Programming

Hey
I have some questions regarding the old unix if command.
(see man pageman.cat-v.org/unix-6th/1/if

Here I uploaded the source code: (it's a bit too long, to put it here)
pastebin.com/bj0Hvfrw

Now my questions:

1.) Line 14: The function exp() is called with no arguments. But the function is declared as exp(s), so it needs an argument. Why is this working?


2.) What's happening there with exp() -> e1() -> e2() -> e3()... I think it's called recursive descent parsing, but I don't really get it. Could you help me a bit there.

Could you show us some code and usage of this 'old if.c' and explain what you are trying to do with it? That pastebin stuff is nonsense.

I am trying to understand the whole source code. I would like to know what each functions does, what each line does. So I picked some lines I do not understand and wrote the questions here.

Why is it nonsense? It's the acutal source code of the command if (unix ver.6).
How to use the command in unix is described in the posted man page.

I meant no offense.
I'm just going along with what you said. The questions you raised supports the fact that it doesn't make any sense, right?

if (exp()) is asking if the function exists; it's not calling that function.
p1 is not assigned to e1(), e2(), etc; p1 is a pointer to (the address of) those functions. By doing that it's redirecting the search for special characters (-r, -w, -c, etc.). The logic is: "If you don't find what you want here, I'll provide you another place to look", and so on.

A typical if statement would be something like...
if ( x = y ) { then do something
The source you provided checks for the characters '{} () = !=' and other options. In other words it's just checking for proper syntax and storing arguments for some later action. That's about all it's good for. Eventually it will return true or false or error.

The source code is old K&R C, without function declarations.

Don't write code like that, and don't ever modify old K&R C by adding function declarations - unless you like getting into the intricacies and implications of C variable promotion rules, and how they may have changed over the years.

Original K&R C just took all arguments to a function, promoted them so they'd all be the same size, and stuffed them on the stack.

I think, if arguments aren't declared after the first function definition line:

Code:

main(argc, argv) <--defintion
char *argv[];  <-- argument declaration
{
    ....

then the argument implicitly defaults to "int".

Basically, in K&R C all functions are called as variable argument functions with every argument promoted to the same size, and the arguments aren't type-checked. Ever. And argument declarations in the function definitions only tell the function how to interpret the data in the variable passed - whatever that value may be, with, again, no type checking.

A "declaration" is code that tells the compiler what something is - think of it as a customs declaration for a bottle of booze - you're telling customs that you have a bottle of booze somewhere in your luggage, and what it is. It's not the bottle itself.

A "definition" is code that IS the function or variable. It's the bottle itself.

K&R C has pretty much no declarations. No one knows what anything else is. Try making drinks without knowing in advance what's in every bottle of booze...

Thank you ongoto and thank you achenle for your great explanation

I worked really hard on understanding the code and I nearly got everything.
This is the last part I do not understand:

Code:

if(eq(a, "{")) { /* execute a command for exit code */
    if(fork()) /*parent*/ wait(&ccode);
    else { /*child*/
        doex(1);
        goto err;
    }
    while((a=nxtarg()) && (!eq(a,"}")));
    return(ccode? 0 : 1);
}

As described in the man-page (if page from Section 1 of the unix-6th manual), if we put the command in brackets "if expr { command } ", we can obtain his exit code.

So we fork the current process, and then wait for our child process to finish? But where is our child process continuing his work? After the fork, we will go into the while-loop and and just skip some arguments and then return with ccode? Where was ccode changed? What is ccode?

Could you please explain me this the given code snippet? And elaborate on ccode?

The man page of wait: wait page from Section 2 of the unix-6th manual
The source code: [C] code - Pastebin.com

Thank you very much

Part of an if statement can be to call (fork) an external function; e.g. sed. You could say a == (the results of) sed /something/ for example. ccode is just the name of a variable. Judging by the way it's used here, ccode could be described as a container for:
The child process will be running in it's own "shell" program space and then return it's status (success or failure).
return(ccode? 0 : 1); is a ternary expression. (You can look up ternary). ? indicates a test, 0 is returned if test is true, 1 if false. I think in this case, true would indicate no errors.