FYI - Have you seen Tail.Fody? about language-ext HOT 8 CLOSED

Any thoughts before I close this?

from language-ext.

Sorry, haven't really had time to follow this up.

Have you seen Tail.Fody

Not that specific plugin no, but I have seen Fody. I think Tail.Fody whilst useful for the end-user, isn't really that useful for language-ext. Also I believe that RyuJIT is supposed to do a better job of dealing with tail-calls, so it's probably moot. It's nearly never an optimisation in C# because loops will always be faster than recursion. I think for the users of Language-Ext they would prefer the speed benefits of a pure functional interface hiding imperative behind-the-scenes (as long as the behaviour is referentially transparent - as with the fold implementations).

There's another for adding null checks to all methods, except for where you specify, eg. OptionUnsafe.
https://github.com/Fody/NullGuard

That's definitely more interesting, but might be a painful process of retrofitting [AllowNull] where it is allowed. I'd be quite cautious about this, because of possible breaking changes. I dunno, I'm torn, because it's obviously a very good idea, and I can't stand the fact that the if(arg == null) checks breaks expression based code.

from language-ext.

If you're interested, I could put NullGuard just into OptionT and UnsafeOptionT, and see what you think from there?

from language-ext.

My understanding is that you can't opt-in, you can only opt-out by using
[AllowNull], so that would be a project-wide change. That is my concern.
There's a ton of code that would need to be considered. Unless you know
otherwise?

It seems to defeat the point having to opt-in.

On Mon, 9 Nov 2015 at 08:20 Will Mooar [email protected] wrote:

If you're interested, I could put NullGuard just into OptionT and
UnsafeOptionT, and see what you think from there?

—
Reply to this email directly or view it on GitHub
#62 (comment).

from language-ext.

The opt-in approach would let us get used to it before we risk turning it on across the lot.
Or we rely on the unit tests and hit the big red button...

I'll check and get back to you re project-wide opt-out.

from language-ext.

It seems to me to be the kind of thing that would only be valuable project-wide. Otherwise when you look at any cs file you won't know whether it's got the protection it needs [without looking at some project settings]. If you're prepared to take it on project-wide that'd be great, but I wouldn't underestimate the amount of work and the cautious approach you'd need to take.

from language-ext.

Yes, I hear you there. Just read their outstanding issues and am concerned that maybe it isn't ready for Roslyn yet, which we depend on. Probably shelve it for now methinks, sorry to waste your time there!

(I'll keep an eye on it, it wouldn't surprise me if a patch comes sometime in the coming weeks)

from language-ext.

Thanks Will, I don't think it's time wasted. It definitely has potential.
Let's keep an eye on it.

On Mon, 9 Nov 2015 at 08:55 Will Mooar [email protected] wrote:

Closed #62 #62.

—
Reply to this email directly or view it on GitHub
#62 (comment).

from language-ext.