I have a function which works with a std::ostream. I need to support using a C file handle (FILE*). Should I be creating my own subclass of std::ostream which delegates to a FILE*?
As Ben Voigt points out, you want to subclass streambuf. There are pages on the University of Southern California's website which have the documentation, header, and source for a GNU implementation of a streambuf subclass (stdiobuf) that wraps a FILE*. It has some dependencies on the library it is a part of (GroovX), but those should be easily to remove (I would begin by removing all references to GVX_TRACE).
Interestingly, it also provides a minimalistic subclass (stdiostream) of std::iostream, in spite of what Ben Voigt said. But this does not seem to be necessary, as the rdbuf ("read buffer"/set the stream buffer) method which the stdiostream class uses to connect the stdiobuf class to a stream object is publicly accessible.
You can find more about subclassing streambuf here (look particularly at the bottom of the page, which discussing the virtual functions). The implementation linked above overrides sync, underflow (to support input) and overflow (to support output).
Further notes about the linked implementation:
The init method uses the setg and setp methods to set the pointers for the input and output sequences.
The line const int num = pptr()-pbase(); is calculating the number of characters to flush by subtracting the base output pointer from the current output pointer ("put pointer").
The variable unhelpfully named om is the mode parameter.
The variable named fd is the file descriptor.
No, ostream is not meant to be derived from. The way the iostreams library allows customization is by supplying a streambuf pointer when creating an ostream. streambuf has a lot of virtual functions so you can change its behavior.
You need to derive either directly from streambuf or from the existing filebuf subclass. You probably only need to provide the overflow function, the defaults for all the others should work ok.
Related
I need some guidance or pointers understanding how to implement a custom ostream. My requirements are:
A class with a '<<' operator for several data types.
The intention is to send output to database. Each "line" should go to a separate record.
Each record most important field would be the text (or blob), but some other fields such as time, etc. can be mostly deduced automatically
buffering is important, as I don't want to go to database for every record.
First, does it worth deriving from ostream? What do I get by deriving from ostream? What if my class simply implements few operator<< methods (including some custom data types). Which functionality do I get from ostream?
Assuming what I want is a class derived from ostream, I need some guidance understanding the relationship between the ostream and the streambuf classes. Which one do I need to implement? Looking at some samples, it appears that I don't need to derive from ostream at all, and just give the ostream constructor a custom streambuf. Is that true? is that the canonical approach?
Which virtual functions at the custom streambuf do i need to implement? I've seen some samples (including this site: here and here, and few more), some override the sync method, and other override the overflow method. Which one should I override? Also, looking at the stringbuf and filebuf sources (Visual Studio or GCC) both those buffer classes implement many methods of the streambuf.
If a custom class derived from streambuf is required, would there be any benefit deriving from stringbuf (or any other class) instead of directly from streambuf?
As for "lines". I would like at least when my users of the class using the 'endl' manipulator to be a new line (i.e. record in database). Maybe - depends on effort - every '\n' character should be considered as a new record as well. Who do my custom ostream and/or streambuf get notified for each?
A custom destination for ostream means implementing your own ostreambuf. If you want your streambuf to actually buffer (i.e. don't connect to the database after each character), the easiest way to do that is by creating a class inheriting from std::stringbuf. The only function that you'll need to override is the sync() method, which is being called whenever the stream is flushed.
class MyBuf : public std::stringbuf
{
public:
virtual int sync() {
// add this->str() to database here
// (optionally clear buffer afterwards)
}
};
You can then create a std::ostream using your buffer:
MyBuf buff;
std::ostream stream(&buf)
Most people advised against redirecting the stream to a database, but they ignored my description that the database basically has a single blob field where all text is going to.
In rare cases, I might send data to a different field. This can be facilitated with custom attributes understood by my stream. For example:
MyStream << "Some text " << process_id(1234) << "more text" << std::flush
The code above will create a record in the database with:
blob: 'Some text more text'
process_id: 1234
process_id() is a method returning a structure ProcessID. Then, in the implementation of my ostream, I have an operator<<(ProcessID const& pid), which stores the process ID until it gets written. Works great!
The simplest way is to inherit std::streambuf and override just two methods:
std::streamsize xsputn(const char_type* s, std::streamsize n) – to append a given buffer with size provided to your internal buffer, std::string for example;
int_type overflow(int_type c) – to append a single char to your internal buffer.
Your streambuf can be constructed from whatever you want (DB connection for example). After append something into the internal buffer you may try to split it into lines and push something into DB (or just buffer an SQL statements to execute later).
To use it: just attach your streambuf to any std::ostream using constructor.
Simple! I've done something like this to output strings to syslog – everything works fine with any custom operator<< for user defined classes.
my2c - I think you are tackling this the wrong way. A stream may sound like a nice idea, but you'll need a way to indicate the end of the row too (and then what if someone forgets?) I would suggest something along the lines of how the java PreparedStatements and batches work, as in provide a set of methods which accept the types and a column index, then a "batch" method which explicitly makes it clear that you are indeed batching that row and then an execute to push the batch in.
