This has been reported already upstream at http://code.djangoproject.com/ticket/6433. Unfortunately both patches (yours and the one on that ticket) have slightly interesting bugs...
For example, the patch on that ticket would parse 8.3RC3 -> [8, 33] which is (clearly) wrong. Yours has an issue that isn't currently a problem in postgres, but maybe at sometime in the future, if you had the string 10.1RC3 your version would end up with [1, 0, 1] which is obviously not the desired effect! I'm looking in to this bug, and will be tracking upstreams dealings with it. Thanks! -- Brett Parker -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
