So here's an interesting question, How would I make something kinda like a wrapper for cout?
I want to be able to add it into a dll so I can throw it into my programs. but the basic syntax of it should be
Mything::mesage << "I'm some text" << im_an_int << someclass << mything::endl;
or
Mything::mesageandlog << "I'm going to print to console, and to a file!" << mything::endl;
I can handle most of the internal logic but as to what I should put to even do this. kinda stumped.
Possibly make a static stream member in my class called message, then have an event fire when its written too that runs it through a method?
Idk, I looked around and found something sortA similar, but as for throwing it into a dll I'm at a loss. (How to write a function wrapper for cout that allows for expressive syntax?)
because this requires me to use extern and a variable, but how would I make it static so I can just straight call it without creating a variable?
Bit of clarification, something like this:
mydll.h
#include <iostream>
namespace mynamespace {
extern struct LogMessage{};
template <typename T>
LogMessage& operator<< (LogMessage &s, const T &x) {
SetStdHandle(STD_OUTPUT_HANDLE, GetStdHandle(STD_OUTPUT_HANDLE));
SetConsoleTextAttribute(GetStdHandle(STD_OUTPUT_HANDLE),FOREGROUND_BLUE);
std::cout << "[IF] ";
SetConsoleTextAttribute(GetStdHandle(STD_OUTPUT_HANDLE),FOREGROUND_WHITE);
//LogTimestamp(); --ill impliment this.
std::cout << x << endl;
//writeStreamToLogfile(s); --and ill handle this.
return s;
}
}
driverprogram.h
#include <mydll.h>
#include <iostream>
int _tmain(int argc, _TCHAR* argv[])
{
mynamespace::LogMessage << "Something: << std::endl;
}
expected output:
"[IF] [00:00:00] Something
You can create a struct, that has a << operator
struct OutputThing
{
template< class T >
OutputThing &operator<<( T val )
{
std::cout<<val;
return *this;
}
};
Now whenever you want to log, you will have to instance the object.
OutputThing()<<"x ="<<x;
If you want to avoid the repeated construction and destruction of the object, you can make it a singleton.
struct OutputThingSingleton
{
static OutputThingSingleton& GetThing()
{
static OutputThingSingleton OutputThing;
return OutputThing;
}
template< class T >
OutputThingSingleton &operator<<( T val )
{
std::cout<<val;
return *this;
}
private:
OutputThingSingleton()
{};
};
So the call now looks like
OutputThingSingleton::GetThing()<<"x ="<<x;
Which you could shorten using a macro.
This will work across multiple dlls, however depending on how it is used you can have multiple instances of the singleton existing. This would work fine as long as you don't want to maintain any state in your singleton. If you do need to ensure a single instance, you can compile it in its own dll. Any other binary that uses this dll will share the single instance 'owned' by the dll.
First of all, just to give fair warning, I'm pretty sure this won't work in a DLL. You want to put it into a header (as it's shown here).
Second, it's probably a little more elaborate than you were considering. In particular, it defines a multi-output stream class that works like any other stream. Essentially any normal overload of operator<< should work fine with it.
Unlike a normal stream operator, however, the output goes to multiple streams, and each line of output (on all the streams) is preceded by a prefix (currently set to the value "[FIX]", but it just uses the content of a string, so whatever you put in that string should work. A more polished/finished implementation would probably allow you to set the prefix with something like a manipulator, but this (currently) doesn't support that.
Finally, it does some variadic template trickery, so you can specify the output files as either file names or existing ostream objects, or a combination thereof (e.g., see demo main at end).