Any stream based operation will rely on type (typically) to indicate which column to fill - but what if you have two ints? IMO, as a user, it doesn't feel like a natural way of inserting records into a database...
To add a new source or destination of character input/output to the iostreams mechanism, you should create a new streambuf class. The task of the stream buffer classes is to communicate with the 'external device' that will store the characters and to provide buffering facilities.
The problem with using iostreams to communicate with your database is that a database table does not match with the concept of a character sequence. A bit like pushing a round peg in a square hole. A streambuf only operates on characters. That is the only thing ever presented to it. This means the streambuf has to parse the character stream presented to it to find the field and record separators.
If you decide to go this route, I predict you will end up writing a CSV-to-SQL converter in your streambuf, just to get it working.
You will probably be better of with just adding a few operator<< overloads to your class(es). You could look at the Qt framework for ideas here. They also have the possibility to use operator<< to add items to a collections and such.
I often use the Boost Interprocess library, which provides an I/O stream class, bufferstream, which is basically similar to std::stringstream, except it operates over an arbitrary user-provided memory buffer.
I've used this stream class extensively, and I've also used it's source code to learn a lot about implementing/extending C++ I/O streams in general.
However, one piece of implementation code has struck me as very weird, and I'd like to see if anyone knows exactly why the Boost designer implemented it this way.
Specifically, it's a function in the streambuffer class boost::interprocess::basic_bufferbuf. This class inherits from std::basic_streambuf, and overrides all the virtual protected methods to provide specific behavior.
The source code can be browsed online here.
Specifically, the function basic_bufferbuf::seekpos() is implemented as follows:
virtual pos_type seekpos(pos_type pos, std::ios_base::openmode mode
= std::ios_base::in | std::ios_base::out)
{ return seekoff(pos - pos_type(off_type(0)), std::ios_base::beg, mode); }
Generally, stream buffer implementations which override std::basic_streambuf::seekpos() just call std::basic_streambuf::seekoff with the starting position set to std::ios_base::beg, so that's normal.
But why does the implementor write pos - pos_type(off_type(0)) here? Why subtract zero from the starting position? What possible use is that?
I suspect it must have something to do with unusual off_type or pos_type types, but I can't imagine what. Generally, pos_type and off_type are just signed integral types.
So why subtract zero here?
How can I override a C++ standard-library class function? In my application, I use ofstream objects in many different places of code. And now I want to open files in a different permission mode in Linux, Ubuntu. But open function of ofstream has no parameter to specify the permission mode of the file it creats.
Now I want to override open() function of ofstream class so it will get a parameter to specify the permissions for user access.
For starters, to clarify your terminology, the STL usually refers to the subset of the C++ standard library containing the containers, iterators, and algorithms. The streams classes are part of the C++ standard library, but are usually not bundled together with the STL. Some purists will insist that there is no such thing as the STL in the C++ standard library (since the STL is, technically speaking, a third-party library that was adopted into the standard), but most C++ programmers will know what you mean.
As for your question, there is no way within the standard to specify permission modes with ofstream. If you want to create your own custom stream class that is like ofstream but which supports permissions, your best bet would be to do the following:
Create a subclass of basic_streambuf that allows you to open, write, and possibly read files while specifying Unix permissions. The streams classes are designed so that the details of communicating with physical devices like disk, networks, and the console are all handled by the basic_streambuf class and its derived classes. If you want to make your own stream class, implementing a stream buffer would be an excellent first step.
Define your own class that subclasses basic_ostream and installs your custom basic_streambuf. By default, the basic_ostream supports all of the standard output routines by implementing them in terms of the underlying basic_streambuf object. Once you have your own stream buffer, building a basic_ostream class that uses that streambuf will cause all of the standard stream operations on that class (such as <<) to start making the appropriate calls to your streambuf.
If you'd like more details on this, an excellent reference is Standard C++ IOStreams and Locales. As a shameless plug, I have used the techniques from this book to build a stream class that wraps a socket connection. While a lot of the code in my stream won't be particularly useful, the basic structure should help you get started.
Hope this helps!
This is not answering your question directly as I wouldn't advise overriding ofstream::open.
Instead couldn't you use the first suggestion in this post? Open the file as you normally would to get the correct permissions, and then construct an ofstream from the file descriptor.
#include <iostream>
#include <fstream>
class gstream: public std::ofstream
{
void open(const std::string& filename, ios_base::openmode mode,int stuff)
{
//put stuff here
}
};
int main() {
gstream test;
//io stuff
return 0;
}
seems to work here.
Another option would be to create a wrapper class that contains an 'ofstream' object and has the interface you want, and passes the work onto its 'oftstream' member. It would look like this.
I have a function which works with a std::ostream. I need to support using a C file handle (FILE*). Should I be creating my own subclass of std::ostream which delegates to a FILE*?
