Question

well i have to process a large chunk of text, analysing it linear from begin to end. And i wonder what is the better approach for this: using char* or std::string.
while using char* i can alter the pointer to a position further in the string eg.

//EDIT later: mallocing some space for text
char str[] = "text to analyse";
char * orig = str;
//process
str += processed_chars; //quite fast
//process again
// later: free(orig);

but using string i might have to use std::string::erase – but it create a copy, or move bytes or something (i don’t know the actual implementation)

string str = "text to analyse";
//process
str = str.erase(0,processed_chars);

or is there a way to alter the std::string’s hidden pointer?

EDIT: as Sylvain Defresne requested here more code:

class tag {
public:
    tag(char ** pch) {
        *pch = strstr(*pch,"<");
        if(pch == NULL) return;

        char *orig = *pch+1;
        *pch = strstr(*pch,">");
        if(pch == NULL) return;
        *pch+=sizeof(char); //moving behind the >

        //process inner tag data

        if(*(*pch-2)!='/'){// not selfclose
            while (!(**pch == '<' && *(*pch+1) == '/')){ //sarch for closing tag
                tag* kid = new tag(pch);
                sublings.push_back(*kid);
            }

            *pch = strstr(*pch,">");
            if(pch == NULL) return;
            *pch+=sizeof(char); //moving behind the >

            //add check if the clothing tag is matching

        }
    }
}

i use it for recursive xml-like notation parsing

char str[] ="<father><kid /></fatherr>";
char * pch = str;
tag *root = new tag(&pch);

this code is ugly as hell, i am just starting with low-level pointer arithmetic and stuff, used visual components till now so don’t judge too hard

Answer 1

With std::string, you would probably use std::string::iterator. Your code would be:

std::string str = "text to analyse";
std::string::iterator iter = str.begin();

// process
iter += processed_chars;