First, the header:
#ifndef LOGGER_H_INC_
#define LOGGER_H_INC_
#include <iostream>
#include <streambuf>
#include <vector>
#include <fstream>
class logger: public std::streambuf {
public:
logger(std::streambuf* s): sbuf(s) {}
~logger() { overflow('\n'); }
private:
typedef std::basic_string<char_type> string;
int_type overflow(int_type c) {
if (traits_type::eq_int_type(traits_type::eof(), c))
return traits_type::not_eof(c);
switch (c) {
case '\n':
case '\r': {
prefix = "[FIX]";
buffer += c;
if (buffer.size() > 1)
sbuf->sputn(prefix.c_str(), prefix.size());
int_type rc = sbuf->sputn(buffer.c_str(), buffer.size());
buffer.clear();
return rc;
}
default:
buffer += c;
return c;
}
}
std::string prefix;
std::streambuf* sbuf;
string buffer;
};
namespace multi {
class buf : public std::streambuf {
std::vector<std::streambuf *> buffers;
public:
typedef std::char_traits<char> traits_type;
typedef traits_type::int_type int_type;
buf() {}
void attach(std::streambuf *s) { buffers.push_back(s); }
void attach(std::ofstream &s) { buffers.push_back(s.rdbuf()); }
int_type overflow(int_type c) {
bool eof = false;
for (std::streambuf *buf : buffers)
eof |= (buf->sputc(c) == traits_type::eof());
return eof ? traits_type::eof() : c;
}
};
class logstream : public std::ostream {
std::vector<std::ofstream *> streams;
buf outputs;
logger log;
void attach(std::ostream &s) { outputs.attach(s.rdbuf()); }
void attach(char const *name) {
std::ofstream *s = new std::ofstream(name);
streams.push_back(s);
outputs.attach(s->rdbuf());
}
template <typename T, typename...pack>
void attach(T &t, pack&...p) {
attach(t);
attach(p...);
}
public:
template <typename...pack>
logstream(pack&...p) : log(&outputs), std::ostream(&log) { attach(p...); }
~logstream() {
for (auto d : streams) {
d->close();
// Bug: crashes with g++ if delete is allowed to execute.
//delete d;
}
}
};
}
#endif
Then the demo of how to use it:
#include "logger"
int main(){
multi::logstream l(std::cout, "c:/path/log.txt");
l << "This is a prefixed string\n";
}
Obviously the header is fairly large, but the code to use it seems (at least to me) about as simple as you can hope for -- create an object, specifying where you want the output to go, just a normal stream -- except that you can specify more than one. Then write to it like you would to any other stream, and the output goes to all of the specified outputs, with each line preceded by the specified prefix.
Related
I want to use file streams and consol output stream. In constructor, I want to initialize with either file or consol output stream depending on parameter passed to constructor. Then I will have another function in class which will redirect output to that stream. what will be the code for it? I am trying with the below code which is not working.
Any other design suggestions are welcome.
class Test
{
private:
std::ios *obj;
std::ofstream file;
std::ostream cout1;
public:
// Test(){}
Test(char choice[])
{
if(choice=="file")
{
obj=new ofstream();
obj->open("temp.txt");
}
else
obj=new ostream();
}
void printarray()
{
for(int i=0;i<5;i++)
(*obj)<<"\n \n"<<"HI"
}
};
Something like this should work:
#include <iostream>
#include <fstream>
#include <string>
class Test
{
private:
std::ofstream file;
std::ostream& obj;
public:
// Use overloaded constructors. When the default constructor is used,
// use std::cout. When the constructor with string is used, use the argument
// as the file to write to.
Test() : obj(std::cout) {}
Test(std::string const& f) : file(f.c_str()), obj(file) {}
void printarray()
{
for(int i=0;i<5;i++)
obj<<"\n " << "HI" << " \n";
}
};
int main()
{
Test a;
a.printarray();
Test b("out.txt");
b.printarray();
}
PS Look at the changes to printarray. What you were trying, with %s, is good for the printf family of functions but not for std::ostream.
Any other design suggestions are welcome.
Two of these members are useless:
std::ios *obj;
std::ofstream file;
std::ostream cout1;
You can't do anything with a std::ios, a std::ostream that isn't associated with a streambuf is useless, and you never use file or cout1 anyway!
You want:
std::ofstream file;
std::ostream& out;
as shown in R Sahu's answer, and write to out.
Test(char choice[])
{
if(choice=="file")
This doesn't work, you need to use strcmp to compare char strings. You should probably use std::string not char*.
