Difference between revisions of "LinuxMCE sqlCVS commit Procedure"

Revision as of 15:34, 19 March 2008

NOTE: This is the tentative procedure for handling sqlCVS commits

• Step 1: Create a Mantis ticket for the commit.
  • This should describe in detail what the purpose of the commit is.
  • Each commit should only fix one bug or add one feature.
  • If sql2cpp must be run this should be noted in the ticket as well.
  • If there is any code associated with the commit it should be attached to the ticket.

• Step 2: Validate your sqlCVS setup with "sqlCVS -v"
  • If there are any problems please fix these before doing the commit.

• Step 3: Commit the change to sqlcvs.linuxmce.org with a comment that refers to the Mantis ticket.
  • Use the command line tool "/usr/bin/pluto/sqlCVS", and commit using "-U anonymous~nopass".
  • The comment and the reference to the Mantis ticket will allow the person validating the commit to find the commit in sqlCVS.
  • Now monitor the ticket for any reported problems.

• Step 4: Someone other than the person committing the fix will validate the sqlCVS commit.
  • If any problems are found going through these steps note this on the Mantis ticket and stop.
    • Step A: pull the anonymous commit.
    • Step B: Run "sqlCVS -v"
    • Step C: Check the commit for any duplicate vendor or product names.
    • Step D: Check the mantis ticket to see if the description is complete.
    • Step E: If sql2cpp must be run with this database, do that as well and commit the code.
    • Step F: If there is any associated code, make sure that gets committed as well.
    • Step G: Commit the change under your own validated login.
    • Step H: Note the commit batch in the Mantis ticket and close the ticket.