Re: confused about string and lambda args
On Friday, 17 January 2020 at 07:57:16 UTC, mark wrote: On Thursday, 16 January 2020 at 17:11:11 UTC, Adam D. Ruppe wrote: [...] The string thing probably shouldn't be used anymore. I suggest you always use the => form instead. The string thing is a legacy version that was before the language had =>. [...] Thanks for that very clear explanation. I have attempted to fix that part of the tour: https://github.com/dlang-tour/english/pull/304 I (hopefull) deleted the above and replaced it with: https://github.com/dlang-tour/english/pull/305
Re: confused about string and lambda args
On Thursday, 16 January 2020 at 17:11:11 UTC, Adam D. Ruppe wrote: [...] The string thing probably shouldn't be used anymore. I suggest you always use the => form instead. The string thing is a legacy version that was before the language had =>. [...] Thanks for that very clear explanation. I have attempted to fix that part of the tour: https://github.com/dlang-tour/english/pull/304
Re: confused about string and lambda args
On Thu, Jan 16, 2020 at 05:03:33PM +, mark via Digitalmars-d-learn wrote: [...] > auto wordCharCounts4 = words // I added this and it won't compile > .map!(a => a.count); // Error: no property count for type string > writeln(wordCharCounts4); You need to import std.algorithm to get `count`. > I don't understand why both syntaxes work for .length but only the > string form for .count? Because .length is a property of strings, whereas .count is actually not a string property, but a function that's being called via UFCS: Unified Function Call Syntax, in which when the compiler sees something like: obj.func(x, y, z); but `obj` doesn't have a member named `func`, then it will try to rewrite it into: func(obj, x, y, z); instead. T -- Knowledge is that area of ignorance that we arrange and classify. -- Ambrose Bierce
Re: confused about string and lambda args
On Thursday, 16 January 2020 at 17:03:33 UTC, mark wrote:
auto wordCharCounts = words // I added this and it works
fine
.map!"a.length";
writeln(wordCharCounts);
The string thing probably shouldn't be used anymore. I suggest
you always use the => form instead.
The string thing is a legacy version that was before the language
had =>.
I don't understand why both syntaxes work for .length but only
the string form for .count?
It is because of imports.
So the string version passes the string to the library, which
pastes it into some skeleton code and makes a function out of it.
It basically does:
string code = "import some_stuff; (a) { return " ~
your_string ~ "; }";
mixin(code);
Note it does this INSIDE the library.
It is that `import some_stuff;` that accounts for this
difference. The string one pastes in some library imports so some
functions are available. The => form does not.
Since the string one is inside the lib, it can NOT see your own
functions from your module! But since the lib imports a few other
library modules, it may be able to see things your module didn't
import.
The better way to do it is to use your => format, but go ahead
and import the necessary module.
I believe `count` is located in `import std.algorithm;`. So add
that to your module and it should work now.
confused about string and lambda args
I'm looking at https://tour.dlang.org/tour/en/gems/range-algorithms (IMO the example code is far too long and complicated.) But here's the thing: auto wordCharCounts = words // I added this and it works fine .map!"a.length"; writeln(wordCharCounts); auto wordCharCounts2 = words // I added this and it works fine .map!(a => a.length); writeln(wordCharCounts2); auto wordCharCounts3 = words // this is in the tutorial .map!"a.count"; writeln(wordCharCounts3); auto wordCharCounts4 = words // I added this and it won't compile .map!(a => a.count); // Error: no property count for type string writeln(wordCharCounts4); I don't understand why both syntaxes work for .length but only the string form for .count?
