On Wed, Apr 9, 2008 at 10:46 AM, Karel Zak <kzak@xxxxxxxxxx> wrote: > > Well, it seems we have worked on the same thing: > http://thread.gmane.org/gmane.linux.utilities.util-linux-ng/1371 > > So... see the patch below. It's based on both patches. Well first, sorry about the collision and extra work. I only subscribed after making my patches. My mistake. I like the patch you just posted. I only have two comments: 1. The use of the variable name "fd" for a FILE* is a little confusing to me, as I would normally expect "fd" to stand for "file descriptor" and be a small integer returned by open(2), dup(2) etc. 2. Just putting "\n" in the scan format is not enough to verify that a \n is actually in the input. See this example program: orbital:misc$ cat newline-scan.c #include <stdio.h> int main (int argc, char *argv[]) { int a, b, nf; for (;;) { nf = scanf("%d %d\n", &a, &b); if (EOF == nf) { break; } else if (nf != 2) { fprintf(stderr, "Only converted %d fields\n", nf); return 1; } else { printf("a=%d, b=%d\n", a, b); } } return 0; } orbital:misc$ gcc -Wall newline-scan.c -o newline-scan orbital:misc$ echo 1 2 3 4 | ./newline-scan a=1, b=2 a=3, b=4 The intent of my messing around with the %[ format directive and checking the "newline" variable was to ensure that there were only two fields per line; in other words that the format of the input is correct. However, your version will, I think, accept input files which don't have newlines in the expected place. That could be a useful feature for people who want to be able to edit the timings file for example. Anyway, I think which form of the code one prefers is largely a matter of how much validation one believes is useful here. I'm happy to leave that choice up to you. Thanks, James. -- To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
