db-derby-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Apache Wiki <wikidi...@apache.org>
Subject [Db-derby Wiki] Update of "DerbyCommitHowTo" by BryanPendleton
Date Sat, 04 Mar 2006 00:20:38 GMT
Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Db-derby Wiki" for change notification.

The following page has been changed by BryanPendleton:
http://wiki.apache.org/db-derby/DerbyCommitHowTo

New page:
= Committing patches to the Derby Subversion Repository =

Oyvind Bakksjo compiled this set of notes to guide Derby committers through the low-level
mechanics of committing a change to Derby. These notes are best read within the larger context
of the DerbyCommitProcess.

Committing contributions from others is not as straightforward as one might think. This mainly
stems from the fact that there is an svn diff command, but there is no svn patch command.
There are a number of things to remember; if not done carefully, you might end up doing partial
commits that may break the build. This page attempts to give a recipe for safely committing
code contributions. 

* Make sure you have a clean sandbox:
  * In the trunk directory, run svn status - it should not list anything. If it does, you
may want to run svn revert -R . to remove all local modifications, or use/check out a different
sandbox. 
  * Run ant clobber.
* Check your sandbox' svn revision with svn info.
* Align your sandbox' svn revision with the patch's:
  * If the contributor has indicated on which revision the patch was created (with svn diff),
run svn update -r REVISION. 
  * Otherwise, you may choose to run svn update -r { DATE }, using the date of the contribution
of the patch. NOTE: svn is very picky about date formats. See [http://svnbook.red-bean.com/en/1.0/ch03s03.html
these notes on valid date formats].
  * If you think there has been no changes in the patch's area since it was contributed, you
can try on the head: svn update. 
* Apply the patch in the trunk directory: patch -p0 < PATCH_FILE.
* Make sure there were no conflicts when applying the patch. On unix, you can run find . -name
"*.rej" -print. If a *.rej file is found, you should either resolve the conflict or ask the
submitter to submit a new patch against the newest revision. If you omit this step, you'll
still catch this problem before committing (when you compare your svn status output to the
contributor's), but it's a good idea to catch this early on. 
* Run svn status, and check for AddedOrDeletedFilesInDiff.
* By now your svn status should be equal to the contributor's. Now compare the diff by running
{{{svn diff | diff - PATCH_FILE}}}. The diffs you see could be the following:
  * Revision numbers, if your sandbox was not aligned with the contributor's (see above)
  * Subversion file properties: You will have to set svn properties for AddedOrDeletedFilesInDiff.

  * Actual code diffs: Inspect and figure out what/why.
Files may be examined in different order by {{{svn diff}}} on your machine and the contributor's;
not sure if                   there's anything we can do about that.
* Update to the head with svn update. Check that there are no conflicts.
* Build the code with ant all and ant buildjars.
* Run tests if you are not confident that the contributor's/reviewer's actions are sufficient.
Check that tests pass. If they do not:
  * Check the nightly/tinderbox for the same failure.
  * Verify that the failure is not caused by the patch.
* Commit the patch with {{{svn commit}}}. Use either {{{--message}}}, {{{--file}}} or {{{--editor-cmd}}}.
Include the following in the commit message:
  * The ID of the JIRA issue. Make sure you use the format DERBY-NNN so that JIRA picks it
up.
  * Some text explaining what the patch does (typically snipped from the JIRA issue).
  * The contributor's name/email.
* Send out an email to let people know you have committed the patch. This is done automatically
if you post the information as a comment to the JIRA issue.


Mime
View raw message