Book Review: A Wanted Man, Lee Child

A Wanted ManI’d read a few negative reviews of this Jack Reacher thriller before I read the book. I agree that it’s different to the majority of the books in the series: for one thing, the author concentrates on the cerebral side of Reacher’s character and concentrates too much on his exceptional memory for phone numbers and mathematical abilities (these aren’t pertinent to the plot so why keep referring to them?); for another, there’s far less action in this book (usually, Reacher’s run into trouble at a bar within the first few pages, here we’re into the finale before he gets into a fight). It’s not really up to the usual standards, but it’s still a better read than a couple of other books I failed to get into recently.
20130418-193203.jpg

2 Comments

Filed under Book Review

C++: functional programming and type classes

Functional C++ posted an article introducing type classes for C++.

Leave a comment

Filed under C++, Programming

C++: noexcept specification generated for destructors in C++11

Andrzej Krzemienski’s blog has this post on the subject of noexcept:

Even if you explicitly declare your destructor without any exception specification, the compiler will add one for you: the same as it would add if it had to implicitly generate the destructor. If you do not like the exception specification generated by the compiler, you have to explicitly specify it.

So if there’s a reason to throw from a destructor (which is typically frowned upon), you’d better declare noexcept(false).

Leave a comment

Filed under C++, Programming

C++: Resumable functions, Async and Await

Interesting post from Jens Weller on the new concurrency features in C++ 11 and C++ 14. He gives examples of code written using standard library async, futures and “.then()” with lamdbas, compared to the simplicity of expressing the same functionality with the proposed resumable functions (that is, resumable and await as language keywords).

Whilst resumable and await won’t be part of the language until at least C++17, Weller points out that Microsoft may put implementations into Visual Studio well before that. Here’s the paper, N3722.

Leave a comment

Filed under C++, Programming

Goldman Sachs automated trading error

Bloomberg reports that Goldman Sachs is the latest firm to suffer from a programming error in their automated trading software:

An internal system that Goldman Sachs uses to help prepare to meet market demand for equity options inadvertently produced orders with inaccurate price limits and sent them to exchanges

Participants in trades caused by error can appeal, and exchanges are reviewing and busting trades. It seems a large number of trades occurred because quotes entered the market at a price of $1:

Of the 500 biggest options trades in the first 15 minutes markets were open today, 405 of them were for tickers starting with H through L and priced at $1, according to data compiled by Trade Alert LLC and Bloomberg

.

Leave a comment

Filed under Finance

How to implement a “virtual” operator<

In order to support a collection of heterogenous objects, I implemented a std::set<std::shared_ptr<Base>> then inherited the derived classes from Base.  That works fine, except that the ordering and uniqueness is controlled by the object addresses, not the values of the derived objects.  Ideally, one would make operator< virtual on Base, but that won’t work here because the signature of operator< for each of the derived classes is different:

bool Base::operator<( const Base& rhs ) const
{  /* implementation */ }

bool A::operator<(const A& rhs ) const
{ /* implementation */  }

bool B::operator<(const B& rhs ) const
{ /* implementation */  }

Instead, this solution follows the “Non-Virtual Interface” idiom and leaves operator< non-virtual. Instead, operator< delegates to a separate virtual method, compares_less. In order to provide a strict-weak-ordering, the objects of distinct types need to be handled too – it’s not enough to fall back to comparing pointer addresses (the addresses could provide an inconsistent ordering).  Here, inter-type comparisons are handled by ordering on the type name (supplied by another virtual method).

