On Feb 5, 2008 6:01 PM, chasd chasd@silveroaks.com wrote:
User reports: "Email causes loop"
One issue is that some of the reports may be on the users mail list, some may be on the forums. I don't know of a way to get those two ( and others I don't know about ) reconciled. If two users on the mail list report a bug, and two users on the forums report the same bug, it is the same as if _only_ two users report the bug and not four. Obviously Trac is where that should be consolidated, but at what point is a bug "confirmed" by the community ?. Four users run into the bug ? Three ? Five ?
1+ can be confirmed. As long as someone can upload an email to reproduce this, it's confirmed by a single person.
The difficulty lies within those reports "RoundCube doesn't work in Firefox", then a dozen people say, "works for me" or "yeah, same problem". ;-)
It doesn't take much to confirm a bug, it's just about the reproduction. As long as that is given, we can always fix it.
What we would like is:
X amount of people test and say, "Correct, I can reproduce this behavior." and maybe upload a bunch of emails to testing.roundcube.net in the process and let us know about it.
OK, I've used RC for over a year and never knew about that site. Of course, I haven't had a need.
Believe me, I am glad.
For others in need (this is linked from the trac homepage): http://trac.roundcube.net/wiki/Howto_ReportIssues
I also frequently pass around this link on IRC and on the mailinglists.
I know we don't always have the fastest turn-around on bugs and tickets but generally what we also lack is more feedback.
Yes, I think what feedback you have is too diffuse ( spread out ).
Someone opens an issues and is never to be seen again.
I call those drive-by bug reports.
I'd welcome drive-by patches. ;-)
The user has an issue, fires off an e-mail to some list, and thinks that is enough. They may even unsubscribe from the mail list at that point because of "too many messages." I know of one acquaintance of mine that joined a list to report a bug ( not RC ), and then set his spam filter to block all the traffic from the list. Didn't even bother to unsubscribe. Those types of reports we can't help, even if it is an actual bug. I would hope that if the bug is serious enough, someone else will run into it, and they will have what it takes to champion it to Trac and provide enough information for the bug to be verified and possible fixed.
Open source software is based on community, which means ( at least to me ) the end users must take an active role in reporting and following through on bugs. Most users are used to commercial software, where it not only is not expected for the end user to provide feedback, but it is actively discouraged in many cases.
So IMHO any messages reporting bugs on the Users list or on the forums should be funneled to
Yes, please.
But only to a certain extend. We also get the "I can't send email"-stuff which in 99.99999% of all cases is config related. But I guess there are various things how we can improve that etc..
so bug reporting is consistent for the developers.
And when it works for us, we close with WORKSFORME. Which doesn't rule out that the error is still stuck somewhere in RoundCube.
In order to "reward" those who spend their time to provide decent feedback, is there any feedback from the developer(s) when a bug has been verified ?
I try to confirm them too. I am trying to check Trac's timeline every day and work through issues. It obviously doesn't work always like that, but generally that's what I try. And when I say, "I", I really mean "we". I know are doing the same.
OK, what about RFE's though ? They can't be "confirmed" in any way. Popularity is not always a good indicator of priority either.
Would viable RFE's get included on
http://trac.roundcube.net/roadmap
or
http://www.roundcube.net/about
under planned features ?
How will those RFE's not determined to be viable be handled or notified ? I assume all RFEs should have a Trac ticket.
Could have a trac ticket, I suppose. There is nothing wrong with it. But currently it doesn't always work like that.
Cheers, Till _______________________________________________ List info: http://lists.roundcube.net/dev/