We have 554 unconfirmed bugs older than 90 days. Most of these are likely to be obsolete. It would be really easy to write a script to ping the people who submitted those bugs, and ask them to update the bug if it's still valid. If there's no reply in two weeks, we could mark the bug 'resolved abandoned'.
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
I'd like to send this once packages are ready.
Comments?
On Fri, Sep 30, 2005 at 11:13:46PM -0700, Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days. Most of these are likely to be obsolete. It would be really easy to write a script to ping the people who submitted those bugs, and ask them to update the bug if it's still valid. If there's no reply in two weeks, we could mark the bug 'resolved abandoned'.
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
I'd like to send this once packages are ready.
I am not sure that I like this approach. It might appear to our users that we do not care.
The timelines are a bit strict too. ;)
Ciao, Marcus
I agree,
ppl take the time to notify bugs and what you are basically saying is
"we havenot even got around to looking at the bug you submitted, we are hoping it may have gone away but just incase there is still a bug in our software could you please spend some more time to look at it again , update the bug report and then we will have a smaller list of old bugs that we may still not look at. But at least the bug stats will look good.
"If you dont reply within two weeks , we'll ignore the bug anyway and mark it as resolved.
"Thanks for your time in supporting our software. "
Maybe not the best approach.
On Sat, 01 Oct 2005 12:43:02 +0200, Marcus Meissner marcus@jet.franken.de wrote:
On Fri, Sep 30, 2005 at 11:13:46PM -0700, Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days. Most of these are likely to be obsolete. It would be really easy to write a script to ping the people who submitted those bugs, and ask them to update the bug if it's still valid. If there's no reply in two weeks, we could mark the bug 'resolved abandoned'.
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
I'd like to send this once packages are ready.
I am not sure that I like this approach. It might appear to our users that we do not care.
The timelines are a bit strict too. ;)
Ciao, Marcus
Marcus Meissner wrote:
I am not sure that I like this approach. It might appear to our users that we do not care.
I bet we can word the message to avoid that appearance.
We could even have the message simply say
Are you still interested in this bug?
rather than asking them to retest the bug. That would have nearly the same effect, and sounds a lot nicer. I've updated the page to reflect this more friendly wording.
The timelines are a bit strict too. ;)
The message makes it clear that the bug can be reopened at any time. I think that takes away the strictness you're worried about.
How do you feel about it with those changes? - Dan
On Sat, Oct 01, 2005 at 11:07:39AM -0700, Dan Kegel wrote:
Marcus Meissner wrote:
I am not sure that I like this approach. It might appear to our users that we do not care.
I bet we can word the message to avoid that appearance.
We could even have the message simply say
Are you still interested in this bug?
rather than asking them to retest the bug. That would have nearly the same effect, and sounds a lot nicer. I've updated the page to reflect this more friendly wording.
The timelines are a bit strict too. ;)
The message makes it clear that the bug can be reopened at any time. I think that takes away the strictness you're worried about.
How do you feel about it with those changes?
Actually I think a bunch of volunteers has gone over most of the issue manually now...
This lead to more instantely closed bugs ;)
So an automated approach is no longer necessary I guess.
Ciao, Marcus
On 10/5/05, Marcus Meissner meissner@suse.de wrote:
Actually I think a bunch of volunteers has gone over most of the issue manually now...
This lead to more instantely closed bugs ;)
So an automated approach is no longer necessary I guess.
There does seem to be a lot of progress recently...
Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days. Most of these are likely to be obsolete. It would be really easy to write a script to ping the people who submitted those bugs, and ask them to update the bug if it's still valid. If there's no reply in two weeks, we could mark the bug 'resolved abandoned'.
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
For what it's worth, I like the proposal.
As a user, I'd like to be told that it's been 6 months and nobody's gonna look at my bug it if I don't do something proactively myself. IMHO much better to ping users so they can keep their bugs alive if they still exist rather than the bug entries just bitrotting and nobody knowing what state they're in.
Thanks, /me
On Sat, 01 Oct 2005 15:01:33 +0200, Molle Bestefich molle.bestefich@gmail.com wrote:
Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days. Most of these are likely to be obsolete. It would be really easy to write a script to ping the people who submitted those bugs, and ask them to update the bug if it's still valid. If there's no reply in two weeks, we could mark the bug 'resolved abandoned'.
