http://d.puremagic.com/issues/show_bug.cgi?id=7917

           Summary: --inline option fails for complex expressions
           Product: D
           Version: unspecified
          Platform: x86_64
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody@puremagic.com
        ReportedBy: russel@winder.org.uk


--- Comment #0 from Russel Winder <russel@winder.org.uk> 2012-04-15 08:45:43 PDT ---
Ignoring all issues of whether the code is good code or even idiomatic code, I am getting compilation problem with using --inline on complex expressions.

So, for example, http://www.russel.org.uk/Bazaar/Pi_Quadrature/pi_d2_parallelMapSequentialReduce.d works fine without --inline but with it fails with:

pi_d2_parallelMapSequentialReduce.d(32): Error: function
std.parallelism.TaskPool.amap!(partialSum).amap!(Result).amap cannot get frame
pointer to execute

However, perhaps more problematic is that
http://www.russel.org.uk/Bazaar/Pi_Quadrature/pi_d2_sequentialMap.d
http://www.russel.org.uk/Bazaar/Pi_Quadrature/pi_d2_sequentialMapParallelReduce.d
 work fine without the --inline option but fail with it giving the error:

Internal error: toir.c 178

It is not clear whether this latter is at all related to the former.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917


bearophile_hugs@eml.cc changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |bearophile_hugs@eml.cc


--- Comment #1 from bearophile_hugs@eml.cc 2012-04-15 15:19:26 PDT ---
(In reply to comment #0)

> However, perhaps more problematic is that
> http://www.russel.org.uk/Bazaar/Pi_Quadrature/pi_d2_sequentialMap.d
> http://www.russel.org.uk/Bazaar/Pi_Quadrature/pi_d2_sequentialMapParallelReduce.d
>  work fine without the --inline option but fail with it giving the error:
> 
> Internal error: toir.c 178

I suggest you to minimize that program to the minimal program that shows the same error, and show it here.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917



--- Comment #2 from Russel Winder <russel@winder.org.uk> 2012-04-16 04:41:09 PDT ---
Given the code:

import std.algorithm ;
import std.range ;
import std.stdio ;
import std.typecons ;

double partialSum ( immutable Tuple ! ( int , int , double ) data ) {
  return 1.0 ;
}

void execute ( immutable int numberOfTasks ) {
  immutable n = 1000000000 ;
  immutable delta = 1.0 / n ;
  immutable sliceSize = n / numberOfTasks ;
  immutable pi = reduce ! ( ( a , b ) => a + b ) ( 0.0 , map ! ( partialSum ) (
    map ! ( i => tuple ( i , cast ( int ) sliceSize , cast ( double ) delta ) )
( iota ( numberOfTasks ) ) ) ) ;
}

int main ( immutable string[] args ) {
  execute ( 1 ) ;
  return 0 ;
}


The problem lies with the cast of sliceSice.  Change the parameter to a literal and it goes away.  Change the definition of sliceSize and it goes away.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917



--- Comment #3 from bearophile_hugs@eml.cc 2012-04-16 04:56:08 PDT ---
(In reply to comment #2)
> import std.algorithm ;
> import std.range ;
> import std.stdio ;
> import std.typecons ;
> 
> double partialSum ( immutable Tuple ! ( int , int , double ) data ) {
>   return 1.0 ;
> }
> 
> void execute ( immutable int numberOfTasks ) {
>   immutable n = 1000000000 ;
>   immutable delta = 1.0 / n ;
>   immutable sliceSize = n / numberOfTasks ;
>   immutable pi = reduce ! ( ( a , b ) => a + b ) ( 0.0 , map ! ( partialSum ) (
>     map ! ( i => tuple ( i , cast ( int ) sliceSize , cast ( double ) delta ) )
> ( iota ( numberOfTasks ) ) ) ) ;
> }
> 
> int main ( immutable string[] args ) {
>   execute ( 1 ) ;
>   return 0 ;
> }

Maybe you are able to trim away some more stuff from that code, to minimize it some more.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917



--- Comment #4 from Russel Winder <russel@winder.org.uk> 2012-04-16 05:08:36 PDT ---
As far as I can see in the time available any further reduction loses the error message.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917



--- Comment #5 from bearophile_hugs@eml.cc 2012-04-17 11:15:02 PDT ---
(In reply to comment #4)
> As far as I can see in the time available any further reduction loses the error message.


A first reduction:

import std.algorithm: map;
import std.typecons: tuple, Tuple;
int foo(Tuple!(int)) {
    return 1;
}
void main() {
    int x = 2;
    map!foo(map!(_ => tuple(x))([3]));
}

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917



--- Comment #6 from bearophile_hugs@eml.cc 2012-04-17 11:37:14 PDT ---
I am hitting the same "Internal error: toir.c 178" even without -inline with this reduced code:



import std.algorithm: sort;
struct Foo(int m) {
    int x;
}
void bar(int m)(Foo!m[] foos) {
    void spam() {
        //sort!((Foo!m a, Foo!m b) => true)(foos); // no error
        sort!((a, b) => true)(foos);
    }
}
void main() {
    alias Foo!2 F2;
    bar([F2(1)]);
    alias Foo!3 F3;
    bar([F3(2)]);
}

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917


SomeDude <lovelydear@mailmetrash.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lovelydear@mailmetrash.com
           Platform|x86_64                      |All
         OS/Version|Linux                       |All


--- Comment #7 from SomeDude <lovelydear@mailmetrash.com> 2012-04-21 12:31:03 PDT ---
Also happens on 2.059 Win32

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917


Denis <verylonglogin.reg@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |verylonglogin.reg@gmail.com


--- Comment #8 from Denis <verylonglogin.reg@gmail.com> 2012-05-03 14:25:30 MSD ---
This issue is derived from Issue 7955. Everything you need to create this issue from Issue 7955 is to wrap lambda template in a nested function.

Code from Issue 7955 comment #1 with addition of `spam`:
---
void f(alias fun)() { }

void g(T)() {
    void spam() {
        f!(a => a)();
    }
}

void main() {
    g!int();
    g!long();
}
---

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7917


Walter Bright <bugzilla@digitalmars.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |bugzilla@digitalmars.com
         Resolution|                            |WORKSFORME


--- Comment #9 from Walter Bright <bugzilla@digitalmars.com> 2013-10-03 18:06:03 PDT ---
This compiles without error in 2.064 head.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------