Fix mailbox-get-all, and make mailbox timeouts throw a wait-timeout error instead of a string

basis/concurrency/conditions/conditions.factor

@@ -20,10 +20,12 @@ IN: concurrency.conditions
         ]
     ] dip later ;
 
+ERROR: wait-timeout ;
+
 : wait ( queue timeout status -- )
     over [
         [ queue-timeout [ drop ] ] dip suspend
-        [ "Timeout" throw ] [ cancel-alarm ] if
+        [ wait-timeout ] [ cancel-alarm ] if
     ] [
         [ drop '[ _ push-front ] ] dip suspend drop
     ] if ;

basis/concurrency/mailboxes/mailboxes-tests.factor

@@ -1,6 +1,6 @@
 IN: concurrency.mailboxes.tests
-USING: concurrency.mailboxes concurrency.count-downs vectors
-sequences threads tools.test math kernel strings namespaces
+USING: concurrency.mailboxes concurrency.count-downs concurrency.conditions
+vectors sequences threads tools.test math kernel strings namespaces
 continuations calendar destructors ;
 
 { 1 1 } [ [ integer? ] mailbox-get? ] must-infer-as
@@ -75,3 +75,15 @@ continuations calendar destructors ;
 [ ] [ "d" get 5 seconds await-timeout ] unit-test
 
 [ ] [ "m" get dispose ] unit-test
+
+[ { "foo" "bar" } ] [
+    <mailbox>
+    "foo" over mailbox-put
+    "bar" over mailbox-put
+    mailbox-get-all
+] unit-test
+
+[
+    <mailbox> 1 seconds mailbox-get-timeout
+] [ wait-timeout? ] must-fail-with
+    

basis/concurrency/mailboxes/mailboxes.factor

@@ -49,7 +49,7 @@ M: mailbox dispose* threads>> notify-all ;
 
 : mailbox-get-all-timeout ( mailbox timeout -- array )
     block-if-empty
-    [ dup mailbox-empty? ]
+    [ dup mailbox-empty? not ]
     [ dup data>> pop-back ]
     produce nip ;