Yeah you could do `&(&1)` or so for an inline identity, though definitely
not as clear. :-)
Constant is not as succinct though: `&(fn->&1 end)`
And still not clear. ^.^
On Thursday, November 17, 2016 at 12:14:39 PM UTC-7, Drew Olson wrote:
>
> Ignore my comment about not being about to use t
While we're wishing for things, I've also wanted constant/1 a few times:
constant(3).()
# => 3
Especially for things like creation of Agents.
On Thu, Nov 17, 2016 at 1:08 PM, Drew Olson wrote:
> I've also often wished for an identity/1 in Kernel, especially because you
> cannot easily represen
I've also often wished for an identity/1 in Kernel, especially because you
cannot easily represent it with the short function syntax.
On Thu, Nov 17, 2016 at 12:33 PM, OvermindDL1 wrote:
> `identity : 'a -> 'a` is the most common function declaration like that in
> almost every functional langua
Ignore my comment about not being about to use the shorter function syntax.
I was confused about the precedence of &.
On Thu, Nov 17, 2016 at 1:08 PM, Drew Olson wrote:
> I've also often wished for an identity/1 in Kernel, especially because you
> cannot easily represent it with the short functi
`identity : 'a -> 'a` is the most common function declaration like that in
almost every functional language, would be nice to have in kernel.
On Thursday, November 17, 2016 at 10:48:20 AM UTC-7, Louis Pop wrote:
>
> Hiya
>
> I'd like an identity function in Kernel, that's for sure.
>
> I think a
Hiya
I'd like an identity function in Kernel, that's for sure.
I think an identity function being the default argument to filter would be
quite nice. I find reject a bit strange, I don't set why it exists when we
have filter.
Cheers,
Louis
On Thu, 17 Nov 2016, 17:40 Tallak Tveide, wrote:
> I
I assume you mean the default function should be
'''
fn x -> !x end
'''
To reject any falsey value. It does seem reasonable, but I dont think it is
nearly as readable as 'compact', so I vote nay to this one...
In fact, isn't this a better option? (Or equally good)
'''
def filter(enum, fun \\ f
Every couple of months, we get a request for Enum.compact. So far, the main
argument against it is that Enum.reject list, &is_nil/1 is sufficient.
When you compare the difference in a pipe, you can see a significant
difference in readability.
It really is a pervasive use case.
list
|> nil_or_res
