Stephane wrote:

Now do you have a fix?

1) Tried to reproduce it without success. My fault, looks like I stopped in the middle of the startup process since the font update took to long and the startup logic in Clipboard (who resets the class variable) is processed after that. So it was still pointing to the mac clipboard in this case - so it's not a problem of the one click image or Pharo by default. But it would be good if the Clipboard startup is processed much earlier (before such annoying "long runners"). The font update at image startup is now disabled by default in newer Pharo images but can be enabled again. So I've added the Clipboard to the end of the first processed startup elements so it gets called earlier. The change is in "System-Support-torsten_bergmann.33.mcz" (based on 32) one has to call "SystemDictionary initialize" after that. 2) Regarding the Mac host windows menu. First question is if it should really be part of the standard image or a loadable package for Mac? Whats the status of the windows port, if any? However - the test is currently red on Windows: a) Yes - there is code in each test method to return if not on mac, but it will already crash in #setUp on a non Mac platform like Windows b) If this check is also added to #setUp to return early it will still be red since a few more test methods like #testMenuHandles are not guarded c) testCharacterChanging sends the unknown method #calculateShouldBeFrom:using: Maybe we should move HostSystemMenusTest>>isMacintosh to the class side and implement HostSystemMenusTest class>>isAbstract like this: isAbstract ^self isMacintosh not This makes the unit test an abstract one on non-mac platforms (and therefore green ;), we can remove all the checks in the test methods and someone with a mac (John?) could fix the test code on the intended platform... Bye Torsten -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01