[ http://opensource.atlassian.com/projects/hibernate/browse/OGM-12?page=com... ] Emmanuel Bernard reassigned OGM-12: ----------------------------------- Assignee: Nicolas Helleringer -- 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/OGM-12?page=com... ] Nicolas Helleringer commented on OGM-12: ---------------------------------------- Implemented String Mapped storage for BigDecimal, BigInteger, Byte, Boolean, URL To be counter checked https://github.com/nicolashelleringer/hibernate-ogm/branches/OGM-12 -- 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/OGM-12?page=com... ] Nicolas Helleringer commented on OGM-12: ---------------------------------------- Maybe allowing the user to choose how he want to store is the path to go (with underlying storage capability description juste like SQL dialect in HHH) For the types that do have a 'standard' string encoding provided by Java that garanties the two way from and to string coherence, I think we should stay on that. But, therefor the question is, if we do want to e.g. store something from Java and then (the data beeing replicated to another computer / technology) be read in .NET (it could be !), then the question is : - Is the way that Java does stringify the type a standard outside the Java world or is it limited to it ? I shall have a look on that on the types I have allready done (.NET and C++ focus I think). -- 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/OGM-12?page=com... ] Emmanuel Bernard commented on OGM-12: ------------------------------------- Hey A few remark on your use of Git / GitHub. You've merge my work from master into your topic branch. This is something we don't do as it creates non linear history and makes it ahrder to read in the end. http://in.relation.to/Bloggers/HibernateMovesToGitGitTipsAndTricks When you work on a topic branch, always rebase this work (git rebase master) before pushing it. Note that while you've published the branch to github, it's ok to use rebase as a topic branch is really a work in progress. To work around your problem, you can create a new branch and cherry pick the two core commits of your work and make sure you rebase in the future. Wrt GitHub, when you want to share some work, go to your topic branch on github (after pushing it) and you should see a Pull Request button on the top right. This is the preferred way to share as it lists the relevent commits and offer a nice diff option as well as a way to discuss the code. -- 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/OGM-12?page=com... ] Emmanuel Bernard commented on OGM-12: ------------------------------------- So you've raised interesting questions. Here is my take on them. Today we end up serializing "primitive" Java objects. Of course that's not ideal to reach platform independence as these are still actual Java objects. Some work is being done on the Infinispan side to make Hot Rod, the communication protocol, somewhat data gnostic by reusing things like Google Protocol Buffer or equivalent. It might make sense to exploit this work at some point. In an ideal world we would push data into infinispan as platform independent values I'm not sure that for OGM v1 we want to *properly* address this issue. So it's fine by me if we limit the Java types we end up serializing into the grid to help the future migration if needed. In my opinion, the following are ok: - core numeric types (even if that can be subject to discussion) - String - Boolean - Byte - byte[] Any other you can think of? As for BigInteger / BigDecimal, if we can find a platfgorm agnostic representation in String or something else, that would be interesting. Worse case scenario, we need ot make sur eBigInteger / BigDecimal .toString are properly specified and won't change. -- 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/OGM-12?page=com... ] Emmanuel Bernard commented on OGM-12: ------------------------------------- Hey Nicolas, I've applied your pull request with minor adjustments (license header, some test split) and move BooleanType to a pass through type rather than a string mapped type. Can I ask you a couple of things before closing this issue: - can you add some JavaDoc on Big* telling that their toString representation is the standard serialized formed as defined by the Big* JavaDoc - can you list the supported types in a wiki page under http://community.jboss.org/en/hibernate/ogm describing how they are persister (serialized or transformed) That would be awesome :) -- 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/OGM-12?page=com... ] Emmanuel Bernard edited comment on OGM-12 at 4/13/11 2:55 AM: -------------------------------------------------------------- Another thing we should do for types is to test that the types actually do what they claim they do. We can look in the grid and ensure that the type is truly serialized or truly a string representation. ManyToOneTest test has a test that check that the cache is empty, we could reuse the same concept on the BuiltInTypeTest to make sure the value in the map is truly what it should. was (Author: emmanuel): Another thing we'll for types is to test that the types actually do what they claim they do. We can look in the grid and ensure that the type is truly serialized or truly a string representation. ManyToOneTest test has a test that check that the cache is empty, we could reuse the same concept on the BuiltInTypeTest to make sure the value in the map is truly what it should. -- 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/OGM-12?page=com... ] Nicolas Helleringer commented on OGM-12: ---------------------------------------- JavaDoc part done Wiki will follow after my holidays ;) For test insuring that type descriptor do what they say they do I do agree (Had put a word on the topic in the last code commit of OGM-12 impl) -- 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/OGM-12?page=com... ] Emmanuel Bernard resolved OGM-12. --------------------------------- Resolution: Fixed Fix Version/s: 3.0-next Applied, Thanks Niko. -- 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