end line

[ema897]

I split a line into token. I have to compare with other word, but I notice that when I arrive at the end of the line, last token is for example:

endline\000\000 ecc..

Is there a way to eliminate this terminal char?

[Ganado]

I think more information is needed here. C-style strings are always supposed to be null-terminated. If you're reading past the null-termination, you're potentially reading into illegal memory. If you're converting things into std::strings, show us how you're doing that.

Show us exactly how you are splitting the line and what the data in and out is.

[jonnin]

it may be easier to code and more effiicent to strstr() or find() the compare word against the long string than to try to do it one by one over a tokenized group. Go back before you tokened may be better.
you can of course replace all 0 with space except the last one, in a copy. This should give you back the same thing you had before you tokened it, so that seems a bit like too much work somewhere (as above).

string ignores zeros entirely; it operates off the size variable alone.

[ema897]

string line;
string word;
myfile in;
int i=0;
bool flag=false;
myfile.getline(in,line);
while(i<line.size()){
        word.clear();
        if (isspace(line[i])){
            while(isspace(line[i])) i++;
        }
        if(ispunct(line[i])){
            if (line[i]=='('||line[i]==')') {
                     word += line[i];
                     flag = true;
            }else{
                    cout<<"error'"<<endl;
                    return;
                }
        }
        if(!flag) {
            while (!ispunct(line[i]) && !isspace(line[i])) {
                word += line[i];
                i++;
            }
            if(ispunct(line[i])){
                if (!(line[i]=='('||line[i]==')')){
                    cout<<"error"<<endl;
                    return ;
                }
            }
        }
        i++;
//compare
}
//other things 

[ema897]

Reading line:
a and b or c

for each iteration, string word will be:
a
and
b
or
c\000\000

[Ganado]

I get "error" printed when I modified your code to compile (and I used cin as input).

However, when I change your line[i] calls to line.at(i), an std::out_of_range exception was thrown. I didn't look hard enough to know where exactly you're going out of bounds, but you definitely are somewhere.

I suggest adding more bounds checks in your code. For example, in your loop on lines 22-25, you keep incrementing i when line[i] is NOT punct AND is NOT space. What happens if the last character is 'a'?

At some point, you're incrementing i past the bounds of your array before the next iteration of the outer loop.

[ema897]

getline(_my_file,line);
    while(i<line.size()){
        word.clear();
        if (isspace(line.at(i))){
            while(isspace(line.at(i))) i++;
        }
        if(i<line.size())
        if(ispunct(line.at(i))){
            if (line.at(i)=='('||line.at(i)==')') {
                     word += line.at(i);
                     flag = true;
            }else{
                    cout<<"error'"<<endl;
                    return ;
                }
        }
        if(!flag) {
            while (!ispunct(line.at(i)) && !isspace(line.at(i))) {
                word += line.at(i);
                i++;
                if(i==line.size()) break;
            }
            if(i<line.size())
            if(ispunct(line.at(i))){
                if (!(line.at(i)=='('||line.at(i)==')')){
                    cout<<"error'"<<endl;
                    return ;
                }
            }
        }
        i++;
        flag=false;
}

this way could be good?