www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - cannot access frame of function

reply Alex <sascha.orlov gmail.com> writes:
Hi all,
given this code:

import std.algorithm.iteration : sum, cumulativeFold;

void main()
{
	double[5] a;
	a = 0;
	foreach(el; a) assert(el == 0);
	a[0] = 1.0;
	a[1] = 2.0;
	a[2] = 3.0;
	a[3] = 4.0;
	a[4] = 5.0;
	foreach(el; a) assert(el != 0);
	auto asum = a[].sum;
	auto jProbs = a[].cumulativeFold!((a, b) => (a + b)/asum);
}

the last line does not compile.

I found
http://forum.dlang.org/post/mailman.4097.1499105927.31550.digitalmars-d-bugs puremagic.com
and
https://issues.dlang.org/show_bug.cgi?id=11886

What I do not understand, how my example differs from the fixed 
bug?
Sep 18 2017
parent reply user1234 <user1234 12.hu> writes:
On Monday, 18 September 2017 at 14:45:25 UTC, Alex wrote:

 Hi all,
 given this code:

 import std.algorithm.iteration : sum, cumulativeFold;

 void main()
 {
 	double[5] a;
 	a = 0;
 	foreach(el; a) assert(el == 0);
 	a[0] = 1.0;
 	a[1] = 2.0;
 	a[2] = 3.0;
 	a[3] = 4.0;
 	a[4] = 5.0;
 	foreach(el; a) assert(el != 0);
 	auto asum = a[].sum;
 	auto jProbs = a[].cumulativeFold!((a, b) => (a + b)/asum);
 }

 the last line does not compile.

 I found
 http://forum.dlang.org/post/mailman.4097.1499105927.31550.digitalmars-d-bugs puremagic.com
 and
 https://issues.dlang.org/show_bug.cgi?id=11886

 What I do not understand, how my example differs from the fixed 
 bug?


asum is a lazy range and the error comes from this.
Let's say that if you replace  asum by 1.23 then the code 
compiles.
Sep 18 2017
parent reply ag0aep6g <anonymous example.com> writes:
On 09/18/2017 08:25 PM, user1234 wrote:

 On Monday, 18 September 2017 at 14:45:25 UTC, Alex wrote:


[...]

 import std.algorithm.iteration : sum, cumulativeFold;

 void main()
 {
     double[5] a;




[...]>>     auto asum = a[].sum;
[...]

 asum is a lazy range and the error comes from this.


asum is not a range. It's a double.


 Let's say that if you replace  asum by 1.23 then the code compiles.


Doesn't work for me. This still fails compilation with the same error:

----
import std.algorithm.iteration : sum, cumulativeFold;
void main()
{
     double[5] a;
     auto asum = 1.23;
     auto jProbs = a[].cumulativeFold!((a, b) => (a + b)/asum);
}
----
Sep 18 2017
next sibling parent user1234 <user1234 12.hu> writes:
On Monday, 18 September 2017 at 18:49:54 UTC, ag0aep6g wrote:

 On 09/18/2017 08:25 PM, user1234 wrote:

 On Monday, 18 September 2017 at 14:45:25 UTC, Alex wrote:


 [...]

 import std.algorithm.iteration : sum, cumulativeFold;

 void main()
 {
     double[5] a;




 [...]>>     auto asum = a[].sum;
 [...]

 asum is a lazy range and the error comes from this.


 asum is not a range. It's a double.


 Let's say that if you replace  asum by 1.23 then the code 
 compiles.


 Doesn't work for me. This still fails compilation with the same 
 error:

 ----
 import std.algorithm.iteration : sum, cumulativeFold;
 void main()
 {
     double[5] a;
     auto asum = 1.23;
     auto jProbs = a[].cumulativeFold!((a, b) => (a + b)/asum);
 }
 ----


Aw right, i had put 1.23 directly in the lambda. So, OP, you can 
dismiss my comment.
Sep 18 2017
prev sibling parent reply Alex <sascha.orlov gmail.com> writes:
On Monday, 18 September 2017 at 18:49:54 UTC, ag0aep6g wrote:

 Doesn't work for me. This still fails compilation with the same 
 error:

 ----
 import std.algorithm.iteration : sum, cumulativeFold;
 void main()
 {
     double[5] a;
     auto asum = 1.23;
     auto jProbs = a[].cumulativeFold!((a, b) => (a + b)/asum);
 }
 ----


So, this is a bug, isn't it? I assume, I should file it then...
Sep 18 2017
parent reply Jordan Wilson <wilsonjord gmail.com> writes:
On Monday, 18 September 2017 at 21:58:39 UTC, Alex wrote:

 On Monday, 18 September 2017 at 18:49:54 UTC, ag0aep6g wrote:

 Doesn't work for me. This still fails compilation with the 
 same error:

 ----
 import std.algorithm.iteration : sum, cumulativeFold;
 void main()
 {
     double[5] a;
     auto asum = 1.23;
     auto jProbs = a[].cumulativeFold!((a, b) => (a + b)/asum);
 }
 ----


 So, this is a bug, isn't it? I assume, I should file it then...


I have this same issue. Anyone have any thoughts on a workaround?

For example, my attempt at an exponential moving average doesn't 
compile:
auto ema(Range,T) (Range rng, int period, T seed) {
     return rng.cumulativeFold!((a,b) => 
a+(2.0/(period+1))*(b-a))(seed.to!double);
}

So I ended up with this:
auto ema(Range,T) (Range rng, int period, T seed) {
     struct EMA(Range) {
         double currentValue;
         double weighting;
         Range rng;
         this (Range r, int p, double s) {
             currentValue = s;
             weighting = 2.0 / (p+1);
             rng = r;
         }
         auto front() { return currentValue; }
         auto popFront() {
             rng.popFront;
             if (!rng.empty){
                 currentValue += (rng.front - 
currentValue)*weighting;
             }
         }
         auto empty() { return rng.empty; }
     }
     return EMA!Range(rng,period,seed.to!double);
}

Thanks,

Jordan
Mar 22 2018
parent Alex <sascha.orlov gmail.com> writes:
On Friday, 23 March 2018 at 03:38:22 UTC, Jordan Wilson wrote:

 I have this same issue. Anyone have any thoughts on a 
 workaround?


For completeness:
https://issues.dlang.org/show_bug.cgi?id=17841
Mar 23 2018