// LessThan.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"

#include <iostream>
#include <set>
#include <string>

template<typename Container>
void dump( const Container& elements )
{
  for ( auto it = elements.begin(); it != elements.end(); ++it )
  {
    if ( it != elements.begin() )
    {
      std::cout << ", ";
    }
    std::cout << (*it);
  }
  std::cout << std::endl;
}

class Base
{
public:
  bool operator<( const Base& rhs ) const
  {
    return this->compares_less( rhs );
  }

  virtual std::ostream& dump( std::ostream& os ) const = 0;

  virtual bool compares_less( const Base& rhs ) const = 0;
  virtual std::string type_name() const = 0;
};

class A : public Base
{
  int a_;
public:
  A(int a) : a_(a)
  {}

  int a() const { return a_; }

  virtual std::ostream& dump( std::ostream& os ) const
  {
    os << a_;
    return os;
  }

  virtual bool compares_less( const Base& rhs ) const
  {
    if ( auto pA = dynamic_cast<const A*>( &rhs ) )
    {
      return a_ < pA->a_;
    }
    return this->type_name() < rhs.type_name();
  }

  virtual std::string type_name() const
  {
    return "A";
  }
};

class B : public Base
{
  std::string b_;
public:
  B( std::string b ) : b_(b)
  {}

  std::string b() const { return b_; }

  virtual bool compares_less( const Base& rhs ) const
  {
    if ( auto pB = dynamic_cast<const B*>( &rhs ) )
    {
      return b_ < pB->b_;
    }
    return this->type_name() < rhs.type_name();
  }

  virtual std::string type_name() const
  {
    return "B";
  }

  virtual std::ostream& dump( std::ostream& os ) const
  {
    os << b_;
    return os;
  }
};

std::ostream& operator<<( std::ostream& os, const std::shared_ptr<Base>& b )
{
  b->dump( os );
  return os;
}

struct LessThan
{
  bool operator()( const std::shared_ptr<Base>& lhs, const std::shared_ptr<Base>& rhs ) const
  {
    return *lhs < *rhs;
  }
};

int _tmain(int argc, _TCHAR* argv[])
{
  std::set<std::shared_ptr<Base>, LessThan> elements;

  elements.insert( std::make_shared<A>( 1 ) );
  elements.insert( std::make_shared<A>( 10 ) );
  elements.insert( std::make_shared<B>( "World" ) );
  elements.insert( std::make_shared<A>( 7 ) );
  elements.insert( std::make_shared<B>( "C++ says" ) );
  elements.insert( std::make_shared<A>( 3 ) );
  elements.insert( std::make_shared<B>( "Hello" ) );
  elements.insert( std::make_shared<A>( 5 ) );

  dump( elements );

  return 0;
}

Here’s the output, showing that the A’s and B’s are ordered as expected:
VirtualLessThanOutput

Leave a comment

Filed under C++, C++ Code, Programming

Herb Sutter GotW94: AAA Style (Almost Always Auto)

Herb Sutter’s Guru of the Week article on Almost Always Auto is a tour de force covering everything you always wanted to know about auto.

Leave a comment

Filed under C++, Programming

C++ Meet-Up London: Ben Hanson on Lexertl

I attended the London C++ Meet-Up on Tuesday 13th August where Ben Hanson presented his work on Lexertl.

The next London C++ meet up is Wednesday, 11 September 2013 at Google Campus, 5 Bonhill St, London, England EC2A.

Leave a comment

Filed under C++, Programming

Bruce Dawson on a VS Compiler bug and heap tracing tools

Bruce Dawson found a compiler bug – and demonstrated a couple of good tools for tracing heap allocations.

Leave a comment

Filed under Programming

Scott Meyers Video – Universal References

Scott Meyers posted a video of his Universal References presentation from C++ and Beyond 2012. There’s also an article on the subject in the ACCU magazine Overload Issue 111, but I found the video much clearer (and more entertaining).

Meyers coined the term ‘Universal References’ for instances where the type of a variable or parameter is declared T&& and where T is a deduced type. It may look like an rvalue-reference, but it can also match lvalue-references:

20130813-122646.jpg

so it’s usually a mistake to overload on both T&& and const T& in a template function. Another key point is that, whereas an overload with an rvalue-reference would call std::move (e.g. a move constructor where T is not deduced), an overload with a universal-reference should call std::forward (because you don’t yet know if it’s an rvalue- or lvalue-reference):

20130813-123009.jpg

He also summarises the reference-collapsing rules as “lvalue-references are infectious”, a handy mnemonic from Stephan T. Lavavej.

1 Comment

Filed under C++, Programming, Video