Policy on trackers in Bugzilla revisited

Stewart Gordon smjg_1998 at yahoo.com
Sun Mar 17 14:54:53 PDT 2013


There seems to be disagreement between various users on the propriety of 
trackers.  These are bug reports that don't describe a single bug, nor a 
feature request, but are used to group together related issues.

Trackers (also known as meta bugs or umbrella bugs) are used heavily on 
Mozilla's own Bugzilla, as you can see here:

https://bugzilla.mozilla.org/buglist.cgi?keywords=meta&resolution=---

I shall also quote for convenience the description from the keywords 
page there:

"A placeholder bug for tracking the progress of other bugs. Meta bugs 
are made dependent on other bugs so that interested parties can be kept 
up-to-date with status via one bug, without having to receive all the 
mails related to all the bugs related to the development of a particular 
area."


Just before the concept was brought here, there was a thread on it
http://www.digitalmars.com/d/archives/digitalmars/D/bugs/7640.html
in which nobody expressed any objection, and then a few trackers were 
then filed.

However, since that time, two or three people have been killing off 
random trackers, seemingly because they personally don't like the 
concept.  The arguments have included:


1. That's what keywords are for.

It's hard to draw the line between the what keywords are for and what 
trackers are for.  But there are a number of functional differences:

- anybody can create a tracker, whereas only someone with the necessary 
access to the Bugzilla configuration can create a keyword

- trackers can track other trackers, but keywords can't track other keywords

- trackers can generate an email notification when one of the tracked 
bugs changes status

But in practice, some groupings of bugs are more suited to trackers, and 
others to keywords.  For instance, if there are likely to be a large 
number of bugs in a given category, a keyword would probably be better. 
  But on a Bugzilla installation having a million or so bug reports on 
it, loads of keywords that each group together 5-10 related issues would 
bloat the keyword list, and so trackers would be more likely to be used 
for these.


2. It doesn't help in getting them fixed/Nobody fixing D bugs has found 
them useful.

Bugzilla isn't about the few people who are using Bugzilla primarily to 
find and fix bugs.  It's about maintaining a list of bugs in a given 
software product or line of products.  The majority of users are using 
it primarily to file bugs, discuss bugs and to some extent perform 
housekeeping tasks such as flagging duplicates.

Thinking about it I suppose that, for the most part, the people who find 
trackers useful are those who use them for tracking.  That is, to keep 
up to date with the progress in fixing bugs in a certain area of the 
software.


3. Such lists of bugs belong on a wiki, not on Bugzilla.

Having such lists on a wiki or other separate website might have some 
advantages, such as being able to present the information in a better 
way than is possible under Bugzilla.  But for this to work we would need 
to make sure people are aware of this external facility.  There are 
further logistical challenges when it comes to maintaining it given that 
it would be separate from Bugzilla.  A simple wiki page that lists bugs 
does nothing to make it easier for users to see how many of them are 
resolved.

I suppose a web-based tool could be written to provide tracking 
functionality.  At the most basic level, it would store lists of bug 
numbers, and use the CSV output mode from Bugzilla to count up bugs from 
these lists by status.  A more sophisticated tool of this sort would 
include an email bot that subscibes to updates on the tracked bugs and 
relays them to interested parties.

But I suppose most people in the Mozilla bug-reporting community would 
argue that this is reinventing the wheel, and you might as well just 
have trackers within Bugzilla.


4. The status field of a tracker is meaningless/likely to be 
meaninglessly changed.

There was a brief exchange about it on this thread
http://forum.dlang.org/thread/5112D61B.5010905@digitalmars.com?page=2

on which it was pointed out that trackers will tend to be closed and 
reopened as all the bugs it tracks are fixed and then a new one is 
opened.  Obviously this isn't ideal, but it's only a small maintenance 
overhead (unless you get rid of it by deciding to just leave the tracker 
open all the time).

Perhaps better would be a new Bugzilla feature whereby an issue can be 
designated as a tracker, such that its status automatically depends on 
the status of its dependent bugs.

But even in the absence of this feature, it doesn't strike me as a 
reason to ban trackers.


Anyway ... I suppose what all this goes to show is that trackers have 
their pros and cons.  If they were all bad, they certainly wouldn't have 
thrived on b.m.o for all these years.

As such, while some people may not like trackers, others find them 
useful, and they don't seem to be doing any harm.  As such, my 
recommendation would be to keep them now that we have them.

What do other people think?

Stewart.

-- 
My email address is valid but not my primary mailbox and not checked 
regularly.  Please keep replies on the 'group where everybody may benefit.


More information about the Digitalmars-d mailing list