Otto Moerbeek wrote:
> This angel has problem with KNF:

Ow yes: int\nmain(...


> space between function name and arg list. Also, return (expr);
> is preferred to return expr;

It strikes me as inconsistent. I'd expect either ``return 0;'' or
``return(0)'' but not ``return (0)''. Any pointer to the reason
behind this? I can't find it in style(9).


> Apart from that, formatting changes combined with functional
> changes are frowned upon.

Errr OK, I'll keep it in mind.


> But the main point: you did not make clear why stderr would be
> better.

Because it's an errormessage. Errormessages should be sent to
stderr. I tried login in to a disabled account to see if the
errormessage was not echoed but that was also not the case. That
why I was informing if there were any other objections to sending
the message to stderr. It was not a rhetorical question, nor was
my patch a suggestion nor a recommendation. It was merely meant to
clarify what I had in mind.

But now after taking your remarks in account, this is what the
code should look like in the end. After applying one patch for the
functional change and one for the format. (without the license for
briefness):

[snip: licensenote]
#include <sys/types.h>
#include <fcntl.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>

/* Distinctly different from _PATH_NOLOGIN. */
#define _PATH_NOLOGIN_TXT "/etc/nologin.txt"

#define DEFAULT_MESG "This account is currently not available.\n"

/*ARGSUSED*/
int
main(int argc, char *argv[])
{
int nfd;
ssize_t nrd;
char nbuf[BUFSIZ];

nfd = open(_PATH_NOLOGIN_TXT, O_RDONLY);
if (nfd == -1) {
write(STDERR_FILENO, DEFAULT_MESG, strlen(DEFAULT_MESG));
return (1);
}

while ((nrd = read(nfd, nbuf, sizeof(nbuf))) != -1 && nrd != 0)
write(STDERR_FILENO, nbuf, nrd);

close(nfd);
return (1);
}

What say you?



# Han