use log_warn to report the failure of canonpath

reset path buffer before writing to it otherwise path, which is filled by the garbage that's on the stack, may not end with a NUL byte. main_enqueue fails with "malformed data" if the path sent doesn't end with a NUL. (i.e. the fact that in every case the path is NUL-terminated is not relevant. For semplicity, I'm working with paths long PATH_MAX bytes and only checking that path[PATH_MAX-1] is NUL when handling the imsgs.) Issue reported by Dirk-Wilhelm Peters, thanks!

use a strictier pledge for `amused add' now that we're using canonpath instead of realpath(3), we can drop the rpath pledge promise.

use canonpath instead of realpath(3) so we get the same behaviour with `add' and `load'

use canonpath (from kern_pledge.c) instead of home grow "equivalent"

drop a local variable don't need to keep `files' around, res->files is enough. it would have prevented the previous bug (probably.)

fix `amused add' failure introduced in previous commit don't advance the files pointer, we need the original pointer when receiving the replies.

refactor: group imsg handling the approach with one-function-per-imsg leads to too much code splatted across the file.

amused monitor: allow to pass a list of event as filter it's easier / simpler for scripts to do $ amused monitor next,prev,jump rather than $ amused monitor | egrep --line-buffered 'next|prev|jump'

add back missing memzero

use a more readable strncmp to look for prefixes also while here stick an `else' before the second clausole. We're supposed to trim out only '> ' and ' ', but the missing else would also trim out ' ' after a '> '!

get rid of realpath during load it has a non-ignorable cost over NFS (sdk@ reported ~30 seconds to load 64k songs, and up to 5 minutes over wifi!) and don't provide us any real gain: files can still vanish after being imported or may appear later. The only advantage of realpath was that it would clean up the path from segments like "/./" and resolve the ".." components, but that's a minor issue anyway. prodded by, discussed with and tested by sdk@, thanks a lot!

drop now unused #include <sndio.h> with the recent refactoring, sndio is only visible in player.c

fix argc check argc can be -1 if we called noarg with argc=0, because optind is 1.

correctly handle arguments for sub-commands things like "--" should be skipped etc, easier to rely on getopt(3).