[23:08]<bslivka> Woo hello [23:28]<zumbrunn> hi bslivka [23:28]<bslivka> Great day [23:29]<bslivka> I'm thinking I should probably post my little TestBug app to helma's bugzilla [23:29]<bslivka> the behavior exibited just doesn't seem right. [23:30]<zumbrunn> I didn't try out the TestBug.zip yet [23:30]<zumbrunn> I'll try tomorrow morning [23:30]<zumbrunn> (in a few hours) [23:42]<zumbrunn> or nope... I'm trying right now... [23:44]<zumbrunn> I see that main fails after the app is restarted [23:44]<zumbrunn> but the object seems to be persisted [23:44]<zumbrunn> calling add seems to add two nodes each time [23:54]<bslivka> Yep [23:55]<bslivka> if you look at the code, you'll see that I hava a mountpoint [23:55]<bslivka> then another mountpoint on that mountpoint [23:55]<bslivka> then a collection on that outer mountpoint [23:56]<bslivka> add.hac adds an object to the outer collection [23:56]<bslivka> which causes that object, and its parent, the outer mountpoint to be persisted [23:56]<bslivka> but not the mountpoint's parent [23:56]<bslivka> so the chain is broken [23:57]<bslivka> in memory helma still knows what's connected to what, but since the inner mountpoint isn't persisted properly, it doesn't rebuild the graph correctly [23:57]<bslivka> it starts from scratch [23:57]<bslivka> that is, when you restart helma [23:58]<bslivka> the workaround is to, one time, say root.homepage=new HomePage() [23:58]<bslivka> which forces it to be persisted, and given a numbered id [23:59]<bslivka> One thing I don't understand though is that unless you do that, the HomePage mountpoint's _id is "homepage" [23:59]<bslivka> not a number