Thomas Zimmermann <[email protected]> writes:

> I sorted the bugs in SHR trac into the several milestones after i was away 
> for 
> some months.
>
> But now i would be interested in: Which SHR-unstable bugs are the most 
> annoying ones for you?
>
> I tried the 21. lite image and i would say that are the following one:
>  * Shutdown with quickstettings needs 2 minutes, because of a timeout.
>  * Connecting to GPRS doesn't work because fsogsmd crashes.
>  * Next button in first start wizzard doesn't work, so it's hard to get 
> stared.
>
> But what do you think?

Are you asking for yourself, or for the SHR team as a whole?

It seems a bit strange to ask, because

- if it isn't already obvious to you which are the most serious bugs,
I'm sure you will get conflicting answers from the people on this list

- given a pool of bugs that are roughly equally serious, I think it
would make sense for you and your team colleagues to pick off whichever
ones you personally feel most comfortable about tackling.

Now it may be that that strategy may leave a couple of hard bugs that no
one in the current team feels like taking on.  But that's OK, because by
clearing up the other ones you will have clarified the overall picture,
and you'll motivate someone really hardcore to come in and tackle those
last remaining bugs.

Also I'd like to mention testing (again, having recently raised it
somewhat inappropriately in another thread).

For me at the moment, arguably the #1 bug is that it seems to happen too
frequently that things that were working are randomly broken.  For
example, the mailing list at the moment has discussions about GPS and
GPRS having broken, whereas previously they were working.  Yeah, I know
it's called "unstable", but do you want SHR to stay unstable forever?

It doesn't have to be like that; I think you just need to start building
up a regression test suite.

(I hope I'm not being unfair here by ignoring test suites that you
already have.  As an example, I just took a look at
http://git.shr-project.org/git/?p=phonefsod.git;a=summary, and I don't
see anything in the tree there that looks like a test suite.)

So, in summary, I'd suggest that

- as a team, you decide from now on to write a regression test for
  everything that you fix, and put in place whatever basic
  infrastructure is needed for that

- after that, you each work on whichever bug you feel most effective -
  based on a combination of your ability and how important it seems.

I hope this comes across as trying to be constructive (and not
condescending).  You guys have already achieved wonders to get SHR to
where it is today.  If only it's possible to fix the few remaining
important issues, without breaking anything else, I think you're very
close to completing a reliable and exciting daily phone system.

Regards,
        Neil
_______________________________________________
Shr-User mailing list
[email protected]
http://lists.shr-project.org/mailman/listinfo/shr-user

Reply via email to