is a mistake (and arguably a bug), not something we should...

> Well, but wouldn't that make it harder to deprecate `slurp` in favor of this by Phobos' standards? As far as I know there is no plan to deprecate or...

I think the real issue here is the name `Nullable`. If it were called `Optional` or `Maybe`, nobody would expect assigning `null` to work like this in the first place,...

I would much, much rather improve `match` itself than avoid using it. Duplication like this increases maintenance burden and risks introducing bugs (in fact, I think it has introduced one...

Do we also want to deprecate binding `var.member` to a template `alias` parameter, since it has the same behavior?

Spec PR was merged. Time to revisit this?

I believe @12345swordy is referring to this issue: https://issues.dlang.org/show_bug.cgi?id=21690

IMO it would be much better to either move the proverbial "needle" back to 0% and keep it there, or commit to having *all* default arguments evaluated at the call...