No such method error refreshing StatefulManyToMany about squeryl

[max-l]

NoSuchMethodError are very often due to incomplete/stale compilation, or wrongfull
use of asInstanceOf, I don't think there asInstanceOfs in that area, can you validate the 1st hypothesis ?

[alecw]

Hi Max,

I did clean build of both Squeryl and my project. Same result. It's puzzling.

-Alec

[max-l]

Wait a minute.... the $eq method is used by the HashMap, and that is what the complain is about.... KeyedEntity implements the required, equals+ hashCode stuff
but if you are using an instanceOf like :

ZamboniSchema.workflowToSteps.leftStateful(this.asInstanceOf[Workflow])

the equality and hashCode might not be implemented....

[alecw]

Hi Max,

Note that the problem is not the $eq method. It is the $plus$plus$eq method. I tried without the asInstanceOf and still had the problem. I tried reproducing the problem with SchoolDb2 courseAssignments but I was unable to. The thing that is strange to me about the exception is the the L at the beginning of the class name in the function signature. I wonder if this is caused by the fact that this many-to-many relationship relates a table to itself.

I'm afraid I can't spend any more time on this right now. I'm just going to not use the stateful flavor for now. At some point I guess you or I could try to create a test case with a table referring to itself.

Thanks, Alec

[max-l]

Djeez, I didn't realize it was the the ++= ($plus$plus$eq)...

if you get a chance, Line 189 in :
http://github.com/max-l/Squeryl/blob/master/src/main/scala/org/squeryl/dsl/ManyToMany.scala

replace :
_map ++= relation.associationMap

by :
for(e <- relation.associationMap)
_map.put(e._1, e._2)

The ++= relies on some implicit conversion magic, the for loop has much lower content of type system magic....

[alecw]

Hi Max,

Map having an add method? Who would think of such a crazy thing? ;-)

This change fixes the problem:

    for(e <- relation.associationMap)
      _map += e

Thanks, Alec

[alecw]

So much for using markup. Here it is again:

for(e <- relation.associationMap)
  _map += e

[max-l]

djeeez... I meant put, I corrected the comment, fortunately
github's issue tracking allows editing the comments !

[max-l]

replaced +== by put in master.