www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - type conversions

reply "WhatMeWorry" <kc_heaser yahoo.com> writes:
I'm trying to get my head around D's type conversion. What is the
best way to convert a string to a char array? Or I should say is
this the best way?

string s = "Hello There";
char[] c;

c = string.dup;


Also, what is the best way to explicitly convert a string to an
int?  I've been looking at Library Reference (Phobos) but I'm
stuck.
Apr 29 2012
next sibling parent "Era Scarecrow" <rtcvb32 yahoo.com> writes:
On Sunday, 29 April 2012 at 23:42:39 UTC, WhatMeWorry wrote:

 I'm trying to get my head around D's type conversion. What is 
 the
 best way to convert a string to a char array? Or I should say is
 this the best way?

 string s = "Hello There";
 char[] c;

 c = string.dup;


 Also, what is the best way to explicitly convert a string to an
 int?  I've been looking at Library Reference (Phobos) but I'm
 stuck.


I think this is right...

import std.stdio;
import std.conv;

int main(string args[]){
	string str = "Hello world!";
	char[] c = to!(char[])(str);
	
	string num = "1234";
	int i = to!int(num);
	
	writefln("%s\t- %s", typeid(str), str);
	writefln("%s\t- %s", typeid(num), num);

	writefln("%s\t- %s", typeid(c), c);
	writefln("%s\t- %s", typeid(i), i);

return 0;
}
Apr 29 2012
prev sibling next sibling parent Andrej Mitrovic <andrej.mitrovich gmail.com> writes:
On 4/30/12, WhatMeWorry <kc_heaser yahoo.com> wrote:

 I'm trying to get my head around D's type conversion. What is the
 best way to convert a string to a char array? Or I should say is
 this the best way?

 string s = "Hello There";
 char[] c;

 c = string.dup;


Well it depends . Why do you need a char[]? If you're interfacing with
C you most likely need a 0-terminated string and not a char[].


 Also, what is the best way to explicitly convert a string to an
 int?


import std.conv;
int i = to!int("12345");
Apr 29 2012
prev sibling parent reply Jonathan M Davis <jmdavisProg gmx.com> writes:
On Monday, April 30, 2012 01:42:38 WhatMeWorry wrote:

 I'm trying to get my head around D's type conversion. What is the
 best way to convert a string to a char array? Or I should say is
 this the best way?
 
 string s = "Hello There";
 char[] c;
 
 c = string.dup;


dup will return a mutable copy of an array. idup will return an immutable copy 
of an array. They will both always copy. If you want to convert without having 
to make a copy if the array is of the constancy that you want already (e.g. if 
a templated function is templated on string type, and it could be any 
constancy of char[]), then use std.conv.to.

auto c = to!(char[])(str);

If str was already char[], then it will just be returned, whereas if it's 
immutable(char)[], then it would dup it and return that.


 Also, what is the best way to explicitly convert a string to an
 int?  I've been looking at Library Reference (Phobos) but I'm
 stuck.


Use std.conv.to:

auto i = to!string("1234");

std.conv.to is what you use for pretty much any conversion.

- Jonathan M Davis
Apr 29 2012
parent reply Ary Manzana <ary esperanto.org.ar> writes:
On 4/30/12 8:08 AM, Jonathan M Davis wrote:

 On Monday, April 30, 2012 01:42:38 WhatMeWorry wrote:

 I'm trying to get my head around D's type conversion. What is the
 best way to convert a string to a char array? Or I should say is
 this the best way?

 string s = "Hello There";
 char[] c;

 c = string.dup;


 dup will return a mutable copy of an array. idup will return an immutable copy
 of an array. They will both always copy. If you want to convert without having
 to make a copy if the array is of the constancy that you want already (e.g. if
 a templated function is templated on string type, and it could be any
 constancy of char[]), then use std.conv.to.

 auto c = to!(char[])(str);

 If str was already char[], then it will just be returned, whereas if it's
 immutable(char)[], then it would dup it and return that.


 Also, what is the best way to explicitly convert a string to an
 int?  I've been looking at Library Reference (Phobos) but I'm
 stuck.


 Use std.conv.to:

 auto i = to!string("1234");

 std.conv.to is what you use for pretty much any conversion.

 - Jonathan M Davis


Can the documentation of std.conv be fixed?

http://dlang.org/phobos/std_conv.html#to

I mean, all the toImpl methods are documented, but in "to" it clearly 
says "Client code normally calls to!TargetType(value) (and not some 
variant of toImpl." I think all the documentation should be in "to". Now 
it sounds like you know what "to" does... but people read documentation 
because they don't know what it does.

There's also no need to document all the different parse methods in 
different places. Just one place is enough and simpler to read.
Apr 30 2012
parent "Jonathan M Davis" <jmdavisProg gmx.com> writes:
 Can the documentation of std.conv be fixed?
 
 http://dlang.org/phobos/std_conv.html#to
 
 I mean, all the toImpl methods are documented, but in "to" it clearly
 says "Client code normally calls to!TargetType(value) (and not some
 variant of toImpl." I think all the documentation should be in "to". Now
 it sounds like you know what "to" does... but people read documentation
 because they don't know what it does.


The comment on to could obviously be changed, but it wouldn't be all that fun 
to try and not have toImpl in the documentation. You'd have to do something 
like

version(StdDdoc)
{
 /++
 ddoc comment
 +/
 T to(T, S)(S value)
 if(...)
 {assert(0);}
}
else
{
 T toImpl(T, S)(S Value)
 if(...)
 {...}
}

for every toImpl, which would create undesirable code duplication on top of 
being ugly. You can certainly make the changes and create a pull request for 
them if you want to though. Another approach would simply be to put better 
examples in the documentation.

The weird thing is that there's a huge comment with a variety of examples near 
the top of the file that looks an awful lot like it's intended to be the ddoc 
comment for the module, but it's not actually at the top of the module, nor is 
it on any symbol (and it specifically has a space after its first asterisk so 
that it's not a ddoc comment). It's also, right above to, so maybe it was 
intended to be on to at one point. Either way, it's a bit weird.

- Jonathan M Davis
Apr 30 2012