As Ben Voigt points out, you want to subclass streambuf. There are pages on the University of Southern California's website which have the documentation, header, and source for a GNU implementation of a streambuf subclass (stdiobuf) that wraps a FILE*. It has some dependencies on the library it is a part of (GroovX), but those should be easily to remove (I would begin by removing all references to GVX_TRACE).
Interestingly, it also provides a minimalistic subclass (stdiostream) of std::iostream, in spite of what Ben Voigt said. But this does not seem to be necessary, as the rdbuf ("read buffer"/set the stream buffer) method which the stdiostream class uses to connect the stdiobuf class to a stream object is publicly accessible.
You can find more about subclassing streambuf here (look particularly at the bottom of the page, which discussing the virtual functions). The implementation linked above overrides sync, underflow (to support input) and overflow (to support output).
Further notes about the linked implementation:
The init method uses the setg and setp methods to set the pointers for the input and output sequences.
The line const int num = pptr()-pbase(); is calculating the number of characters to flush by subtracting the base output pointer from the current output pointer ("put pointer").
The variable unhelpfully named om is the mode parameter.
The variable named fd is the file descriptor.
No, ostream is not meant to be derived from. The way the iostreams library allows customization is by supplying a streambuf pointer when creating an ostream. streambuf has a lot of virtual functions so you can change its behavior.
You need to derive either directly from streambuf or from the existing filebuf subclass. You probably only need to provide the overflow function, the defaults for all the others should work ok.
Some days ago, I decided that it would be fun to write a streambuf subclass that would use mmap and read-ahead.
I looked at how my STL (SGI) implemented filebuf and realized that basic_filebuf contains a FILE*. So inheriting from basic_filebuf is out of the question.
So I inherited from basic_streambuf. Then i wanted to bind my mmapbuf to a fstream.
I thought the only thing that I would have to do would be to copy the implicit interface of filebuf... but that was a clear mistake. In the SGI, basic_fstream owns a basic_filebuf. No matter if I call basic_filestream.std::::ios::rdbuf( streambuf* ), the filestream completely ignores it and uses its own filebuf.
So now I'm a bit confused... sure, I can create my own mmfstream, that would be the exact copy/paste of the fstream but that sounds really not DRY-oriented.
What I can't understand, is: why does fstream is so tightly coupled with filebuf, so that it is not possible to use anything else than a filebuf? The whole point of separating streams and bufs is that one can use a stream with a different buffer.
Solutions:
=> filestream should rely on the implicit interface of filebuf. That is, fstream should be templated by a streambuf class. That would allow everyone to provide its own streambuf subclass to a fstream as long as it implements filebuf's implicit interface. Problem: we cannot add a template parameter to fstream since it would break template selectors while using fstream as template template parameter.
=> filebuf should be a pure virtual class without any additional attributes. So that one can inherit from it without carrying all its FILE* garbage.
Your ideas on the subject ?
In the IO streams' design, most of the actual streams' functionality (as opposed to the stream buffers' functionality) is implemented in std::basic_istream, std::basic_ostream, and their base classes. The string and file stream classes are more or less just convenience wrappers which make sure a stream with the right type of buffer is instantiated.
If you want to extend the streams, you almost always want to provide your own stream buffer class, and you almost never need to provide your own stream class. .
Once you have your own stream buffer type, you can then make it the buffer for any stream object you happen to have around. Or you derive your own classes from std::basic_istream, std::basic_ostream, and std::basic_iostream which instantiates your stream buffer and pass it to their base classes.
The latter is more convenient for users, but requires you to write some boiler-plate code for the buffer's instantiation (namely constructors for the stream class).
To answer your question: File streams and file buffer are coupled so tightly because the former only exists to ease the creation of the latter. Using a file stream makes it easy to set it all up.
Using your own stream class to wrap construction of your own stream buffer shouldn't be a problem, since you shouldn't be passing around file streams anyway, but only (references) to the base classes.
Check out mapped_file in the Boost.Iostreams library. I've never used used it myself, but it seems like it might already do what you need.
EDIT: Oops, reread your questions and I see you're doing this for fun. Perhaps you can draw inspiration from Boost.Iostreams?
fstream in itself is not a big class. It inherits from basic_stream to provide support for all the << and >> operations, contains a specialized steambuf that have to be initialized, and the corresponding constructors to pass the parameters to the streambuf constructor.
In a sense, what you wrote about your templated solution is OK. But basic_stream can also be derived into a tcp_stream for example. In that case, the constructors of fstream are a bit useless. Thus you need to provide a new tcpstream class, inheriting from basic_stream with the correct parameters for the constructors to be able to create the tcp_stream. In the end, you wouldn't use anything from fstream. Creating this new tcpstream is a matter of writing 3 or 4 functions only.
In the end, you would derive from the fstream class without any real reason to. This would add more coupling in the class hierarchy, unneeded coupling.
The whole point of std::fstream is that it is a _F_ile based std::stream. If you want an ordinary std::stream backed by your mmstreambuf, then you should create a mmstreambuf and pass it to std::stream::stream(std::streambuf*)