<< Back to previous view

[JDBC-106] README should not specify 3rd party driver versions Created: 11/Mar/15  Updated: 18/May/15  Resolved: 18/May/15

Status: Resolved
Project: java.jdbc
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Trivial
Reporter: Sean Corfield Assignee: Sean Corfield
Resolution: Completed Votes: 1
Labels: None


 Description   

Per https://github.com/clojure/java.jdbc/pull/39 the README is constantly getting out of date regards latest third party driver versions.

Better to remove all versions and just link to Maven Central so folks can quickly click through to find the latest versions for themselves.



 Comments   
Comment by Sean Corfield [ 18/May/15 11:57 AM ]

Updated README to link to Maven Central.





[JDBC-107] metadata-result leaks ResultSets Created: 12/Mar/15  Updated: 18/May/15  Resolved: 18/May/15

Status: Resolved
Project: java.jdbc
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Defect Priority: Major
Reporter: Niels van Klaveren Assignee: Sean Corfield
Resolution: Declined Votes: 0
Labels: None


 Description   

When working with Oracle DatabaseMetaData, I noticed out of cursors exceptions (ORA-01000) when iterating over tables and requesting their .getIndexInfo and .getImportedKeys information. These errors tend to point to leaked resultsets. Investigating the metadata-result code I noticed this does happen, and the result set are just leaked silently on the other databases I tend to use.

A fix would include a default result-set-fn doall in the :or keys clause, and binding the true clause of the (if (instance? java.sql.ResultSet rs-or-value) to a new var, .close the rs-or-value, and return the new var.



 Comments   
Comment by Sean Corfield [ 12/Mar/15 11:57 AM ]

Thanks Niels. I'll take a look at that shortly. I may have some time next week.

Comment by Sean Corfield [ 18/May/15 11:38 AM ]

I finally got around to looking at this – apologies for the delay – and I'm not convinced metadata-result should close the ResultSet since that is passed in by the user. java.jdbc itself only ever calls metadata-result in a test so I updated that with the way I think it should be handled, by the user:

(sql/with-db-metadata [metadata db-spec]
  (with-open [result (.getTables metadata ...)]
    (let [table-info (sql/metadata-result result)]
      ... do stuff with table-info ...)))

The point here is that the user creates the ResultSet, not java.jdbc, so the user should close it correctly.





[JDBC-108] Remove Double Counting of Options in insert-sql Created: 18/May/15  Updated: 18/May/15  Resolved: 18/May/15

Status: Resolved
Project: java.jdbc
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Enhancement Priority: Minor
Reporter: Earl St Sauver Assignee: Sean Corfield
Resolution: Completed Votes: 0
Labels: None

Attachments: Text File doubleCountEstsauver.patch    
Patch: Code

 Description   

insert-sql ~(line 965) iterates through the options to count the number of rows + Col/Vecs.



 Comments   
Comment by Sean Corfield [ 18/May/15 11:23 AM ]

Thanks Earl. Since you're not on the clojure.org/contributing list, I can't use the patch directly but I went ahead and removed the duplicated calls to `count` that you pointed out.





Generated at Fri May 22 12:49:27 CDT 2015 using JIRA 4.4#649-r158309.