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
from that link:
If you're still interested in this bug, please click on the following link: %url% and add a comment saying you're still interested, giving more information if you have it (e.g. whether it still happens with the latest version of Wine).
This may help to get some fresh input on old bugs and for that it may be of value but I think the basic attitude here is wrong.
It is not a case of whether one user is still "interested" , if there is a bug it is a bug , it needs confirming or refuting (hence unconfirmed , ie never looked at.)
There may be many others who have the same issue but have not reported a bug since a bug is already open. If the original poster has lost interest (after seeing the bug ignored for six months this is very likely) this does not mean others are not interested .
Also once bugs are marked as "resolved" in any sense they certainly wont even get read by anyone likely to really resolve the issue.
There is also the strong likelyhood that after a long delay the original poster may have had to change their email address due to spam flooding or a multitude of other reasons and they will never even get the "ping".
Is there any reason to think that a bug will get more attention if it is refreshed?
"Resolved - abandonned" does not seem any more valid than marking "Resolved - Unconfirmed"
In short, sending a ping to those who can still be contacted may be a good way to get some updated info and presumably some closures by the bug poster but is probably not helpful if it marks unresolved issues as resolved.
;)
wino@piments.com wrote:
This may help to get some fresh input on old bugs and for that it may be of value but I think the basic attitude here is wrong.
The goal of autoresolving is to automate exactly what QA volunteers currently do manually. If you don't like how the QA volunteers are dealing with stale bug reports, feel free to jump in and deal with them differently. Or if you feel the autoresolve proposal doesn't match what the volunteers are doing, please say so.
The autoresolve proposal currently gets two things wrong, I think:
1) it only sends email to the original submitter. It should cc all the people who have asked to be cc'd on the bug, of course.
2) it won't handle any bugs which are stale but which a QA volunteer inquired about recently.
Both of those are probably fairly easy to solve.
But I'm not planning on actually doing an autoresolve run any time soon, since the volunteers seem to be making good headway manually. - Dan
p.s. I do believe I'm getting too cranky to actually interact with other people on this list. Perhaps I should go away until, say, my two-year-old son has learned to sleep through the night! That'll teach me to wait until I'm old and feeble before having my first kid next time :-)
On Fri, 30 Sep 2005, Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days.
[...]
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
Note that the second time you run the query you should only close unconfirmed bugs that are more than 104 days old and unchanged (this was ambiguous in your proposal). That's because unconfirmed bugs that are between 90 and 103 days old have not been notified at that point, or have not had their two weeks to react.
Francois Gouget wrote:
On Fri, 30 Sep 2005, Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days.
[...]
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
Note that the second time you run the query you should only close unconfirmed bugs that are more than 104 days old and unchanged (this was ambiguous in your proposal). That's because unconfirmed bugs that are between 90 and 103 days old have not been notified at that point, or have not had their two weeks to react.
That comment reads a lot like a unit test. :-D
//Jakob
In the same can we close the older bugs dated below 2003, with no activity.
Thanks and regards, Vijay
On 10/1/05, Jakob Eriksson jakov@vmlinux.org wrote:
Francois Gouget wrote:
On Fri, 30 Sep 2005, Dan Kegel wrote:
We have 554 unconfirmed bugs older than 90 days.
[...]
Detailed proposal, including the text of the message to be sent, are at http://kegel.com/wine/qa/autoresolve.html
Note that the second time you run the query you should only close unconfirmed bugs that are more than 104 days old and unchanged (this was ambiguous in your proposal). That's because unconfirmed bugs that are between 90 and 103 days old have not been notified at that point, or have not had their two weeks to react.
That comment reads a lot like a unit test. :-D
//Jakob
Francois Gouget wrote:
Note that the second time you run the query you should only close unconfirmed bugs that are more than 104 days old and unchanged (this was ambiguous in your proposal). That's because unconfirmed bugs that are between 90 and 103 days old have not been notified at that point, or have not had their two weeks to react.
Good point. I've updated the page.
-
Dan Kegel -
Dan Kegel -
Francois Gouget -
Jakob Eriksson -
Marcus Meissner -
Marcus Meissner -
Molle Bestefich -
peter＠piments.com -
Vijay Kiran Kamuju -
wino＠piments.com