Apache OpenOffice (AOO) Bugzilla – Issue 94564
duplicate issue, same submitter/reporter, sent fast: auto reject
Last modified: 2017-05-20 09:19:24 UTC
I didn't know I was submitting the same issue twice in one hour. I thought the first attempt hadn't gotten through, because my connection to the Internet at that moment was spotty, and I didn't get an acknowledgment, so I kept trying. The 2 posts were about 20 minutes apart. As far as I know, both issues that I sent were completely identical. Result: Someone at your end had to read 2 issues, process 2 issues, and mark one as a duplicate of the other. You don't need extra work. Suggestion: Could your issue tracker notice that an issue is a duplicate of another by the same reporter? Then, instead of the standard acknowledgment, post a semi-acknowledgment, like "Your newer post appears to be an absolute duplicate of Issue . . . [#] already received. The earlier Issue will be retained and the newer one not posted. The OOo community is aware of the earlier Issue. Should you wish to add information or describe a different issue, please edit and post a new comment or Issue. Thank you." I suggest a 2-hour window, which is long enough to accommodate differences in system time, including old accumulated errors and savings time and time zone differences, between nearby terminals. A short window would mean no need to run a search for duplicates; if the absolute duplicate is of a 3-year-old issue, post it as you do now, in case modern times are a new context for the issue. Instead, a design approach could be to retain a copy of each posted issue for 2 hrs, to compare a new post to the 2-hr bank, and to delete from the 2-hr bank whenever a new post finds anything in the bank older than 2 hr, so you don't need to run a daemon or loop just to delete over-aged issues from the bank. If anyone deletes a post, the deletion process has to check the time of origin; if a post could be 2 hours or less old, the deletion process would also delete the copy. For more flexibility, near-duplicates should also be handled. An approach to them could add a comparison page showing fields that differ, e.g., Platform as Unknown or PC, and report that all other fields are identical, and then ask the submitter whether they prefer the old or new posting, possibly providing access to the comments field of the preferred issue for anything the submitter would like to clarify, then delete the unwanted issue and post the desired one, with today's standard acknowledgment. In the event that someone has already replied to or changed the earlier post, such as if it has been reassigned to a different person, then the submitter of the near-duplicate issue would not be given the comparison page but instead would be advised that the new issue appears to be a near-duplicate of older Issue [#] which has already had some action taken on it or a reply added to it and the tracker wold reject the new issue. The length of the window should be written to a single place in your issue tracker program with everything else referring to that one place, so you can adjust the window or bank with experience. Thank you. -- Nick
obsolete