[jboss-jira] [JBoss JIRA] (AS7-1893) JDBC cache store configuration improvements

[ https://issues.jboss.org/browse/AS7-1893?page=com.atlassian.jira.plugin.s... ] Richard Achmatowicz edited comment on AS7-1893 at 2/26/12 7:48 PM: ------------------------------------------------------------------- Manik, Paul Tristan set up the JDBC store configuration in such a way that: - if you define an entry table only, you get a string based store - if you define a bucket table only, you get a binary based store - if you define both types of tables, you get a mixed store I assume you would rather see something which makes the creation of the different types of stores more explicit, such as: {noformat} <!-- define a string-based store --> <string-keyed-jdbc-store fetch-state= passivation= preload= purge= shared= singleton= > <string-keyed-table ...> ... </string-keyed-table> <string-keyed-jdbc-store> <!-- define a binary-based store --> <binary-keyed-jdbc-store fetch-state= passivation= preload= purge= shared= singleton= > <binary-keyed-table ...> ... </binary-keyed-table> <binary-keyed-jdbc-store> <!-- define a mixed store --> <mixed-keyed-jdbc-store fetch-state= passivation= preload= purge= shared= singleton= > <string-keyed-table ...> ... </string-keyed-table> <binary-keyed-table ...> ... </binary-keyed-table> <mixed-keyed-jdbc-store> {noformat} As you point out, in the current implementation, one can specify specific details for the underlying database tables, such as column names and column types, and table name prefixes. When you say that the current implementation exposes implementation details, do you mean that you simply want to change the names so that we aren't tied to a particular implementation, but keep the underlying implementation the same (i.e. get rid of the word bucket-table and replace it with binary-keyed-table, but use a bucket-based cache store anyway). Also, is the exposure to TableManipulation attributes ok as is? In other words, retain the ability to set the column names and types for each column? was (Author: rachmato): Manik, Paul Tristan set up the JDBC store configuration in such a way that: - if you define an entry table only, you get a string based store - if you define a bucket table only, you get a binary based store - if you define both types of tables, you get a mixed store I assume you would rather see something which makes the creation of the different types of stores more explicit, such as: {noformat} <!-- define a string-based store --> <string-keyed-jdbc-store fetch-state= passivation= preload= purge= shared= singleton= > <string-keyed-table ...> ... </string-keyed-table> <string-keyed-jdbc-store> <!-- define a binary-based store --> <binary-keyed-jdbc-store fetch-state= passivation= preload= purge= shared= singleton= > <binary-keyed-table ...> ... </binary-keyed-table> <binary-keyed-jdbc-store> <!-- define a mixed store --> <mixed-keyed-jdbc-store fetch-state= passivation= preload= purge= shared= singleton= > <string-keyed-table ...> ... </string-keyed-table> <binary-keyed-table ...> ... </binary-keyed-table> <mixed-keyed-jdbc-store> {noformat} As you point out, in the current implementation, one can specify specific details for the underlying database tables, such as column names and column types, and table name prefixes. When you say that the current implementation exposes implementation details, do you mean that you simply want to change the names so that we aren't tied to a particular implementation, but keep the underlying implementation the same (i.e. get rid of the word bucket-table and replace it with binary-keyed-table, but use a bucket-based cache store anyway). Also, is the exposure to TableManipulation attributes ok as is? -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.jboss.org/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira