patch 9.0.0950: the pattern "\_s\zs" matches at EOL

Problem:    The pattern "\_s\zs" matches at EOL.
Solution:   Make the pattern "\_s\zs" match at the start of the next line.
            (closes #11617)
diff --git a/src/regexp_nfa.c b/src/regexp_nfa.c
index b829a3a..0228cfc 100644
--- a/src/regexp_nfa.c
+++ b/src/regexp_nfa.c
@@ -1742,7 +1742,8 @@
 			    break;
 			}
 		    }
-		    semsg(_(e_nfa_regexp_unknown_operator_percent_chr), no_Magic(c));
+		    semsg(_(e_nfa_regexp_unknown_operator_percent_chr),
+								  no_Magic(c));
 		    return FAIL;
 	    }
 	    break;
@@ -4013,7 +4014,8 @@
 
     for (j = 0; j < sub->in_use; j++)
 	if (REG_MULTI)
-	    fprintf(log_fd, "*** group %d, start: c=%d, l=%d, end: c=%d, l=%d\n",
+	    fprintf(log_fd,
+		    "*** group %d, start: c=%d, l=%d, end: c=%d, l=%d\n",
 		    j,
 		    sub->list.multi[j].start_col,
 		    (int)sub->list.multi[j].start_lnum,
@@ -4790,7 +4792,7 @@
 		{
 		    sub->list.multi[subidx].start_lnum = rex.lnum;
 		    sub->list.multi[subidx].start_col =
-					  (colnr_T)(rex.input - rex.line + off);
+					 (colnr_T)(rex.input - rex.line + off);
 		}
 		sub->list.multi[subidx].end_lnum = -1;
 	    }
@@ -6243,7 +6245,7 @@
 		}
 		else if (!vim_iswordc_buf(curc, rex.reg_buf)
 			   || (rex.input > rex.line
-				&& vim_iswordc_buf(rex.input[-1], rex.reg_buf)))
+			       && vim_iswordc_buf(rex.input[-1], rex.reg_buf)))
 		    result = FALSE;
 		if (result)
 		{
@@ -6373,7 +6375,7 @@
 
 	    case NFA_NEWL:
 		if (curc == NUL && !rex.reg_line_lbr && REG_MULTI
-						 && rex.lnum <= rex.reg_maxline)
+						&& rex.lnum <= rex.reg_maxline)
 		{
 		    go_to_nextline = TRUE;
 		    // Pass -1 for the offset, which means taking the position
@@ -7025,7 +7027,7 @@
 			&& rex.lnum == 0
 			&& clen != 0
 			&& (rex.reg_maxcol == 0
-			    || (colnr_T)(rex.input - rex.line) < rex.reg_maxcol))
+			  || (colnr_T)(rex.input - rex.line) < rex.reg_maxcol))
 		    || (nfa_endp != NULL
 			&& (REG_MULTI
 			    ? (rex.lnum < nfa_endp->se_u.pos.lnum