www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Using shorthand *= leads to unexpected result?

reply Michael <michael toohuman.io> writes:
It may be that I'm doing something wrong here, but after updating 
DMD to the latest version, my simulations started producing some 
very odd results and I think I've pinpointed it to a sign 
inversion that I was making. Here is some code from dpaste to 
demonstrate the behaviour I get vs. the behaviour I expected:
https://dpaste.dzfl.pl/9bd7aea75fb2

Any ideas? Am I getting something wrong or is this some sort of 
regression with the latest DMD?
May 15 2016
next sibling parent reply Saurabh Das <saurabh.das gmail.com> writes:
On Sunday, 15 May 2016 at 13:01:45 UTC, Michael wrote:

 It may be that I'm doing something wrong here, but after 
 updating DMD to the latest version, my simulations started 
 producing some very odd results and I think I've pinpointed it 
 to a sign inversion that I was making. Here is some code from 
 dpaste to demonstrate the behaviour I get vs. the behaviour I 
 expected:
 https://dpaste.dzfl.pl/9bd7aea75fb2

 Any ideas? Am I getting something wrong or is this some sort of 
 regression with the latest DMD?


Strangely, if I substitute the variable type 'auto' for 'float' 
or 'real', it gives the expected output of -1. Variable type 
'double' gives the incorrect value of 1.

This is strange behaviour. Perhaps someone more knowledgeable can 
shed light. If it is a regression, it is a very serious 
regression :(

~ sd
May 15 2016
parent reply Michael <michael toohuman.io> writes:
On Sunday, 15 May 2016 at 13:12:44 UTC, Saurabh Das wrote:

 On Sunday, 15 May 2016 at 13:01:45 UTC, Michael wrote:

 It may be that I'm doing something wrong here, but after 
 updating DMD to the latest version, my simulations started 
 producing some very odd results and I think I've pinpointed it 
 to a sign inversion that I was making. Here is some code from 
 dpaste to demonstrate the behaviour I get vs. the behaviour I 
 expected:
 https://dpaste.dzfl.pl/9bd7aea75fb2

 Any ideas? Am I getting something wrong or is this some sort 
 of regression with the latest DMD?


 Strangely, if I substitute the variable type 'auto' for 'float' 
 or 'real', it gives the expected output of -1. Variable type 
 'double' gives the incorrect value of 1.

 This is strange behaviour. Perhaps someone more knowledgeable 
 can shed light. If it is a regression, it is a very serious 
 regression :(

 ~ sd


Well I'm pretty sure the code was working just fine earlier in 
the week at the office, but running the code at home with the 
newest version of DMD started producing these odd results.
May 15 2016
parent Saurabh Das <saurabh.das gmail.com> writes:
On Sunday, 15 May 2016 at 13:25:42 UTC, Michael wrote:

 Well I'm pretty sure the code was working just fine earlier in 
 the week at the office, but running the code at home with the 
 newest version of DMD started producing these odd results.


Typing this function into asm.dlang.org shows a minor difference 
in the assembly generated:

auto f1(double d1)
{
   d1 *= -1.0;
   return d1;
}

With DMD 2.070.2:
pure nothrow  nogc  safe double example.f1(double):
  push   %rbp
  mov    %rsp,%rbp
  sub    $0x10,%rsp
  movsd  %xmm0,-0x8(%rbp)
  xorb   $0x80,-0x1(%rbp)
  rex.W movsd  -0x8(%rbp),%xmm0
  leaveq
  retq
  nopl   0x0(%rax)

With DMD 2.071.0:
pure nothrow  nogc  safe double example.f1(double):
  push   %rbp
  mov    %rsp,%rbp
  sub    $0x10,%rsp
  movsd  %xmm0,-0x8(%rbp)
  xorb   $0x80,-0x8(%rbp)
  rex.W movsd  -0x8(%rbp),%xmm0
  leaveq
  retq
  nopl   0x0(%rax)

The xorb line is different and I conjecture that it is causing 
the bug. I have never used floating-point assembly instructions, 
so I cannot understand what the logic is here. I'll read up and 
try to interpret to confirm.

Hope this helps.
May 15 2016
prev sibling parent reply Marc =?UTF-8?B?U2Now7x0eg==?= <schuetzm gmx.net> writes:
On Sunday, 15 May 2016 at 13:01:45 UTC, Michael wrote:

 It may be that I'm doing something wrong here, but after 
 updating DMD to the latest version, my simulations started 
 producing some very odd results and I think I've pinpointed it 
 to a sign inversion that I was making. Here is some code from 
 dpaste to demonstrate the behaviour I get vs. the behaviour I 
 expected:
 https://dpaste.dzfl.pl/9bd7aea75fb2

 Any ideas? Am I getting something wrong or is this some sort of 
 regression with the latest DMD?


It's a regression introduced here:
https://github.com/dlang/dmd/pull/5534

I've filed a bug report:
https://issues.dlang.org/show_bug.cgi?id=16027
May 15 2016
parent Michael <michael toohuman.io> writes:
On Sunday, 15 May 2016 at 14:12:47 UTC, Marc Schütz wrote:

 On Sunday, 15 May 2016 at 13:01:45 UTC, Michael wrote:

 It may be that I'm doing something wrong here, but after 
 updating DMD to the latest version, my simulations started 
 producing some very odd results and I think I've pinpointed it 
 to a sign inversion that I was making. Here is some code from 
 dpaste to demonstrate the behaviour I get vs. the behaviour I 
 expected:
 https://dpaste.dzfl.pl/9bd7aea75fb2

 Any ideas? Am I getting something wrong or is this some sort 
 of regression with the latest DMD?


 It's a regression introduced here:
 https://github.com/dlang/dmd/pull/5534

 I've filed a bug report:
 https://issues.dlang.org/show_bug.cgi?id=16027


Thanks for filing the bug report, and good find with the 
regression.
May 15 2016