I am trying to create a Log class for my project at school. It needs to either be able to write information to the stdout or to a file depending on the parameters it is passed. I was looking into how to do this and I stumbled upon a thread with a similar question here: Obtain a std::ostream either from std::cout or std::ofstream(file)
The only difference between this thread and my own is that I want to do it inside of a class. Looking at the solution though they use std::ostream out(buf) and construct the ostream on the fly with buf. How can i declare this properly in my Log class to be able to construct the "out" object only once i enter my Log constructor?
I took a quick stab at it below but I am not sure if it is correct or if I am on the right track. Appreciate any help, thanks.
EDIT: I want to be able to do out << "Some string" << endl; after i get this Log class working properly.
EDIT2: An error I am now receiving with the new code below error : 'std::basic_ostream<_CharT, _Traits>::basic_ostream() [with _CharT = char, _Traits = std::char_traits<char>]' is protected
// log.h
#include <string>
#include <fstream>
#ifndef LOG_H_
#define LOG_H_
class Log
{
public:
enum Mode { STDOUT, FILE };
// Needed by default
Log(const char *file = NULL);
~Log();
// Writing methods
void write(char *);
void write(std::string);
private:
Mode mode;
std::streambuf *buf;
std::ofstream of;
std::ostream out;
};
#endif
// log.cpp
#include "log.h"
#include <iostream>
#include <stdlib.h>
#include <time.h>
Log::Log(const char *file)
{
if (file != NULL)
{
of.open(file);
buf = of.rdbuf();
mode = FILE;
}
else
{
buf = std::cout.rdbuf();
mode = STDOUT;
}
// Attach to out
out.rdbuf(buf);
}
Log::~Log()
{
if (mode == FILE)
of.close();
}
void Log::write(std::string s)
{
out << s << std::endl;
}
void Log::write(char *s)
{
out << s << std::endl;
}
You create tmp with std::ostream tmp(buf); and store the address of it in out with this->out = &tmp;. However, tmp will go out of scope at the end of the constructor and the pointer will no longer be pointing at a valid object.
What you should do instead is make out not a std::ostream* but simply a std::ostream:
std::ostream out;
Then in your constructor, once you've got the buf ready, you can give it to out by doing out.rdbuf(buf);.
Response to edit:
The std::ostream doesn't have a default constructor - it has to take a buffer pointer. My mistake. However, the fix is simple. Use your constructor's member initialization list to pass a null pointer (nullptr in C++11, 0 or NULL in C++03):
Log::Log(const char *file)
: out(nullptr)
{
// ...
}
My current implementation, simplified:
#include <string>
#include <memory>
class Log
{
public:
~Log() {
// closing file-descriptors, etc...
}
static void LogMsg( const std::string& msg )
{
static std::unique_ptr<Log> g_singleton;
if ( !g_singleton.get() )
g_singleton.reset( new Log );
g_singleton->logMsg( msg );
}
private:
Log() { }
void logMsg( const std::string& msg ) {
// do work
}
};
In general, I am satisfied with this implementation because:
lazy instantiation means I don't pay unless I use it
use of unique_ptr means automatic cleanup so valgrind is happy
relatively simple, easy-to-understand implementation
However, the negatives are:
singletons aren't conducive to unit-testing
dissonance in the back of my mind for introducing a pseudo-global (a bit of a code smell)
So here are my questions directed towards those developers who are successful in exorcising all singletons from their C++ code:
What kind of non-Singleton implementation do you use for application-wide logging?
Is the interface as simple and accessible as a Log::LogMsg() call above?
I want to avoid passing a Log instance all over my code, if at all possible - note: I am asking because, I, too, want to exorcise all Singletons from my code if there is a good, reasonable alternative.
First: the use of std::unique_ptr is unnecessary:
void Log::LogMsg(std::string const& s) {
static Log L;
L.log(s);
}
Produces exactly the same lazy initialization and cleanup semantics without introducing all the syntax noise (and redundant test).
Now that is out of the way...
Your class is extremely simple. You might want to build a slightly more complicated version, typical requirements for log messages are:
timestamp
level
file
line
function
process name / thread id (if relevant)
on top of the message itself.
