Hi,
I have a question about the algorithm to search the metadata for a named
query as defined in section 14.5 "Query Factory in PersistenceManager
interface" (see description of method newNamedQuery).
(1) The spec says that if the named query is not found in already-loaded
metadata, it examines metadata files until the query is found. I read
this that JDO looks into .jdo files even in case the class metadata was
loaded from a different .jdo file earlier in the search order. Suppose
you have a package.jdo including metadata of class Bar and this does not
define any named query. Then there is a class .jdo file Bar.jdo defining
a named query. Chapter 18 defines that the metadata from the package.jdo
file will be used instead of the one from the class .jdo file.
However, the algorithm for searching named queries reads as if it should
find the named query in the class jdo file. I think this is confusing,
especially because the named query might be written with respect to the
metadata of the class .jdo file and this might conflict with the metdata
that is loaded from the other file. I propose that once the metadata for
a class is loaded no other .jdo file should be examined for searching
named queries.
(2) The spec says the search order for named queries includes files
named based the query name. But the example as given in the spec is
based on the class name: it includes the file Bar.jdoquery and Bar is
the class name, not the query name. I think using the class name makes
more sense and propose to keep the example, but change the description.
What do you think?
Regards Michael
--
Michael Bouschen [EMAIL PROTECTED] Engineering GmbH
mailto:[EMAIL PROTECTED] http://www.tech.spree.de/
Tel.:++49/30/235 520-33 Buelowstr. 66
Fax.:++49/30/2175 2012 D-10783 Berlin
