diff options
Diffstat (limited to 'sink.c')
| -rw-r--r-- | sink.c | 13 |
1 files changed, 11 insertions, 2 deletions
@@ -720,8 +720,14 @@ int open_sink(struct query *ctl, struct msgblk *msg, send_bouncemail(msg, "Some addresses were rejected by the MDA fetchmail forwards to.\r\n", *bad_addresses, from_responses); - /* local notification only if bouncemail was insufficient */ - if (!(*good_addresses) && total_addresses > *bad_addresses) + /* + * It's tempting to do local notification only if bouncemail was + * insufficient -- that is, to add && total_addresses > *bad_addresses + * to the test here. The problem with this theory is that it would + * make initial diagnosis of a broken multidrop configuration very + * hard -- most single-recipient messages would just invisibly bounce. + */ + if (!(*good_addresses)) { #ifdef HAVE_SNPRINTF snprintf(addr, sizeof(addr)-1, "%s@%s", run.postmaster, ctl->destaddr); @@ -735,6 +741,9 @@ int open_sink(struct query *ctl, struct msgblk *msg, SMTP_rset(ctl->smtp_socket); /* required by RFC1870 */ return(PS_SMTP); } + + if (outlevel >= O_VERBOSE) + error(0, 0, _("no address matches; forwarding to %s."), run.postmaster); } /* |