Uber whitespace trimmer

[alcoheca]

Hi, me again!

I wish to tidy up strings from user input. I have done a bit of research and have most of this working except the tricky bit - converting all/any whitespace between words to 1 ' ' char.

For instance, using hyphen to represent spaces..
---John-----MacDonald-

should end up as

John-MacDonald

Here is my code, it is malfunctioning and I've been debugging but can't see where the fault lies, except that the str.erase (i) bit doesn't seem to work as I expect. I expect it to remove one character at position i. but it doesn't, it chops off the rest of the string normally.

void trimString(string &str)
{
  // trim Leading whitespace
  // inline int ? TODO

  // str = "  345  89"
  int i = 0;
  do
    if (str[i] != ' ' && str[i] != '\t')
      break;
  while (++i <= str.size());
  str.erase(0, i);  // now str = "345  89"

  // trim trailing whitespace
  i = str.size();
  while (i-- >= 0)
    if (str[i] != ' ' && str[i] != '\t')
      break;
  str.resize(++i); // no whitespace at end so str = "345  89"
  
  // trim each remaining whitespace to 1 character
  i = 0;
  while (++i < str.size()-1 /* No check of last char needed*/) {
    // first i==1 as first char != ' '
    if (str[i] == ' ' || str[i] == '\t') {
      cout << i; // Sanity check, correct value
      ++i;
      cout << i; // Sanity check, correct value
      if (str[i] == ' ' || str[i] == '\t') {
        str.erase(i-1); // now str = "345" ERROR!!!
        //str.erase(--i);
        cout << i;
        continue;
      }
    }
  } // end while
}

[Sturm]

Instead of doing all this crap with the string, why not just create a new string? isspace() is your friend.

[alcoheca]

Instead of doing all this crap with the string, why not just create a new string? isspace() is your friend.

Use if (isspace(ch)) to move past whitespaces. Then when find next non whitespace character, copy a space, then the character, then move on?

[Sturm]

Sure.

[alcoheca]

Still having problems, and the culprit is definelty the string.erase modifier.
Does this modifier have to be called by reference ( I mean by a pointer to a string) and the other statements like newstr do that by default?
Is that the problem?

At the moment
newstr.erase(j) // j=8
wipes the whole string.

string trimString(string str) {
  string newstr;
  int j = 0;
  for (int i=1; i<str.size(); i++) {
    if (isspace(str[i]) && isspace(str[i+1])) {
       newstr[j] = ' ';
       j++;
       i++;
    } 
    else {
      newstr[j] = str[i];
      j++;
    }
  }
  newstr.erase(j);
  return newstr;
}

[Sturm]

Why are you erasing anything in the first place?
here's the idea:

1. Increment an iterator until isspace() is false.
2. Increment another iterator until isspace() is true. You now have a word.
3. Add the word and a space to the new string.
4. Rinse and repeat.

[alcoheca]

Why are you erasing anything in the first place?
here's the idea:

1. Increment an iterator until isspace() is false.
2. Increment another iterator until isspace() is true. You now have a word.
3. Add the word and a space to the new string.
4. Rinse and repeat.

While watching the newstr value in VS debugger,
after newstr[j] = ' ';
the string was updated to be the same length as str and had {" ÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌ"} as it's value.
(First charcter is a space which it should have copied).
From that I presumed I had to resize the string to the correct value (last increment of j).

I can see your method working, so I'll give that a go.

[WaltP]

Still having problems, and the culprit is definelty the string.erase modifier.
Does this modifier have to be called by reference ( I mean by a pointer to a string) and the other statements like newstr do that by default?
Is that the problem?

At the moment
newstr.erase(j) // j=8
wipes the whole string.

Maybe you should look up how string.erase works. This took me 30 seconds to find rather than the two hours you've been posting.

[alcoheca]

Maybe you should look up how string.erase works. This took me 30 seconds to find rather than the two hours you've been posting.

I did read about all the modifiers etc of the string class, here: http://www.cplusplus.com/reference/string/string/

Just my lecturer in uni never wrote in his notes that I had to have using std::string; in the file.

Like you said I've just spent 2 hours redefining an algorithm when the fault was something I mentioned in the first post. I should have trusted my instincts :'(

Still, nothing like practice to keep you keen :P