As such, it is perfectly conceivable to have several objects with different parameters:
// LogSink is a backend consuming preformatted messages
// there can be several different instances depending on where
// to send the data
class Logger {
public:
Logger(Level l, LogSink& ls);
void operator()(std::string const& message,
char const* function,
char const* file,
int line);
private:
Level _level;
LogSink& _sink;
};
And you usually wrap the access inside a macro for convenience:
#define LOG(Logger_, Message_) \
Logger_( \
static_cast<std::ostringstream&>( \
std::ostringstream().flush() << Message_ \
).str(), \
__FUNCTION__, \
__FILE__, \
__LINE__ \
);
Now, we can create a simple verbose logger:
Logger& Debug() {
static Logger logger(Level::Debug, Console);
return logger;
}
#ifdef NDEBUG
# define LOG_DEBUG(_) do {} while(0)
#else
# define LOG_DEBUG(Message_) LOG(Debug(), Message_)
#endif
And use it conveniently:
int foo(int a, int b) {
int result = a + b;
LOG_DEBUG("a = " << a << ", b = " << b << " --> result = " << result)
return result;
}
The purpose of this rant ? Not all that is a global need be unique. The uniqueness of Singletons is generally useless.
Note: if the bit of magic involving std::ostringstream scares you, this is normal, see this question
I'd go with the simple, pragmatic solution:
you want a solution that is globally accessible. For the most part, I try to avoid globals, but for loggers, let's face it, it's usually impractical.
So, we do need something to be globally accessible.
But, we don't want the additional "there can be only one" restriction that a singleton confers. Some of your unit tests might want to instantiate their own private logger. Others might want to replace the global logger, perhaps.
So make it a global. A plain old simple global variable.
This still doesn't fully solve the problem with unit testing, admittedly, but we can't always have everything we want. ;)
As pointed out in the comment, you need to consider the initialization order for globals, which, in C++, is partly undefined.
In my code, that is generally not a problem, because I rarely have more than one global (my logger), and I stick rigidly to a rule of never allowing globals to depend on each others.
But it's something you have to consider, at least.
I really like the following interface since it uses streaming. Of course you can add channels, time and thread information to it. Another possible extension is to use the __FILE__ and __LINE__ macros and add it as parameters to the constructor. You could even add a variadic template function if you do not like the stream syntax. If you want to store some configuration you could add them to some static variables.
#include <iostream>
#include <sstream>
class LogLine {
public:
LogLine(std::ostream& out = std::cout) : m_Out(out) {}
~LogLine() {
m_Stream << "\n";
m_Out << m_Stream.rdbuf();
m_Out.flush();
}
template <class T>
LogLine& operator<<(const T& thing) { m_Stream << thing; return *this; }
private:
std::stringstream m_Stream;
std::ostream& m_Out;
//static LogFilter...
};
int main(int argc, char *argv[])
{
LogLine() << "LogLine " << 4 << " the win....";
return 0;
}
// file ILoggerImpl.h
struct ILoggerImpl
{
virtual ~ILoggerImpl() {}
virtual void Info(std::string s) = 0;
virtual void Warning(std::string s) = 0;
virtual void Error(std::string s) = 0;
};
// file logger.h //
#include "ILoggerImpl.h"
class CLogger: public ILoggerImpl
{
public:
CLogger():log(NULL) { }
//interface
void Info(std::string s) {if (NULL==log) return; log->Info(s); }
void Warning(std::string s) {if (NULL==log) return; log->Warning(s); }
void Error(std::string s) {if (NULL==log) return; log->Error(s); }
//
void BindImplementation(ILoggerImpl &ilog) { log = &ilog; }
void UnbindImplementation(){ log = NULL; }
private:
ILoggerImpl *log;
};
// file: loggers.h //
#include "logger.h"
extern CLogger Log1;
extern CLogger Log2;
extern CLogger Log3;
extern CLogger Log4;
extern CLogger LogB;
/// file: A.h //
#include "loggers.h"
class A
{
public:
void foo()
{
Log1.Info("asdhoj");
Log2.Info("asdhoj");
Log3.Info("asdhoj");
}
