There's already a bug open on CvsChangeLog regarding it's date processing -- I was going to include the "upgrade" to the extension of the AbstractCvs class with those changes.

Is that OK, or do you want another defect logged? I don't mind submitting in stages.

        Ken

At 19:18 2003-02-09, Conor MacNeill wrote:
Ken Gentle wrote:
Well, I've got an edited version of CvsChangeLog that can handle the "start" and "end" date parameters with a little more flexibility.
I've added an optional attribute, "pattern", which is expected to be a SimpleDateFormat pattern string. I've modified the tasks "setStart" and "setEnd" methods to have string parameters (as opposed to the "Date" they currently have), and have taken steps to ensure the old behavior works.
So as I'm shamelessly stealing test case code from Stefan's AbstractCvsTaskTest, I'm seeing that the ChangeLogTask depends completely on the external environment for CVS parameters (CVSROOT, CVSPASS, etc).
Wouldn't it be desirable for the CvsChangeLog task to behave like the CVS and CvsTagDiff tasks?

Yes, I think if we have an abstract class for CVS tasks, it would make sense to use it. Please open a BugZilla bug and attach any changes to it - I'll commit as soon as I can.


Conor



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

============================================================= J. Kenneth Gentle (Ken) | Phone: (610) 255-0361 Gentle Software, LLC | Email: [EMAIL PROTECTED] =============================================================



Reply via email to