harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tim Ellison <t.p.elli...@gmail.com>
Subject Re: Problems about the harmony
Date Fri, 30 Nov 2007 16:44:01 GMT
junguo pu wrote:
> I am the resposibel group leader of Evermore.

Thank you for your testing and feedback -- it is very valuable.

> Having worked on harmony(Linux) for some days, we found that harmony is not
> stable. Here I will say something from the macro aspect.

While it is good that you are brave enough to try running from our HEAD
stream, we don't make any claims about its stability since that is the
active on-going development area.  As you probably know, we produce
regular milestones which we *do* consider to be stable builds.

> 1.first of all it doesn't provide the install version for users. If user
> downloads the version from the web, he should deploy the environment
> himself, or else harmony will crazy you.

We have no versions of Harmony yet targeted at end users, though we have
started some early work at producing packaged installs in preparation.
At the moment, we ask developers (like yourself) to take milestones and
report problems, and we are very grateful for the help in improving the
code based on everyone's feedback.

At this stage we would expect developers to be able to download and
unzip/untar the milestones directly without much trouble.

> 2.I don't know how Harmony control his code. Before when they fix their bugs
> one day, but it will happen again another day. You'll see the previous code.

I expect you are talking about HEAD again, since we are very sensitive
to regressions between milestone releases.  While there may be
regressions in the latest development stream we do try to minimize them
by running continuous integration tests suites, but at the rate and pace
of change in the code some minor breakages may be expected.

If you see a regression that lasts more than a week or so, then it
should definitely be flagged as an issue since we may have missed it.

> 3.The most important thing is we can't debug our program, which is a
> magnificent and complex program, although you can debug a small smoothly.

Based on Gregory's note I hope this problem has been resolved and you
can make progress once more.  Working at the very edge of the latest
code is exciting and a bit dangerous at times!

Thanks Junguo for your views, and keep them coming.  The feedback is
important for everyone to hear.

Regards,
Tim


Mime
View raw message