Tracker: how it is going to work
Hi all,
In this email I will explain the new issue management process we are going
to be using.
We have decided to go for a 3-headed scheme for managing issues. Our JIRA
contains 3 projects related to IDEA:
1. IDEA Feedback (IDEA, http://www.jetbrains.net/jira/secure/BrowseProject.jspa?id=10080)
2. IDEA Backlog (IDEABKL, http://www.jetbrains.net/jira/secure/BrowseProject.jspa?id=10131)
3. IDEA Development (IDEADEV, http://www.jetbrains.net/jira/secure/BrowseProject.jspa?id=10132)
The Feedback project is used for submitting ALL issues by our community members.
The Backlog project is used for keeping issues that we consider as worth
implementing in IDEA but cannot start implementing immediately.
The Development project reflects our development activity and used only for
our development activity planning.
The life cycle of an issue is the following. First it gets submitted to
the Feedback project. After being reviewed, discussed, refined if required,
etc it moves either to the main development project (if it gets planned for
the nearest release or milestone) or to the backlog project. Each planning
stage, either for the next milestone or the next version of IDEA requires
looking through the issues in the backlog project and moving them to the
development project together with necessary things required for development,
like breaking it up to pieces, estimating, prioritizing, and so on.
One exception is bug requests for EAP builds. They usually make it to the
Development project immediately from the Feedback project. It is our responsibility
to keep the number of requests in the Feedback project at minimum.
In addition, we will give a certain number of people from the community full
access to the Feeback project and a limited group will get access to the
Backlog project with an ability to move requests around.
Exception stack traces will still be posted to the old tracker until we find
a better solution.
Thus, to summarize: the Feedback project is the source of incoming information,
Backlog is everything that makes sense to implement, Development is what
we actually are doing currently.
Let me know if you have any questions.
Best regards,
Eugene Belyaev
President, CTO
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
请先登录再写评论。
Eugene Belyaev wrote:
>
How can I link them to newsgroups (in Thunderbird), or email threads?
Alain
You can subscribe to their RSS feeds to get notifications.
Best regards,
Eugene Belyaev
President, CTO
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
>> Our JIRA contains 3 projects related to IDEA:
>>
Sounds great - how often will you be trawling through the backlog? How
far apart are the milestones? Is there a beauty contest (voting, etc.)
where we can influence which requests make it through from Feedback to
Backlog to Development?
R
Eugene
>
Could you please elaborate a little, in the interest of everybody.
I'm clueless. Look at this page:
http://www.jetbrains.net/jira/browse/IDEADEV
No RSS on the page, no indication, nothing. Where, what, for whom. It's
bloated. I'm pissed.
If this web stuff is the only interface between you and us, I'll take
the back seat.
Alain
Alain,
I understand your frustration here. Just let us finish the things, we will
add all the required info to those pages.
Best regards,
Eugene Belyaev
President, CTO
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
>> You can subscribe to their RSS feeds to get notifications.
>>
Cool. This looks like a process somewhat similar to what we're doing
internally, though with the added element of EAP.
Of course, it would still have to be managed correctly to avoid issue
starvation (when the balance between ease of implementation and
importance is not kept).
So I guess the next EAP will report exceptions to JIRA?
Amnon
Amnon,
We are still thinking about what to do with exceptions. Usually they are
of two types:
1. things that can be fixed easily just because a developer overlooked some
condition: like boundary values, something can be null, etc. These can easily
be fixed and should just get fixed. Reporting on these does not make a lot
of sense for the users as well. Another thing is that EAP build exceptions
usually either should get fixed immediately or they become obsolete when
the next build gets published because a lot of code changes all the time.
2. Serious issues that need investigation and often should get accompanied
by a corresponding request that gets planned. So an exception results in
a request with a link to the original exception stack trace.
After a lot of experience in managing exception traces we do not think it
is correct to treat them as regular requests because they have a pretty different
life cycle as I described above. We are thinking of having a special system
for managing exceptions with an easy ability to link them to the tracker.
Best regards,
Eugene Belyaev
President, CTO
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
We have a new space at http://www.jetbrains.net/confluence/display/IDEADEV/Home
that eventually will replace the old intellij.net web site. You may comment
on the pages and we will be creating all the needed pages to help you get
around.
Best regards,
Eugene Belyaev
President, CTO
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
>> You can subscribe to their RSS feeds to get notifications.
>>
Eugene,
Thanks for the explanation.
I see exceptions as the most obvious form of bug, as they are simply not
supposed to happen (or at least not see the light of day, but get caught
and handled internally).
