Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

Control Flow

2 (2.2%)

Arrays, Strings, Pointers, and References

8 (9%)

Functions

4 (4.5%)

Basic object-oriented programming

30 (33.7%)

A bit more advanced OOP (Composition, Operator overloading, Inheritance, Virtual Functions)

18 (20.2%)

Templates

8 (9%)

Other (Explain)

4 (4.5%)

Working with files?  (Streams)

15 (16.9%)

Total Members Voted: 89

[Siquo]

It's kind of... well, would trying to access char* with char[something] work? From this discussion I assume that the answer is no, but I kind of recall some book saying that it would.

Yeah, but it works for a different reason. As eerr said, it's even more confusing if you have different things that actually work the same (and the caveat is: almost always).

[Supermikhail]

Uh. If so... I don't remember anymore in which cases you would use char* at all. It appears more transparent if to store a character array you use char[]. So, why would you want to store an array of anything using a pointer to the first member? The benefits of that I cannot fathom.

[Shades]

Uh. If so... I don't remember anymore in which cases you would use char* at all.

Pretty much any time you don't know the length of the array at initialisation.

[Virex]

You could also just need a pointer to a character somewhere

[Supermikhail]

Uh. If so... I don't remember anymore in which cases you would use char* at all.

Pretty much any time you don't know the length of the array at initialisation.

Don't you use dynamic lists for that? Argh. No. Is an pointer to an array comparable to a dynamic list?

You could also just need a pointer to a character somewhere

Virex, please don't confuse him any more!

Why would that be? I mean if you need a single character, and then you want to pass a pointer to it, why create it with the pointer, not like a character?

[Shades]

Pretty much any time you don't know the length of the array at initialisation.

Don't you use dynamic lists for that? Argh. No. Is an pointer to an array comparable to a dynamic list?

Pretty much I'd either use a stl::string for a char* (or for a char[] for that matter) and some form of stl::vector for dynamic list. Of course I'd also use boost::array for an array too but that is besides the point.

Unfortunately there are still cases where you'd have to use c rather than c++ so you end up using the pointers or companies where 'dangerous third party data types', such as stl, aren't allowed and frankly it's easier to use char* than to right your own clone of decent well tested classes.

Why would that be? I mean if you need a single character, and then you want to pass a pointer to it, why create it with the pointer, not like a character?

The only case I can think of is when for some reason you need another function to set the value of a char but it can't be the return value for one of many reasons, in which case you could pass your local char as a pointer for the function to modify. I can't imagine you'd want to do it that often

[Rose]

if a function needs to set the value of a few chars, then you need to use char* for each

[Supermikhail]

Yeah. It sets the value. But why would you initialise it through char*?

[qwertyuiopas]

You can reassign what a pointer points to, but not an array. Most of the time, string literals are not editable, so a pointer to a string literal cannor be altered as if it was an array. (char *c = "hi"; c[0] = 'n'; does not work, or if it does, you shouldn't rely on it, because many/all good compilers won't.)
A char[] is initialized with a copy of the string literal, as if you had used {'c', 'h', 'a', 'r', '\0'} rather than "char", so it is editable but not reassignable.

So, if you need to edit a string literal mid-function, use a char array. If you don't, but need it to point to different string literals, use a char pointer. If you need to do both, a pointer will work, but only if none of the edited strings are string literals, only char arrays created elsewhere. Or, in C++, simply use a string class rather than a char*/[], , and you don't have to worry about the difference. (Or something like char temp[] = "string"; char *p = temp;, but if you need to do that, then perhaps you should look for another way ot do it...)

[Normandy]

@Supermikhail:
No, nobody uses constructions like that. But you should be aware of how it works, it's rather fundamental.

Just keep in mind the difference between this:

Code: [Select]

const char a = 'a';
char b = a;and this:

Code: [Select]

const char *a = "a";
char *b = a
The first will work, the second will not. In C++, strings and chars are not interchangeable. On the top, you're copying the actual value in a to b, which is why it works. On the bottom, you're copying the pointer, which points to different value types (a points to const char, b points to char), which is why it doesn't work.

Other than that, qwertyuiopas covered it well.

[eerr]

char[] s = "string";
char[] s ={'s','t','r','i','n','g','\0'};

Are identical.
Cstrings are character arrays with \0 at the end.

\0 is the line return character (Enter buton)

[Shades]

\0 is the line return character (Enter buton)

No, \0 (0x00) is a null byte.

Carriage return is \r (0x0D), new line is \n (0x0A) and the enter button is either one, the other or both depending on your OS.
However you are correct in that a null byte marks the end of a cstring.

[nuker w]

Posting to watch. I used to do a bit of C++ and then kinda left it. Might get into it again, now that it seems theres a topic here I can ask questions on.

[Siquo]

I've been stuck for ages in my project now.

After a major rewrite of something totally unrelated, my OpenGL refuses to draw even the tiniest triangle. I can set the background-colour, that's it.

I know I must've screwed up somewhere, but there's so many places where this can go wrong. Bah, it takes all the fun out of programming.

[Shades]

I've been stuck for ages in my project now.

After a major rewrite of something totally unrelated, my OpenGL refuses to draw even the tiniest triangle. I can set the background-colour, that's it.

I know I must've screwed up somewhere, but there's so many places where this can go wrong. Bah, it takes all the fun out of programming.

Depending how much code it is I'd be happy to take a look at it if it would help find the error. I'm not bad with OpenGL stuffs. Also yes debugging is by far the most frustrating part of programming, which is why I unit test.