On Mon, Oct 31, 2011 at 4:10 PM, Alexander Shishkin <[email protected]> wrote: > On Mon, 31 Oct 2011 13:21:01 +0100, Denys Vlasenko <[email protected]> > wrote: >> On Mon, Oct 31, 2011 at 12:25 PM, Denys Vlasenko >> <[email protected]> wrote: >> > On Mon, Oct 31, 2011 at 1:15 AM, Alexander Shishkin <[email protected]> >> > wrote: >> >> Sometimes there's a need to figure out the controlling tty from a shell >> >> script, for example, to obtain a line for getty. >> >> In this case it's easier >> >> to call cttyhack than trying to repeat some of the cttyhack's logic. >> > >> > "cttyhack getty 0 -" should work, right? >> > >> > >> >> + /* If it's a /dev/tty, still make sure it's wired to >> >> std{in,out,err} */ >> >> + fd = open_or_warn(console, O_RDWR); >> >> + if (fd < 0) >> >> + goto ret; >> >> + //bb_error_msg("switching to '%s'", console); >> >> + dup2(fd, STDIN_FILENO); >> >> + dup2(fd, STDOUT_FILENO); >> >> + xmove_fd(fd, STDERR_FILENO); >> > >> > Think what would happen if fd = 1 >> >> Fixed it and applied to git. Thanks. > > Thanks, except for now there is a problem with the serial test going > before the sysfs test, and the former is almost never accurate as very > few serial drivers use "ttyS" names these days. I'd say that the sysfs > test is the most relevant for the cttyhack purposes.
Then we need to stop lying that cttyhack "analyzes stdin with various ioctls"... I restored the order of checks. -- vda _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
