[jboss-user] [JBoss Tools (users)] - Re: Hibernate reverse engineering errors

max.andersen@jboss.com do-not-reply at jboss.com
Sat Mar 7 04:13:34 EST 2009


"bdlink" wrote : OK, I edited the reveng.xml file by hand, eliminating all of the empty string attribures, and that eliminates the exception (cheers!)
  | 

So two bugs found: the editor is leaving empty quotes (bad style) and the reveng.xml parser chokes on empty quotes (which it shouldn't).
Did you say you saw jira's for this already ? (need to create them if they aren't there)

anonymous wrote : 
  | I looked at my DB to see why extra associations were showing up and found duplicate foreign key constraints that mysql added at some time, and when I removed these, I get the associations I would expect.  
  | 

Good - i'm not insane then.

anonymous wrote : 
  | So things are much improved!  There are some remaining issues I am looking at including :
  | 1. Even though I annotate a foreign key with one-to-one, the property is marked with @ManyToOne and the inverse side uses a Set (when it is not excluded).  
  | 

hmmm - the one-to-one detection is rather new; looks like it requires the underlying db to actually have a one-to-one foreign key (i.e. it only listens to reveng.xml about one-to-one attributes not its definition plus the one-to-one between pk and a non pk is not "allowed").

Needs a bug report. Unfortunately I can't fix this for GA.

anonymous wrote : 
  | 2. I have one case where there are two associations between Student and Review (both one Student to many Reviews).  In the ReviewHome class it injects two studentHome components (duplicate names), where one of them should be a different @Role.  (I think this is in JIRA from long ago, but I will check later)
  | 

This is in context of Seam gen, rigth ? This is something to report against seam-gen in Seam - nothing hibernatetools specific about this afaik. If you got a jira ref let me know ;)

anonymous wrote : 
  | The one-to-one issue is obviously more important.
  | 

I can't do it for GA but will look into it straight after.
Until then you can try and implement the oneToone related methods in reverse engineering strategy to think there is a one-to-one.



View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215923#4215923

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215923



More information about the jboss-user mailing list