private:
};
/// file: B.h //
#include "loggers.h"
class B
{
public:
void bar()
{
Log1.Info("asdhoj");
Log2.Info("asdhoj");
LogB.Info("asdhoj");
a.foo();
}
private:
A a;
};
////// file: main.cpp ////////////////
#include "loggers.h"
#include "A.h"
#include "B.h"
#include "fileloger.h"
#include "xmllogger.h"
CLogger Log1;
CLogger Log2;
CLogger Log3;
CLogger Log4;
CLogger LogB;
// client code
int main()
{
std::unique_ptr<ILoggerImpl> filelog1(new CFileLogger("C:\\log1.txt"));
Log1.BindImplementation(*filelog1.get());
std::unique_ptr<ILoggerImpl> xmllogger2(new CXmlLogger("C:\\log2.xml"));
Log2.BindImplementation(*xmllogger2.get());
std::unique_ptr<ILoggerImpl> xmllogger3(new CXmlLogger("C:\\logB.xml"));
LogB.BindImplementation(*xmllogger3.get());
B b;
b.bar();
return 0;
};
// testing code
///////file: test.cpp /////////////////////////////////
#include "loggers.h"
CLogger Log1;
CLogger Log2;
CLogger Log3;
CLogger Log4;
int main()
{
run_all_tests();
}
///////file: test_a.cpp /////////////////////////////////
#include "A.h"
TEST(test1)
{
A a;
}
TEST(test2, A_logs_to_Log1_when_foo_is_called())
{
A a;
std::unique_ptr<ILoggerImpl> filelog1Mock(new CFileLoggerMock("C:\\log1.txt"));
Log1.BindImplementation(*filelog1.get());
EXPECT_CALL(filelog1Mock Info...);
a.foo();
Log1.UnbindImplementation();
}
I have a class that I want to give an output stream as a member to, to wit:
class GameBase {
protected:
ofstream m_OutputWriter;
...
}
There is a method in this class that takes a string argument and opens m_OutputWriter to point to that file, so data may be output to that file by using the standard << operator;
However, what I would like is to make the stream point to cout by default, so that if the output path is not specified, output goes to the console output instead of to a file, and it will be completely transparent by the calling class, who would use
m_OutputWriter << data << endl;
to output the data to the predetermined destination. Yet, I have tried a couple of the other examples here, and none of them exactly seem to fit what I'm trying to do.
What am I missing here?
Why does the stream need to be a member?
struct GameBase {
void out(std::ostream& out = std::cout);
// ...
};
In addition to having an std::ofstream as a member, I would use a function that returns an std::ostream&.
For example:
class GameBase {
std::ofstream m_OutputWriter;
protected:
std::ostream& getOutputWriter() {
if (m_OutputWriter)
return m_OutputWriter;
else
return std::cout;
}
...
}
A fully-functioning example:
#include <iostream>
#include <ostream>
std::ostream& get() {
return std::cout;
}
int main() {
get() << "Hello world!\n";
}
I'd like to compose two (or more) streams into one. My goal is that any output directed to cout, cerr, and clog also be outputted into a file, along with the original stream. (For when things are logged to the console, for example. After closing, I'd like to still be able to go back and view the output.)
I was thinking of doing something like this:
class stream_compose : public streambuf, private boost::noncopyable
{
public:
// take two streams, save them in stream_holder,
// this set their buffers to `this`.
stream_compose;
// implement the streambuf interface, routing to both
// ...
private:
// saves the streambuf of an ios class,
// upon destruction restores it, provides
// accessor to saved stream
class stream_holder;
stream_holder mStreamA;
stream_holder mStreamB;
};
Which seems straight-forward enough. The call in main then would be something like:
// anything that goes to cout goes to both cout and the file
stream_compose coutToFile(std::cout, theFile);
// and so on
I also looked at boost::iostreams, but didn't see anything related.
Are there any other better/simpler ways to accomplish this?
You do have the right design—if you want to do this purely within the stdlib.
One thing: instead of teeing to each streambuf on every output, implement it to use the same put area as one of the streambufs it's given, and copy to the others on overflow and sync. This will minimize virtual calls, which is one of the goals of how streambufs work.