That's why it seems odd that they are not treated like other bugs. The
differentiation between bugs and feature requests is very clear most of
the time, so I think exceptions should have a life cycle similar to a
bug, most of the time a shorter one, as like you say, they become
obsolete, need immediate handling or further investigation (with the
last case being the closest to a 'normal' bug).
I think that if exceptions are treated similarly to bugs, and possibly
classified differently to allow easy filtering, etc. you would achieve
your goal of differentiating them while still maintaining a single place
for all development issues.
In any case, I can attest myself that I will try to closely match the
process you create, and the simpler it is, the less mistakes I will make :)
Amnon
Eugene Belyaev wrote:
In article <cuagi5$qob$1@is.intellij.net>,
Alain Ravet <alain.ravet@biz.tiscali.be> wrote:
Alain,
here is what I do
I go to the project I want to watch, I do somekind of search which makes
sense to me, then at the top of the list of search results there is an
xml link. Click on that. Copy the link to rss, and at the end of the
link you have to add:
&os_username=(your user name)&os_password=(your password)
You can create as many rss feeds as you like.
It's retarded but it works.
R
Exceptions expires very soon and have them as regular bugs is not good idea I'm think. For example many of them may be duplicates and so on.
Keep it couple of EAP and then close that's all.
Another way is to use system which will track developer code location and report that he is in class that caused reported exceptions to look at code and try to fix.
(Just a dream)
TIA,
Dmitry
Robert
>
>...
>It's retarded but it works.
>
>
With patience, time and resource, you can turn about anything into anything:
http://www.americanbamboo.org/images/BambooBicycle.jpg
Alain
Hello Eugene,
EB> We are still thinking about what to do with exceptions. Usually
EB> they are of two types:
As i know, JIRA can file isues from POP inbox. Users can simple authorized
send email into specibied box,
for example, bugs@jetrains.com. This sending may perform without user interraction
or via user email client.
In second case, you need to ask user keep existing letter template and subject
without global changes.
Thanks!
--
Alexey Efimov, Java Developer
Tops BI
http://www.topsbi.ru
In article <266309632434971194786744@news.jetbrains.com>,
Alexey Efimov <aefimov@zebra.ru> wrote:
And cause TONS of duplications because people won't think about looking
first.
R
Alain Ravet wrote:
You can link the XML url as a feed in thunderbirds RSS reader support.
You may need to add &os_user=USERNAME&os_password=PASSWORD to the URL thou
Mark Derricutt wrote:
I tried that, but - please pardon the language - it sucks. For at least two reasons:
- Thunderbird seems to be unable to maintain more than one feed to the same site.
Ok, this is a tool limitation, but still it would annoy the hell outta me to switch
to anything different.
- The items I do get from the feed are always the initial description of an entry.
They do not include any reason or description of the change that caused it to appear
in the feed. Also, additional comments don't get listed at all. I have to click
the link and try to make out the recent change in the JIRA page. This seems to make
it virtually impossible to conveniently follow a discussion if you don't have an email
notification on an issue.
Sure this is no big deal if the new system helps JB to be able to better organize
their work, but the RSS thing can not even remotely replace the current web <-> news
synchronization. At least that's what I experienced. Maybe I'm doing something wrong
though.
Sascha
Sascha Weinreuter wrote:
I hate to say it by Thunderbirds much toted RSS support SUCKS MONKEY NADS.
It's so brittle and a freaking hack thats shoehorned into the interface
in a most hideously manner.
Thunderbird can be set to automatically load the full link, or use the
preview description given in the feed. Try changing that setting...
Mark
In article <cubd93$os6$1@is.intellij.net>,
Sascha Weinreuter <sascha.weinreuter@NOSPAM-cit.de> wrote:
This is not thunderbird only, this is with all RSS
Hello Robert,
RS> And cause TONS of duplications because people won't think about
RS> looking first.
JIRA must detect duplicates and link issues. I think that if issue already
exists, and user get exception, then his MUST post it, for example to have
notification about resolution it.
Thanks!
--
Alexey Efimov, Java Developer
Tops BI
http://www.topsbi.ru
RSS works quite well with Firefox:
Bookmarks / Manage Bookmarks / File / New Live Bookmark
Then enter the URL you got from from the "XML" link
(mayber append that "&os_user..." stuff)
View / Sidebar / Bookmarks
and the RSS feed will appear in your Bookmarks list...
Robert S. Sfeir wrote:
>>Ok, this is a tool limitation, but still it would annoy the hell outta me to
>>switch
>>to anything different.
The only problems I've seen in this area with most aggregators is the
lack of supporting multiple authentication profiles to the same host.
I have about 4 JIRA feeds regularly updated in my RSS tool...