[Pre-pitch] Conditional default arguments

From: Matthew Johnson <matthew@anandabits.com>
Date: November 24, 2017 at 3:55:34 PM PST
To: Dave Abrahams <dabrahams@apple.com>
Subject: Re: [swift-evolution] [Pre-pitch] Conditional default arguments

You posted off-list. If that wasn’t intentional feel free to forward my reply to the list.

Sent from my iPad

Adding language support for defining these more directly would eliminate a lot of boilerplate

I think I understand what you’re trying to do, but it’s not obvious to me that this pattern is common enough to warrant investing language design resources or the increased language complexity of a dedicated feature.

It’s not obvious to me either. It’s not something I come across frequently, but the overloads get painful pretty quickly in the cases where I do.

When I initially added it to the manifesto I put it in the “maybe”section and called it a “pre-pitch” for this reason. I posted on list because Doug had a positive response to the idea. It may well be a case where solving it directly at the language level isn’t appropriate.

I don’t plan to push hard for it if the response is lukewarm. I’m primarily interested in finding out what the response is at this point. :)

When you say “a lot of boilerplate,” how much do you mean?

It’s a combinatoric explosion of overloads so it depends on the number of conditionally defaulted arguments.

The example I gave above is a simplification based on some code I have written. This code already requires an overload set for other reasons. Each of the base overloads requires a set providing emulated default arguments. The total size of the set is uncomfortable.

Is this a pattern you’ve observed outside your own code?

No, but I haven’t been actively looking for it. Part of the reason for posting to the list is to seek feedback and see if others have had similar cases arise.


Sent from my iPhone
Begin forwarded message:

On Nov 24, 2017, at 5:32 PM, Dave Abrahams <dabrahams@apple.com> wrote:

On Nov 24, 2017, at 3:11 PM, Matthew Johnson via swift-evolution <swift-evolution@swift.org> wrote: