this is how printf works

smeezekitty · 09-23-2009, 07:42 PM

i wrote my own verison of printf and wanted to show off
the first param is a function to print a null terminated string

Code:

#define PERCENT_D (10)
#define PERCENT_X (16)

void printf(void (*out)(char *), const char *format, ...){ //void instead of int
int v;
char u[10],t=0;
va_list va;
va_start(va, format);
while(format[t]){
if(format[t]!='%'){u[0]=format[t]; u[1]=0; out(u);} else {
t++;
switch(tolower(format[t])){
case 'u':
case 'd':
v=va_arg(va, int);
itoa(v, u, PERCENT_D);
out(u);
break;
case 's':
out(va_arg(va, char *));
break;
case 'f':
v=va_arg(va, float);
itoa(v, u, PERCENT_D);
out(u);
break;
case 'x':
case 'X':
v=va_arg(va, int);
itoa(v, u, PERCENT_X);
out(u);
break;
default:out("%");
}
}
t++;
}
va_end(va);
}

carbonfiber · 09-24-2009, 12:38 PM

Horrible.

P.S.: Constructivecriticism is my middle name.

smeezekitty · 09-24-2009, 12:47 PM

Quote:

Originally Posted by carbonfiber

Horrible.

P.S.: Constructivecriticism is my middle name.

that isnt consturctive criticism
it would be if you say what is horrable

Sergei Steshenko · 09-24-2009, 01:16 PM

Quote:

Originally Posted by smeezekitty

i wrote my own verison of printf and wanted to show off
the first param is a function to print a null terminated string

Code:

void printf(void (*out)(char *), const char *format, ...){ //void instead of int
int v;
char u[10],t=0;
va_list va;
va_start(va, format);
while(format[t]){
if(format[t]!='%'){u[0]=format[t]; u[1]=0; out(u);} else {
t++;
switch(tolower(format[t])){
case 'u':
case 'd':
v=va_arg(va, int);
itoa(v, u, 9);
out(u);
break;
case 's':
out(va_arg(va, char *));
break;
case 'f':
v=va_arg(va, float);
itoa(v, u, 10);
out(u);
break;
case 'x':
case 'X':
v=va_arg(va, int);
itoa(v, u, 16);
out(u);
break;
default:out("%");
}
}
t++;
}
va_end(va);
}

Why did you write your own 'printf' and why should others consider the code ?

...

Whatever you write, if you have constants like 9, 10, 16, etc. instead of meaningful macro names, your code is bad.

The allowed constants are 0 (like you are starting from the number zero element in an array) and +1, -1 - like 'idx = idx + 1;' - meaning next index/value in an array. And because of ++, -- in "C" one rarely needs +1, -1.

Sometimes 2, 3 - like each second or each third element, but 3 is already suspicious.

Hko · 09-24-2009, 01:29 PM

Quote:

Originally Posted by Sergei Steshenko

Whatever you write, if you have constants like 9, 10, 16, etc. instead of meaningful macro names, your code is bad.

Here it is used as the radix of the conversion, one of the few cases I wouldn't consider bad, let alone "horrible"...

Sergei Steshenko · 09-24-2009, 05:23 PM

Quote:

Originally Posted by Hko

Here it is used as the radix of the conversion, one of the few cases I wouldn't consider bad, let alone "horrible"...

Still,

Code:

#define D_RADIX ((unsigned)10) /* for %d conversions */
#define X_RADIX ((unsigned)16) /* for %x conversions */

, i.e. relationship between symbolic ('D', 'X') and numeric (10, 16) radix. And the relationship is reflected in a meaningful macro name.