[ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-250?pag... ] Emmanuel Bernard updated HSEARCH-250: ------------------------------------- Fix Version/s: (was: 3.1.0.Beta2) 3.1.0.CR1 -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-250?pag... ] Hardy Ferentschik commented on HSEARCH-250: ------------------------------------------- * Refactored SharingBufferReaderprovider to cache reader per directory not per direcory provider since some providers have more than one directory (eg FSSlaveDirectoryProvider). The refactoring ment that now not all directories/readers can be initalised in the initialize() mehod. Some readers can only be initilized when the provider switched to a new directory. * Updated tests Changes should be reviewed though. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-250?pag... ] Hardy Ferentschik commented on HSEARCH-250: ------------------------------------------- I am not to worried about the synchronized method in this case. I think it provides the right semantics and the performance should not be impacted since this method will only be called in the case where the directory provider has more than one underlying directory (eg FSSlaveDirectoryProvider) and then it will only be called once per newly 'discovered' directory. I am also not so happy about the currentReaders map. I don't mind that it is mutuable, but it should be private so that modifications to this map can only be made in the synchronized method. AFAICS the only reason to make the map protected is to be able to test the functionality easier. There is really no need to expose the map, but I will talk to Sanne about it and see what he says. I am sure that the problem could also be solved differently, but the current apporach seemed quite straight forward. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-250?pag... ] Emmanuel Bernard commented on HSEARCH-250: ------------------------------------------ We can add a todo in the code but SharedReaderProvider has other bugs already (no lock when checking isCurrent). This is not a priority -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://opensource.atlassian.com/projects/hibernate/browse/HSEARCH-250?pag... ] Hardy Ferentschik resolved HSEARCH-250. --------------------------------------- Resolution: Fixed -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators.... - For more information on JIRA, see: http://www.atlassian.com/software/jira