On Tue, Jul 25, 2017 at 9:56 PM, Johannes Schindelin <[email protected]> wrote: > > As pointed out by Kang-Che Sung and verified by Jody Bruchon, the "%m" > format placeholder is really not in any open standard. > > Also: contrary to Denys' assumption, there are libc versions out there > which did not follow GLIBC's example to implement this non-POSIX > placeholder. That must be the reason why nothing else in BusyBox' source > code relies on %m but uses helpers such as bb_perror_msg() instead. > > It is very easy, and a good readability improvement, too, to introduce > a new helper in ash's source code.
My personal arguments are: 1. If we want to use printf %m, then make sure we support that wherever possible. Even bb_perror_msg and like could be extended to use %m. 2. It's good to provide fallback for libc that doesn't support %m anyway. That's what platform.c & platform.h in busybox are all about. _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
