autotest: re-organize mailbox storage by UID

an effect of 7ce658d is that we can index messages by UID rather than
content (or more specifically, subject). apart from being cleaner, it
allows duplicated subjects.

src/run-tests.pl

@@ -344,7 +344,7 @@ sub readbox($)
 				print STDERR "message '$f' in '$bn' has no identifier.\n";
 				exit 1;
 			}
-			@{ $ms{$num} } = ($uid, $flg.($sz>1000?"*":""));
+			@{ $ms{$uid} } = ($num, $flg.($sz>1000?"*":""));
 		}
 	}
 	return ($mu, %ms);
@@ -360,8 +360,8 @@ sub showbox($)
 
 	my ($mu, %ms) = readbox($bn);
 	my @MS = ($mu);
-	for my $num (sort { $a <=> $b } keys %ms) {
+	for my $uid (sort { $ms{$a}[0] <=> $ms{$b}[0] } keys %ms) {
-		push @MS, $num, $ms{$num}[0], $ms{$num}[1];
+		push @MS, $ms{$uid}[0], $uid, $ms{$uid}[1];
 	}
 	printbox($bn, @MS);
 }
@@ -487,19 +487,19 @@ sub ckbox($$@)
 	}
 	while (@MS) {
 		my ($num, $uid, $flg) = (shift @MS, shift @MS, shift @MS);
-		if (!defined $ms{$num}) {
+		my $m = delete $ms{$uid};
-			print STDERR "No message $bn:$num.\n";
+		if (!defined $m) {
+			print STDERR "No message $bn:$uid.\n";
 			return 1;
 		}
-		if ($ms{$num}[0] ne $uid) {
+		if ($$m[0] ne $num) {
-			print STDERR "UID mismatch for $bn:$num.\n";
+			print STDERR "Subject mismatch for $bn:$uid.\n";
 			return 1;
 		}
-		if ($ms{$num}[1] ne $flg) {
+		if ($$m[1] ne $flg) {
-			print STDERR "Flag mismatch for $bn:$num.\n";
+			print STDERR "Flag mismatch for $bn:$uid.\n";
 			return 1;
 		}
-		delete $ms{$num};
 	}
 	if (%ms) {
 		print STDERR "Excess messages in '$bn': ".join(", ", sort({$a <=> $b } keys(%ms))).".\n";