On Jun 11, 2007, at 12:12 AM, Pavel wrote:
Just to clarify how do I see that relationship.
First off, I was talking about stable branch, not a release.
Generally it is not possible to fix all major bugs and make a
perfect application in just one release.
Nor it is possible when bugfixing is mixed with development.
In my understanding testing is closely related to branch when it
comes to achieving stability yet avoiding development freeze.
Lets assume functional testing will start for XWiki 1.1 and all
commits (development and bugfixing) would go into head only.
That is, major areas will be covered by functional testing, number
of issues will be identified, and fixed at some point later.
Given that development activities will take place over the same
period of time, the following result is possible:
1. Certain functionality was not covered by testing, since it
appeared post 1.1. Probably it contains some issues.
2. The functionality that was covered by testing may have
regressions, since lots of changes were made since 1.1.
3. Even something that was fixed may have been broken by consequent
changes.
In other words, the more development effort takes place, the less
value from QA cycle remains in the application.
With branch OTOH the outcome would also include more or less stable
1.1 release that ideally has all major bugs fixed.
There is a cost though. The overhead and burden of applying fixes
into 2 places is something that nobody likes. And it slows down the
development to some extent.
That was my vision =).
I agree. As we both said in our answers, the major issues is the
overhead and burden of applying fixes in 2 places. This is an
important issue because we need to go forward and we don't have that
much manpower. Which is why I was saying that if 1) there are major
issues (like security issues) then we should do it - if it's
possible; it may not be if there have been important refactorings to
trunk - and if 2) there are contributors/committers interested in
making it happen.
Regarding the testing, yes you're right, we're trying to mitigate the
risk of using 1.1 by:
* Adding more and more tests
* Delivering new releases every 3 weeks. So if a user using a 1.1
release find a bug and if this bug is added to jira soon enough it'll
be fixed in the next release (thus within 3 weeks)
Anyway, this only represent my POV and I'm happy to change it if
required :)
Thanks
-Vincent
On 6/10/07, Vincent Massol <vincent(a)massol.net>
wrote:
Hi Pavel,
On Jun 9, 2007, at 9:49 PM, Pavel wrote:
Sorry to rise the question again, but now that
functional testing
is about to start, I believe decision on the subject has already
been made.
Why do you say this? I don't see the relationship between that
functional testing and a decision on bug fix release... ;)
I've prepared a mail for you several days ago but I hadn't sent it
so far because I'm not sure about the answer. We take decisions
here together and we haven't discussed this yet so this is a good
time...
This is indeed an important question. I think this is a community
decision and one for the current committers. My personal feeling (I
only speak for myself) is currently the following:
* Most xwiki committers will probably want to work and focus on
1.1. I'm also starting to work on 2.0.
* In order for XWiki 1.1 to become even better we need the maximum
of users to try 1.1.
* I think 1.1 will stay stable enough for users to migrate to it or
at least to try it and migrate if they like it
* Our goal is to increase our automated test suite to reduce the
risk of regression and allow users to safely move to any release we do
* We may need 1.0.x releases but that's hard to do as there aren't
many active committers
* I discussed this with Sergiu and we think there are some
occasions where a bug fix release should be done. One of them is
when there are important security issues discovered and fixed
But I guess the real answer is: if a committer wants to do a bug
release and is volunteering for leading it, I think everyone would
agree. If volunteers like you can/want to contribute to a bug fix
releases then I guess the likelihood that it'll happens will be high.
What do other committers think?
Thanks
-Vincent
On 5/29/07, Pavel <pagrus(a)gmail.com>
wrote:
Vincent et all,
Now that xwiki is 1.0, are there any plans to have a bugfixing-
only branch?
1.1M1 anouncement mentions both bugfixes and new features/changes.
While the
former is good for stability, the latter may not be so.
For me as a person who will need to support xwiki installation it
is an
important question.
Thanks,
Pavel
--
Best regards,
Pavel
--
You receive this message as a subscriber of the xwiki-
users(a)objectweb.org mailing list.
To unsubscribe: mailto:xwiki-users-unsubscribe@objectweb.org
For general help: mailto:sympa@objectweb.org?subject=help
ObjectWeb mailing lists service home page: http://
www.objectweb.org/wws
--
You receive this message as a subscriber of the xwiki-
users(a)objectweb.org mailing list.
To unsubscribe: mailto:xwiki-users-unsubscribe@objectweb.org
For general help: mailto: sympa(a)objectweb.org?subject=help
ObjectWeb mailing lists service home page:
http://www.objectweb.org/
wws
--
You receive this message as a subscriber of the xwiki-
users(a)objectweb.org mailing list.
To unsubscribe: mailto:xwiki-users-unsubscribe@objectweb.org
For general help: mailto:sympa@objectweb.org?subject=help
ObjectWeb mailing lists service home page:
http://www.objectweb.org/
wws