svn commit: r368197 - head/sbin/bectl

[Gleb Smirnoff]

  Kyle,

On Wed, Dec 02, 2020 at 12:08:43PM -0600, Kyle Evans wrote:
K> > K> This should be gated on !cmd->silent, because some paths have
K> > K> consumers that are specifically designed to not have to deal with
K> > K> redirecting stderr. It was quite intentional that this didn't
K> > K> previously print anything.
K> >
K> > AFAIK, the only command that has cmd->silent is "check".
K> >
K> > I can't agree that it should suppress stderr in case of libbe_init()
K> > failure. Failure of the library is something different to failed
K> > check of current system. It is permanent failure, meaning that command
K> > is being with incorrect root argument or ZFS is missing at all. Pretty
K> > much the same as using bectl with incorrect arguments or options.
K> >
K> 
K> The sole purpose of `bectl check` is to weed out if bectl will work
K> (i.e. if libbe_init succeeds) so that scripts like freebsd-update can
K> determine if they're running on a system or with a root (`bectl -r`)
K> that can work with boot environments. It is not meant to give
K> diagnostics like that upon failure, its sole purpose is to know if
K> your script should proceed with doing bectl-y things safely or if it
K> will just be fraught with peril.

Understood. Sorry for my mistake. Is this patch a correct one?

Index: bectl.c
===================================================================
--- bectl.c     (revision 368197)
+++ bectl.c     (working copy)
@@ -585,8 +585,9 @@ main(int argc, char *argv[])
        }
 
        if ((be = libbe_init(root)) == NULL) {
-               fprintf(stderr, "libbe_init(\"%s\") failed.\n",
-                   root != NULL ? root : "");
+               if (!cmd->silent)
+                       fprintf(stderr, "libbe_init(\"%s\") failed.\n",
+                           root != NULL ? root : "");
                return (-1);
        }

-- 
Gleb Smirnoff