[FFmpeg-devel] [PATCH] remove unused and broken test program in libavutil/base64.c

Bobby Bingham uhmmmm
Thu Jan 29 20:59:28 CET 2009


On Thu, 29 Jan 2009 01:20:02 -0800
Baptiste Coudurier <baptiste.coudurier at gmail.com> wrote:

> Hi,
> 
> Marc Mason wrote:
> > Michael Niedermayer wrote:
> > 
> >> Stefano Sabatini wrote:
> >>
> >>>  if (strcmp(encoded, t->encoded) != 0) {
> >> superflous != 0
> > 
> > Redundant (or superfluous) information sometimes make the code or
> > the coder's intention clearer.
> 
> Could you please give an example ?
> I'm really used to read "!strcmp" to match, and when I don't see the
> '!', I know this means "don't match"
> 
> In this case, I also think != 0 is redundant and useless.
> 
> >>>     printf("\n");
> >>>     printf("Encoding/decoding tests on random data\n");
> >> can be merged
> > 
> > For what gain? Cosmetic or computational?
> 
> I also find the merge better.
> 
> > GCC is likely to merge the two statements.
> 
> Does it, or not ?
> 

Doesn't here (gcc 4.3.3, amd64):

printf.c:

#include <stdio.h>

int main()
{
    printf("\n");
    printf("Hello world!\n");
    return 0;
}

relevant output from gcc -O3 -S printf.c:

.LC0:
	.string	"Hello world!"
	.text
	.p2align 4,,15
.globl main
	.type	main, @function
main:
.LFB24:
	subq	$8, %rsp
.LCFI0:
	movl	$10, %edi
	call	putchar
	movl	$.LC0, %edi
	call	puts
	xorl	%eax, %eax
	addq	$8, %rsp
	ret

-- 
Bobby Bingham
Never trust atoms.  Or anything made of atoms.
??????????????????????




More information about the ffmpeg-devel mailing list