Introducing time wasters
gromero at linux.vnet.ibm.com
Sat Sep 22 13:11:59 UTC 2018
On 09/22/2018 08:57 AM, Ivan Gerasimov wrote:
> I think that number of votes for the bug can indicate urgency.
> However this functionality do not seem to used too often (for example a query 'component = core-libs and votes > 1' returned only 8 issues for me).
oh we do have a counter already... I totally missed that.
> Maybe we should use (and pay attention to) the votes more actively?
I really think that having a counter instead of a static label is the best
way to convey the idea we are discussing here, so +1.
...maybe modify "Votes:" counter name or add another named like Mark and
others suggested for the label name, as "affects-maintainers"
(+1 for that BTW). But use a counter and not a static label either way.
> Just an opinion.
> With kind regards,
> On 9/20/18 9:04 AM, jesper.wilhelmsson at oracle.com wrote:
>> OK, I see that this didn't come out as I intended it. The name "timewaster" is clearly not the best since it can be interpreted as "this bug is a waste of time - it shouldn't have been filed" - That was NOT what I meant.
>> My intention was to have a label to indicate that a bug is urgent to fix because a lot of developers is currently spending time chasing down the same issue over and over.
>> It has been suggested to increase the priority instead of using a label. One of the reasons that it ended up as a label is that JBS don't have a way to express urgency. Priority and urgency are two different things and should not be mixed, so raising the priority would send the wrong signal. One example is a broken test in tier 1 that is failing at every run. This bug would cause work for every engineer who is running through our tests, be it through the submit forest or in internal test frameworks in our different organisations. Fixing it is clearly urgent. However, raising the priority would mean that it becomes a blocker for the release. A broken test should not block the release.
>> So, back to the name. Since urgency is what we're after, can we call it "urgent"?
>> As for the "discussed, proposed, and adopted" part - This is where it is proposed and discussed. (Thanks to everyone who engage in the discussion!)
>> It hasn't been adopted yet, it is, as I wrote, an experiment.
>>> On 20 Sep 2018, at 02:04, jesper.wilhelmsson at oracle.com wrote:
>>> As an experiment we are introducing a new label in JBS, timewaster. The label is used to tag bugs that for some reason is wasting engineering time. This could for instance be a bug that occurs frequently in our testing and causes many engineers to investigate the failure in different test runs just to realize it's the same issue as have been seen before, or worse, don't realize it is a known issue and files a duplicate bug in JBS. Bugs that cause tests to fail without a proper explanation may also be considered time wasters since several engineers risk to investigate the failures just to realize there is no information to be found. There are other cases as well and there is some flexibility in the definition. If you see an issue that have been wasting your time, feel free to add an explanation to why you think it's a time waster and add the label.
>>> The fact that a bug is wasting engineering time should be taken into account when a bug is triaged. A time waster has higher urgency than other bugs. This is not really reflected in the priority of a bug so the priority is not necessarily changed due to the higher urgency. When triaging a time waster, or when adding the label after a bug already has been triaged, it's recommended to notify the developers that work on the affected component, e.g. by sending an email to the proper mailing list and make sure the bug is assigned to someone.
>>> If you have a time waster assigned to you, please consider fixing it asap. If you chose to not work on the issue, you should at least be aware that you are choosing to waste more engineering time and others will be affected by this choice.
>>> To see currently open time wasters use the JBS filter Time wasters: https://bugs.openjdk.java.net/issues/?filter=35335
More information about the jdk-dev