Re: [jira] Commented: (DERBY-18) Exposed name matching has bugs when the column name is qualified with a schema name.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Shreyas Kaushik wrote:

> Dan,
>
> I did the patch for this.
>
> From your comments I looked at the getSchemaDescriptor() in
> QueryTreeNode. The schema name was getting
> stored there but not stored in the TableName in the corresponding node.
>
> Since *null* schema implies current schema I just called bind() in
> TableName.java whereever the schema was null.

I'm not sure this is correct, because if you remember you patch to
Derby-13, a TableName with a null schema can represent a correlation
name (which don't have schemas). So the fact a TableName's schema name
is null does not always mean that its schema name maps to the current
schema. Thus only when a TableName represents a implicitly schema
qualifed name (table, view, procedure, function, trigger, index,
constraint name) should its lack of schema name be converted to the
current schema. Possibly the only case that doesn't have an implicit
schema name is the correlation name.

Thus the patch needs to be careful of this overloaded use of TableName.
Does your patch handle this correctly? Without looking at the code in
detail I'm not sure how to handle this correctly, if you haven't made
any progress I can try give and help out tomorrow.

Dan.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFB+c6jIv0S4qsbfuQRAvmcAJ9Nu4H0HzhGxg+S8nOJqBxi1gbbAQCgtWjO
dLSVBATTSeJjkHrin45ygk4=
=3KaU
-----END PGP SIGNATURE-----