Thread: Which style of if loops is most common?

i use k&r but generally add spaces in brackets, like:

Code:

if ( a ) {
(TAB STROKE)i1;
(TAB STROKE)i2;
(TAB STROKE)} else {
(DOUBLE TAB STROKE)i3;
(DOUBLE TAB STROKE)i4;
(DOUBLE TAB STROKE)}
(TAB STROKE)}

i ussually use

Code:

if ( a ) i1 else i2;

when using a single instruction. i edit my programs in notepad and prefer to see more of a "condensed" program than less of an "airy" one.

this is an example:

Code:

/*
==============================================================================
===                                                                        ===
=== MNR source file - version 0.1 build 046 date 18.08.2005 test-c alpha-3 ===
===                        DOS/Windows and UNIX/Linux                      ===
===                                                                        ===
==============================================================================
*/

#include "mnr.h"

//prototypes
char process_com(void);
char test(void);
void process(void);
char cleanup(void);

//global
char *fis, *fos, *com, intr, dash, ign;
FILE *fi, *fo, *dat;

int main(int argc, char *argv[]) {
	printf("\n\n\n===============================================================================\n");
	printf("===              Welcome to the MNR Monstruous Numbers Reader               ===\n");
	printf("===              version 0.1      copyright (C) 2005 moonlord               ===\n");
	printf("=========== FOR TESTING AND EVALUATION ONLY - NOT FOR COMERCIAL USE ===========\n\n");
	switch (argc) {
		case 1: {
			printf("=========== SYNTAX: \"MNR interpretation[switches] file_in file_out\" ===========\n");
			printf("=== interpretation - \"a\" or \"b\" or \"c\" - according to whether the output is ===\n");
			printf("===                  written in American, Britanic or Continental format    ===\n");
			printf("===                  as described in the readme file.                       ===\n");
			printf("=== switches - \"-\" to use dashes between word parts.                        ===\n");
			printf("===          - \"i\" to ignore non-digit characters from file_in; an error is ===\n");
			printf("===            generated otherwise.                                         ===\n");
			printf("===   * switches are case-sensitive.                                        ===\n");
			printf("=== file_in - the number is read from this file.                            ===\n");
			printf("=== file_out - the text is written in this file.                            ===\n");
			printf("===============================================================================\n\n");
			return(0);
			}
		case 2: if (atol(argv[1]) == 443556) {printf(VERSTR); return(0);}
				else {printf("ERR01: Syntax error.\n"); return(1);}
		case 4: {
			com=argv[1]; fis=argv[2]; fos=argv[3];
			if(process_com()) {printf("ERR01: Syntax error.\n"); return(1);}
			switch(test()) {
				case 4: {printf("\nERR04: Error opening data file names.dat.\n"); return(4);}
				case 3: {printf("\nERR03: Error opening output file %s.\n", argv[3]); return(3);}
				case 2: {printf("\nERR02: Error opening input file %s.\n", argv[2]); return(2);}
				case 1: {printf("\nAborted by user.\n"); return(0);}
				}
			if(count_file()) {printf("\nERR08: Non-digit character encountered.\n"); return(8);}
	/**/		reopen_files();
			process();
			switch(cleanup()) {
				case 3: {printf("\nERR07: Error closing data file names.dat.\n"); return(7);}
				case 2: {printf("\nERR06: Error closing output file %s.\n", argv[3]); return(6);}
				case 1: {printf("\nERR05: Error closing input file %s.\n", argv[2]); return(5);}
				}
			printf("Done.\n");
			return(0);
			}
		default: {
			 printf("ERR01: Syntax error.\n");
			 return(1);
			 }
	}
}


//////////
//
// process_com(): Interpretes command.
// Returns: 0 - success;
//          1 - some syntax error;
//

char process_com(void) {
	char *i=com;
	switch(com[0]) {
		case 'a': case 'b': case 'c': {intr=com[0];break;}
		default: return(1);
		}
	while (*++i) {
		switch(*i) {
			case '-': {dash=1;break;}
			case 'i': {ign=1;break;}
			default: return(1);
			}
		}
	return(0);
}


//////////
//
// test(): Tests and opens files (fi,fo,"names.dat").
// Returns: 0 - success;
//          1 - aborted by user;
//          2 - error opening fi (input file);
//          3 - error opening fo (output file);
//          4 - error opening "names.dat";
//

char test(void) {
	struct ffblk fot;
	char conf;
	if (!(findfirst(fos, &fot, 0) == -1)) {
		do {
			printf("%s already exists. Overwrite? (Y/N)", fos);
			conf=toupper(getche());
			printf("\n");
		} while (conf != 'Y' && conf != 'N');
		if (conf == 'N') return(1);
	}
	if ((fi=fopen(fis, "rt")) == NULL) return(2);
	if ((fo=fopen(fos, "wt")) == NULL) return(3);
	if ((dat=fopen("names.dat", "rt")) == NULL) return(4);
	return(0);
}


//////////
//
// process(): Reads from fi, writes in fo.
// Returns: nothing.
//

void process(void) {
	;
}


//////////
//
// cleanup(): Writes buffers and closes files.
// Returns: 0 - success;
//          1 - error closing fi;
//          2 - error closing fo;
//          3 - error closing "names.dat";
//

char cleanup(void) {
	if (fclose(fi) == EOF) return(1);
	if (fclose(fo) == EOF) return(2);
	if (fclose(dat) == EOF) return(3);
	return(0);
	}

it actually does nothing but is in progress. Will announce when it's ready if you're interested...

I like to use spaces before and after brackets:

Code:

void foo (void);
if (a > b) {

} else {

}

Just thought I'd add that from my previous post

i use Allman.

moonlord, does this compile? i think the "t" mode in fopen is not standard:

Code:

	if ((fi=fopen(fis, "rt")) == NULL) return(2);
	if ((fo=fopen(fos, "wt")) == NULL) return(3);
	if ((dat=fopen("names.dat", "rt")) == NULL) return(4);

In fact this is pretty much the style I use:
http://pantransit.reptiles.org/prog/CodingStyle.html

Hmm... The dude that made those coding style guidelines makes some rather dubious claims... though his style is pretty nice. It is pretty close to what I use when writing C, but my C++ code is always Allman unless I am writing a very small demo program for someone... then I take liberties.

>>Also most IDE's auto indent now too, so you'd have to backspace the tab and space twice again.<<
Just change the IDE's config to use spaces instead of tabs upon invoking the auto-indent (assuming its available).


I look after code that is quite old (been running for 10 years+), and many people have changed it over this period of time. Some have used tabs (a mixture of one tab, two tabs, even four tabs for indentation), and some have used spaces (again, varying numbers). It's a mess, and is really aweful to look at sometimes.
My advice: Whatever your own personal preference, if you're amending someone else's code, use their style of indentation and bracketing in preference to your own. That way, future generations of programmers will at least see a common approach throughout.

Originally Posted by Perspective

you use spaces to group your logic, but spaces have no effect on syntax, brackets do. If you group your logic using brackets, you know it will execute the way you've intended it to. If you group your logic with spaces, it will execute based on order of operations despite what the visual cues of your spaces suggest.

Obviously, if parentheses are needed, then in the spirit of not-being-a-moron, they are written. However, in the case of

Code:

(x == y  ||  z == w)

, they are not needed.

Originally Posted by Perspective

ex:
if (a && b || c)

is that if ((a && b) || c) or if (a && (b || c)).

I see nothing wrong with

Code:

((a  &&  b)  ||  c)

.

But this would be if a and b and c are expressions. If a and b are simply integral values, then

Code:

((a && b)  ||  c)

might be used.