error C2513 ifstream declaration

[forestPeterson]

so I need to edit the 'in' statement to 'ios::in' ?

and, why are they redefining 'in'?

[vlad from moscow]

I do not know about which statement you arre speaking. And I do not understand defining in as = in the header above.

[forestPeterson]

/add words from file into lexicon
void Lexicon::addWordsFromFile(string filename)
{
	ifstream ios::in;
	ios::in.open(filename.c_str());

	string line; //read in line
	
	//opens xml text file and reads in activity code, location and date fields
	while (true) {
		getline(ios::in, line); //read each line
		if (ios::in.fail()) break; //boilerplate check for error
		
		line.erase(0, 1); //remove first tab
		if(line == "</ScheduleTasks>") break; //check for end of document
		
		//add readin of location codes and locations to vector array for matching later

		line.erase(0, 1); //remove second tab
		if(line.substr(1, 15) == "ScheduleTask ID") //look for activity
			add (line.substr(42,16)); //store activity code
		line.erase(0, 2); //remove third and fourth tab
		if(line.substr(1, 17) == "Timing LocationID")//look for location
			add1 (line.substr(19,16)); //store location code
		line.erase(0, 1); //remove fifth tab
		if(line.substr(1, 13) == "Planned Begin") //look for dates
			add2 (line.substr(17,26), line.substr(41,50)); //store start date yyyy-mm-dd and store finish date yyyy-mm-dd
	}
	ios::in.close(); //boilerplate close file line
}

is this correct?

[vlad from moscow]

No, it is an invalid declaration

ifstream 'ios::in;


Your shall insert a name of the variable. You can use any other name instead of in.

But iit is much better to remove declaration of the manifest constant "in".

[forestPeterson]

so something like this will work,

void Lexicon::addWordsFromFile(string filename)
{
	ifstream ontology;
	ontology.open(filename.c_str());

	string line; //read in line
	
	//opens xml text file and reads in activity code, location and date fields
	while (true) {
		getline(ontology, line); //read each line
		if (ontology.fail()) break; //boilerplate check for error
		
		line.erase(0, 1); //remove first tab
		if(line == "</ScheduleTasks>") break; //check for end of document
		
		//add readin of location codes and locations to vector array for matching later

		line.erase(0, 1); //remove second tab
		if(line.substr(1, 15) == "ScheduleTask ID") //look for activity
			add (line.substr(42,16)); //store activity code
		line.erase(0, 2); //remove third and fourth tab
		if(line.substr(1, 17) == "Timing LocationID")//look for location
			add1 (line.substr(19,16)); //store location code
		line.erase(0, 1); //remove fifth tab
		if(line.substr(1, 13) == "Planned Begin") //look for dates
			add2 (line.substr(17,26), line.substr(41,50)); //store start date yyyy-mm-dd and store finish date yyyy-mm-dd
	}
	ontology.close(); //boilerplate close file line
}

but changing the foreach.h file and commenting the line
//static const ios::openmode IOS_IN = ios::in;

is better?

[vlad from moscow]

I meant the definition of manifest constant in.

[forestPeterson]

I can hack at this and //#define in = but I'd like to understand why is this here in the first place and why i should comment it out.

[Disch]

but I'd like to understand why is this here

It's there because someone was trying to be cute and redefine the C++ language by abusing macros. This is always a bad idea because it leads to problems like the ones you are having now.

Someone wanted to make C# style foreach loops in C++, and to do that, they tried to hijack 'in' and make it a keyword.

It was foolish. Whoever did it should be ashamed of themselves.

and why i should comment it out.

Because that #define is polluting the namespace, replacing all instances of 'in' with '='.

Unless you are using 'in' for the purpose of writing those foreach loops, that #define is of no value to you, and it is just wrecking other areas of your code. So it should be removed.

[forestPeterson]

that is a good clear explanation for the 'why'

there are three classes that use the foreach class, so if I change foreach then those classes may not work, they are: set.h, vector.h, and bst.h

I looked in bst.h but do not know what to look for to know how 'in' is used; this is the closest I found

/*
 * Method: iterator
 * Usage: iter = bst.iterator();
 * -----------------------------
 * This method creates an iterator that allows the client to
 * iterate through the elements in this binary search tree.  The
 * order of elements produced by the iterator is that of an InOrder
 * walk of the tree.
 *
 * The idiomatic code for accessing elements using an iterator is
 * to create the iterator from the collection and then enter a loop
 * that calls next() while hasNext() is true, like this:
 *
 *     BST<string>::Iterator iter = bst.iterator();
 *     while (iter.hasNext()) {
 *         string key = iter.next();
 *         . . .
 *     }
 *
 * This pattern can be abbreviated to the following more readable form:
 *
 *     foreach (string key in bst) {
 *         . . .
 *     }
 *
 * To avoid exposing the details of the class, the definition of the
 * Iterator class itself appears in the private/bst.h file.
 */
	Iterator iterator();

[forestPeterson]

I noticed that in some of the library classes they used 'cin' rather than 'in' - so I changed to that; I will leave the foreach class alone.

thank you for your help, the compiler errors for fstream are now resolved

[o3hC5Di1]

Hi there,

std::cind is a keyword in c++, I wouldn't change the name to that if I were you, it would only make things more ambiguous.

All the best,
NwN

[forestPeterson]

I used 'ontology'