I'm a D noob. ".map!(a => a.length)" seems like the lambda is passed into the  template. ".map!split" just confuses me. What is split? I thought only types can be after "!". I would guess split is a standard function but then shouldn't it be map!(split)?

    const wordCount = file.byLine()                  // Read lines
                          .map!split                 // Split into words
                          .map!(a => a.length)       // Count words per line
                          .reduce!((a, b) => a + b); // Total word count

On Friday, 8 November 2013 at 05:46:29 UTC, ProgrammingGhost wrote:
> I'm a D noob. ".map!(a => a.length)" seems like the lambda is passed into the  template. ".map!split" just confuses me. What is split? I thought only types can be after "!". I would guess split is a standard function but then shouldn't it be map!(split)?
>
>     const wordCount = file.byLine()                  // Read lines
>                           .map!split                 // Split into words
>                           .map!(a => a.length)       // Count words per line
>                           .reduce!((a, b) => a + b); // Total word count

Do you know C++ templates? C++ func<thing> == D func!(thing).

You can pass anything into a template, not just types. So you are right, "map!split" gives the "split" function into the "map" template and "map!(split)" is the canonical form. D allows you to remove the parens for simple cases, hence "map!split".

When a template argument is only one token long (ie: one number, one type, one string, one name), the parenthesis are optional and can be omitted.

On Friday, 8 November 2013 at 05:46:29 UTC, ProgrammingGhost wrote:
> I'm a D noob. ".map!(a => a.length)" seems like the lambda is passed into the  template. ".map!split" just confuses me. What is split? I thought only types can be after "!". I would guess split is a standard function but then shouldn't it be map!(split)?
>
>     const wordCount = file.byLine()                  // Read lines
>                           .map!split                 // Split into words
>                           .map!(a => a.length)       // Count words per line
>                           .reduce!((a, b) => a + b); // Total word count

For a friendly introduction to D template system please take a look at this: http://nomad.so/2013/07/templates-in-d-explained/

On Friday, 8 November 2013 at 08:59:17 UTC, Gary Willoughby wrote:
> For a friendly introduction to D template system please take a look at this: http://nomad.so/2013/07/templates-in-d-explained/

Then to understand why parens are optional take a look at this: http://nomad.so/2013/08/alternative-function-syntax-in-d/

On Friday, 8 November 2013 at 06:25:15 UTC, qznc wrote:
> On Friday, 8 November 2013 at 05:46:29 UTC, ProgrammingGhost wrote:
>> I'm a D noob. ".map!(a => a.length)" seems like the lambda is passed into the  template. ".map!split" just confuses me. What is split? I thought only types can be after "!". I would guess split is a standard function but then shouldn't it be map!(split)?
>>
>>    const wordCount = file.byLine()                  // Read lines
>>                          .map!split                 // Split into words
>>                          .map!(a => a.length)       // Count words per line
>>                          .reduce!((a, b) => a + b); // Total word count
>
> Do you know C++ templates? C++ func<thing> == D func!(thing).
>
> You can pass anything into a template, not just types. So you are right, "map!split" gives the "split" function into the "map" template and "map!(split)" is the canonical form. D allows you to remove the parens for simple cases, hence "map!split".

Oh I see. Yes I understand C++ templates which is how I guessed that. This FEELS UNUSUAL. Because it seems like it is .map(!split.map(!(...))).reduce...

As if split.map was the template parameter. How does it know if split isn't a class (or if d has them, namespace) and map is a static function? Thats why it confused me.

ProgrammingGhost:

> As if split.map was the template parameter. How does it know if split isn't a class (or if d has them, namespace) and map is a static function? Thats why it confused me.

D doesn't have namespaces, it has modules and packages. map is a higher order function that returns a lazy templated range, a templated struct.

My suggestion is to start from simpler things (simpler syntax) and go up on from there. If you start from the end, you will be confused :-)

Bye,
bearophile

On Friday, 8 November 2013 at 20:15:14 UTC, ProgrammingGhost
wrote:
> Oh I see. Yes I understand C++ templates which is how I guessed that. This FEELS UNUSUAL. Because it seems like it is .map(!split.map(!(...))).reduce...

Simply put, anything more complex than a single name requires !()

     to!char[]("fish");

"test.d(5): Error: cannot resolve type for to!char"