Alternatively, and if you want to only handle stdout & stderr (which is common), run your program through the standard Unix tee program (or the equivalent on your platform), either by doing it yourself when invoking the program, or within the program by forking, setting up the streams as appropriate, etc.
Edit: You got me thinking, and I should know how to get this right. Here's my first approximation. (When this breaks, you get to keep both pieces.)
#ifndef INCLUDE_GUARD_A629F54A136C49C9938CB33EF8EDE676
#define INCLUDE_GUARD_A629F54A136C49C9938CB33EF8EDE676
#include <cassert>
#include <cstring>
#include <streambuf>
#include <map>
#include <vector>
template<class CharT, class Traits=std::char_traits<CharT> >
struct basic_streamtee : std::basic_streambuf<CharT, Traits> {
typedef std::basic_ios<CharT, Traits> Stream;
typedef std::basic_streambuf<CharT, Traits> StreamBuf;
typedef typename StreamBuf::char_type char_type;
typedef typename StreamBuf::traits_type traits_type;
typedef typename StreamBuf::int_type int_type;
typedef typename StreamBuf::pos_type pos_type;
typedef typename StreamBuf::off_type off_type;
basic_streamtee() : _key_buf(0) {}
basic_streamtee(Stream& a, Stream& b) : _key_buf(0) {
this->pubimbue(a.rdbuf()->getloc());
_set_key_buf(a.rdbuf());
insert(a);
insert(b);
}
~basic_streamtee() {
sync();
for (typename std::map<Stream*, StreamBuf*>::iterator i = _bufs.begin();
i != _bufs.end();
++i)
{
StreamBuf* old = i->first->rdbuf(i->second);
if (old != this) {
old->pubsync();
}
}
}
// add this functionality?
// streambufs would be unconnected with a stream
// easy to do by changing _bufs to a multimap
// and using null pointers for the keys
//void insert(StreamBuf* buf);
//void remove(StreamBuf* buf);
void insert(Stream& s) {
sync();
if (!_bufs.count(&s)) {
if (!_key_buf) {
_set_key_buf(s.rdbuf());
}
_bufs[&s] = s.rdbuf(this);
}
}
void remove(Stream& s) {
sync();
typename std::map<Stream*, StreamBuf*>::iterator i = _bufs.find(&s);
if (i != _bufs.end()) {
StreamBuf* old = i->second;
i->first->rdbuf(i->second);
_bufs.erase(i);
if (old == _key_buf) {
_set_key_buf(_bufs.empty() ? 0 : _bufs.begin()->second);
}
}
}
private:
basic_streamtee(basic_streamtee const&); // not defined
basic_streamtee& operator=(basic_streamtee const&); // not defined
StreamBuf* _key_buf;
std::map<Stream*, StreamBuf*> _bufs;
void _set_key_buf(StreamBuf* p) {
//NOTE: does not sync, requires synced already
_key_buf = p;
_update_put_area();
}
void _update_put_area() {
//NOTE: does not sync, requires synced already
if (!_key_buf) {
this->setp(0, 0);
}
else {
this->setp((_key_buf->*&basic_streamtee::pbase)(),
(_key_buf->*&basic_streamtee::epptr)());
}
}
#define FOREACH_BUF(var) \
for (typename std::map<Stream*, StreamBuf*>::iterator var = _bufs.begin(); \
var != _bufs.end(); ++var)
// 27.5.2.4.1 Locales
virtual void imbue(std::locale const& loc) {
FOREACH_BUF(iter) {
iter->second->pubimbue(loc);
}
}
// 27.5.2.4.2 Buffer management and positioning
//virtual StreamBuf* setbuf(char_type* s, std::streamsize n); // not required
//virtual pos_type seekoff(off_type off, std::ios_base::seekdir way,
// std::ios_base::openmode which); // not required
//virtual pos_type seekpos(pos_type sp, std::ios_base::openmode which); // not required
virtual int sync() {
if (!_key_buf) {
return -1;
}
char_type* data = this->pbase();
std::streamsize n = this->pptr() - data;
(_key_buf->*&basic_streamtee::pbump)(n);
FOREACH_BUF(iter) {
StreamBuf* buf = iter->second;
if (buf != _key_buf) {
buf->sputn(data, n); //BUG: ignores put errors
buf->pubsync(); //BUG: ignroes errors
}
}
_key_buf->pubsync(); //BUG: ignores errors
_update_put_area();
return 0;
}
// 27.5.2.4.3 Get area
// ignore input completely, teeing doesn't make sense
//virtual std::streamsize showmanyc();
//virtual std::streamsize xsgetn(char_type* s, std::streamsize n);
//virtual int_type underflow();
//virtual int_type uflow();
// 27.5.2.4.4 Putback
// ignore input completely, teeing doesn't make sense
//virtual int_type pbackfail(int_type c);
// 27.5.2.4.5 Put area
virtual std::streamsize xsputn(char_type const* s, std::streamsize n) {
assert(n >= 0);
if (!_key_buf) {
return 0;
}
// available room in put area? delay sync if so
if (this->epptr() - this->pptr() < n) {
sync();
}
// enough room now?
