Cannot open file created with ofstream

Question

Cannot open .cpp file in gedit editor on Ubuntu 11.10. The file was created with an ofstream object in a program I wrote in C++. The program compiles and runs without any errors, after which it opens a file, reads from it line by line (appending each line to a std::string object) until reaching the end, then returns the string to main(), which then outputs it to a separate file which it creates on the hard drive. However, when attempting to open and display the file in gedit, the editor just hangs and no text is displayed no matter how long I wait, and if I wait too long, when I try to close it, it will say its not responding and I have to force quit it. In the preview icon of the file in the nautilus file browser, it shows the output file as only having a single line of text, while the input file of course shows the whole page filled up with text. That should not be the case, since the output file should contain the same number of lines as the input file, only difference being a few lines of text were changed.

Anyone have any idea what could be causing this??

Note that the input file is quite long, and contains 136871 lines of code, so maybe that has something to do with it.

Here is the code of the program that creates the output file:

#include <iostream>
#include <fstream>
#include <string>

using namespace std;

string fixIssue_SegmentationFault(ifstream& file_handle);

int main() {

cout<< "Thank you for running FixC_html4_tagsIssues.\n"
       "This little program has been written with the intent of fixing\n"
       "the issues in the C_html4_tags.h file of the MAW programming project.\n"
    << endl
    << "Please hit Enter to fix the detected issues." <<endl;
cin.get();

cout<< "Thank you. Please wait as it fixes the issues..." <<endl;
ifstream C_html4_tags_file_handle("/home/mobilexman/Documents/Programming_Projects/MAW/Modules/core/WebCoder_mod/modules/Html_coder/C_html4_coder/C_html4_elements/C_html4_tags/C_html4_tags.cpp");
string output_str = fixIssue_SegmentationFault(C_html4_tags_file_handle);
C_html4_tags_file_handle.close();

ofstream C_html4_tags_replacement_file_handle("/home/mobilexman/Documents/Programming_Projects/MAW/Modules/core/WebCoder_mod/modules/Html_coder/C_html4_coder/C_html4_elements/C_html4_tags/C_html4_tags_replacement.cpp");
C_html4_tags_replacement_file_handle<< output_str.c_str();
C_html4_tags_replacement_file_handle.close();

cout<< "Congratulations. The issues have been fixed." <<endl;

return 0;

}

string fixIssue_SegmentationFault(ifstream& file_handle) {

//////////////////
//Cause of issue:
/////////////////
///////////////////////////////////////////////////////////////////
//map<S_browser, TYPE_attr_values> attr_supported_attr_values_map;
//...
//TYPE_attr_values& attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];
//...
//attr_supported_attr_values.clear();
//attr_supported_attr_values_map.clear();
//...
//attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];
/////////////////////////////////////////////////////////////////////////////
//Explanation:
//Inside constructor of C_html4_tags, I created a map for the supported attr values of each Html attribute I created for each Html tag in the constructor.
//The map uses an S_browser object for its key and an object of TYPE_attr_values (which is a typedef of std::vector<S_html_attr_value>) for its mapped value.
//Therefore, I created a TYPE_attr_values& called 'attr_supported_attr_values' to reference the associated mapped value vector of the 'dummy_browser' object,
//which is just a dummy browser object I created so I could create a list of standard Html 4.01 attributes which exist in the specification for now, and later
//create a list of attributes from that list for each browser, according to which attribute values each browser supports (note: this will be done at the
//next level up: C_html4_elements). The code line
////////////////////////////////////////////////////////////////////////////////////////////////
//TYPE_attr_values& attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];
///////////////////////////////////////////////////////////////////////////////////////////////
//is where I first create that reference and initialize it to the returned reference of the [] operator of the map, which performs an insert operation if the key passed does not already exist inside the map.
//And the code line
//////////////////////////////////////////////////////////////////////////////
//attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];
/////////////////////////////////////////////////////////////////////////////
//is where I reinitialize (or at least that's what I thought it did when I first wrote the code...) the reference to the returned reference again, for each Html attribute after the first attribute.
//I then attempt to clear after each attribute operations both the referenced vector and then the map before reusing the reference 'attr_supported_attr_values'.
//However, that is what created the segmentation fault. It turns out you cannot use the same reference on more than one object, unlike pointers. So it is invalid to reinitialize that reference
//to point to a new object for each attribute, thus creating the problem. Oddly enough, though, doing this didn't actually produce any compiler errors. It compiles ok, but when attempting to run the program,
//it will return a message saying the program unexpectedly exited when running it normal, and you only get the extra info about the segmentation fault when running the debugger on it.  I'm not sure why
//that is, but its probably by design. And so it did take me a bit to figure out why the 'attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];' lines was producing the segmentation
//fault, especially since the attribute it ended on was not actually the second attribute of the first tag, but rather the first attribute of the second tag. I still don't know why that is, but at any rate,
//I believe that after I change the 'attr_supported_attr_values' identifier to a pointer instead of a reference, and adjust all lines using it to treat it as a pointer instead of a reference, the problem
//SHOULD be fixed. But we'll certainly see how it goes...

//////////////////////////
//Steps for fixing issue:
////////////////////////
//1. Read each line of C_html4_tags.cpp until reaching the end, checking each line as we go to see if it contains the target string content. In addition, add the content of each line (regardless of whether it contains the target or not)
//to output_str. Note that the target string content is first "TYPE_attr_values& attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];" which needs to be changed to
//"TYPE_attr_values* attr_supported_attr_values = &attr_supported_attr_values_map[dummy_browser];", and then is the
/////////////////////////////////////
//attr_supported_attr_values.clear();
//attr_supported_attr_values_map.clear();
/////////////////////////////////////////
//lines which needs to be changed to
///////////////////////////////////////
//attr_supported_attr_values->clear();
//attr_supported_attr_values_map->clear();
//////////////////////////////////////////
//2. If the target is found in the current line, perform the necessary changes to the line's content (i.e. buffer_str) before adding it to output_str.
//3. Return output_str.
string output_str;
string buffer_str;
string search_str1 = "TYPE_attr_values& attr_supported_attr_values = attr_supported_attr_values_map[dummy_browser];";
string replacement_str1 = "TYPE_attr_values* attr_supported_attr_values = &attr_supported_attr_values_map[dummy_browser];";
string search_str2 = "attr_supported_attr_values.clear();";
string replacement_str2 = "attr_supported_attr_values->clear();";
string search_str3 = "attr_supported_attr_values_map.clear();";
string replacement_str3 = "attr_supported_attr_values_map->clear();";
size_t pos;
while (getline(file_handle, buffer_str)) {
     if (file_handle.good()) { //i.e. no errors while reading lines
        if ((pos = buffer_str.find(search_str1, 0)) != string::npos) {
           buffer_str.replace(pos, search_str1.size(), replacement_str1);
        }

        else if ((pos = buffer_str.find(search_str2, 0)) != string::npos) {
           buffer_str.replace(pos, search_str2.size(), replacement_str2);
        }

        else if ((pos = buffer_str.find(search_str3, 0)) != string::npos) {
           buffer_str.replace(pos, search_str3.size(), replacement_str3);
        }
        output_str += buffer_str;
     }
}

return output_str;

}

Answer 1

Your problem is that getline extracts and discards the line terminator, meaning that you end up without a single newline in your output. In all likelihood, this will make GEdit very unhappy (resulting in the behaviour you saw), since it has to calculate the line wrapping for a giant file.

Therefore, you should append a newline for each line as you are building the output string.