db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daniel John Debrunner <...@apache.org>
Subject Re: AuthenticationTest failures are mine - I'm looking into it.
Date Fri, 01 Jun 2007 05:43:57 GMT
Daniel John Debrunner wrote:
> I seemed to have caused some errors in AuthenticationTest. I swear it 
> was passing before the commit but then started to fail. :-(

It's actually 543266 that causes the failures. I think that the issue is 
fetching a schema to get a collation when that schema doesn't exist yet. 
No idea what the collation should be at that point!! The authentication 
test hits this because (I think) it prepares a statement with a 
character parameter marker for a procedure call for a user where its 
default schema doesn't exist. Though I'm not sure why the parameter 
needs its collation from the current schema, because the parameter 
marker should take the type of the procedure parameter.



There are few character string types that should take their collation 
type from
compilation schema. I had earlier checked in code for them
to use current schema rather than compilation schema (For reference 
With this commit, I am adding an utility method in ValueNode called 
UsingCompilationSchema(int) which will use the compilation
schema's collation type for it's DTD. And it will use the passed int 
value to se
t its DTD's collation derivation. This utility method will
be used by the subclasses of ValueNode to set their DTD's collation type 
to comp
ilation schema's type wherever required. Note that all the
character string types always take their collation from the compilation 
For instance, persistent character string type column from a
table will take the collation type from the schema their table belongs 
to rather
  than the compilation schema.

I have run the 2 collation related tests, CollationTest and 
CollationTest2 and t
hey run with no problems.

View raw message