if (this->epptr() - this->pptr() >= n) {
std::memcpy(this->pptr(), s, n);
this->pbump(n);
}
else {
FOREACH_BUF(iter) {
iter->second->sputn(s, n);
//BUG: ignores put errors
}
_update_put_area();
}
return n;
}
virtual int_type overflow(int_type c) {
bool const c_is_eof = traits_type::eq_int_type(c, traits_type::eof());
int_type const success = c_is_eof ? traits_type::not_eof(c) : c;
sync();
if (!c_is_eof) {
char_type cc = traits_type::to_char_type(c);
xsputn(&cc, 1);
//BUG: ignores put errors
}
return success;
}
#undef FOREACH_BUF
};
typedef basic_streamtee<char> streamtee;
typedef basic_streamtee<wchar_t> wstreamtee;
#endif
Now, this test is far from complete, but it seems to work:
#include "streamtee.hpp"
#include <cassert>
#include <iostream>
#include <sstream>
int main() {
using namespace std;
{
ostringstream a, b;
streamtee tee(a, b);
a << 42;
assert(a.str() == "42");
assert(b.str() == "42");
}
{
ostringstream a, b;
streamtee tee(cout, a);
tee.insert(b);
a << 42 << '\n';
assert(a.str() == "42\n");
assert(b.str() == "42\n");
}
return 0;
}
Put it together with a file:
#include "streamtee.hpp"
#include <iostream>
#include <fstream>
struct FileTee {
FileTee(std::ostream& stream, char const* filename)
: file(filename), buf(file, stream)
{}
std::ofstream file;
streamtee buf;
};
int main() {
using namespace std;
FileTee out(cout, "stdout.txt");
FileTee err(clog, "stderr.txt");
streambuf* old_cerr = cerr.rdbuf(&err.buf);
cout << "stdout\n";
clog << "stderr\n";
cerr.rdbuf(old_cerr);
// watch exception safety
return 0;
}
You mention having not found anything in Boost.IOStreams. Did you consider tee_device?
I would write a custom stream buffer that just forwards data to the buffers of all your linked streams.
#include <iostream>
#include <fstream>
#include <vector>
#include <algorithm>
#include <functional>
class ComposeStream: public std::ostream
{
struct ComposeBuffer: public std::streambuf
{
void addBuffer(std::streambuf* buf)
{
bufs.push_back(buf);
}
virtual int overflow(int c)
{
std::for_each(bufs.begin(),bufs.end(),std::bind2nd(std::mem_fun(&std::streambuf::sputc),c));
return c;
}
private:
std::vector<std::streambuf*> bufs;
};
ComposeBuffer myBuffer;
public:
ComposeStream()
:std::ostream(NULL)
{
std::ostream::rdbuf(&myBuffer);
}
void linkStream(std::ostream& out)
{
out.flush();
myBuffer.addBuffer(out.rdbuf());
}
};
int main()
{
ComposeStream out;
out.linkStream(std::cout);
out << "To std::cout\n";
out.linkStream(std::clog);
out << "To: std::cout and std::clog\n";
std::ofstream file("Plop");
out.linkStream(file);
out << "To all three locations\n";
}