I'm having two istream-related problems. Please can somebody help... :confused: Problem 1

void loadStuffFromStream (istream& stream)
{
	// goes too far!!
	stream.seekg( length, ios_base::cur );
}

***edited*** This seems to skip to an arbitrary position in the stream, instead of 'current position' + length bytes. Why does it do this? Problem 2

char ch;
int pos;
...
pos = stream.tellg();		// now pos=0
stream.get (ch);
pos = stream.tellg();		// now pos=3585 ???
stream.get (ch);
pos = stream.tellg();		// now pos=3586
stream.get (ch);
pos = stream.tellg();		// now pos=3587
stream.get (ch);
// etc...

Not so much a 'problem' as a matter of curiosity. Why is there a sudden leap in the read-position reported by tellg the first time I read from the stream? (N.B. It is reading the correct characters into ch!)

You'll have to use the read() method of ifstream to do that. Precisely. Which takes us back to my original question ... what structure can I read into instead of an old fashioned char* array? I just thought there might be an elegant way, using vectors or strings or something, that didn't involve allocating chars manually. As I said, maybe I just want the moon on a stick. :)

I don't want it to stop on 'newline'. I want it to read a fixed number of bytes. :confused:

Thanks for the reply. However I don't see how it will know to stop after (say) 20 characters? Without a delimeter, will it read the whole file into the string? I want to do the STL equivalent of fread( mystring, sizeof(char), length, inputfile ); or failing that, use something more elegant than a char[] (which I have to allocate myself) as an intermediate buffer.

At the moment I'm trying to make an effort to use std::string and std::ifstream in my work, instead of using "stdio.h". My main objective is to write elegant code -- performance is not an issue! :) I need to read a lot of blocks (of various sizes) from file, and append them to various string buffers. The problem is that functions like ifstream::read are designed to read into char* style buffers, not strings; however there seems to be some kind of relationship between streambufs and strings that I figured might make my life easier...? :confused:using namespace std; ... string mystring; ifstream inputfile; int length = 20; ... _// is there a better way to do this...?_ char* buffer = new char [length]; inputfile.read (buffer, length); mystring += buffer; delete[] buffer; _// this isn't much better..._ char ch; for (int i=0; i<length; i++) { inputfile >> ch; mystring += ch; _// presumably I could **reserve** extra space before the loop,_ _// to avoid repeated resizing of the string...?_ }
I have the gut feeling that there must be a really elegant way to do this, using some combination of streambuf / ifstream::operator>> / ifstream::read etc. I figured the whole point of using std::string was that it can worry about the allocation side of things behind the scenes. However I'm finding the documentation a little cryptic and generalised. :~ Any ideas? It's quite possible I just want the moon on a stick and I should go back to BASIC... :)

Kewl. Thanks for the info. Given that I'm not interested in the other languages, is there anything else about .NET Professional I should be particularly interested in? MS seem to imply that it has superior debugging facilities compared with the Standard edition, but if we're just talking fancy-pants stuff like remote debugging and such-like, then I'm not too fussed. I just want to write Visual C++ apps on my home PC (ultimately, with a view to distributing my apps). I use .NET Pro at work, so I'm wondering whether, when I buy C++ .NET Standard for home use, I'm suddenly going to realise there are a load of missing features that I badly want! :)

I understand that Microsoft are now giving their Visual C++ Toolkit away for free. :cool: Some websites claim that the compiler is equivalent to the Standard C++ compiler (i.e. not optimizing). On the other hand, Microsoft claim the compiler is the same as their Professional compiler. This seems too good to be true! :omg: Surely one of the main incentives for purchasing .NET Professional is that you get an optimizing compiler? :confused: In other words, what's to stop me just buying C++ .NET Standard (which comes pretty cheap) and, when I've finished developing my app, using the free toolkit to build the final, optimized, distributable exe? In a nutshell: what's the catch? :~