Filing Bugs in Kubuntu

Sat Jan 23 17:35:37 GMT 2010

On Sat, Jan 23, 2010 at 02:43:43PM +0100, Harald Sitter wrote:
[...]
> KDE is not one team doing all and everything, but loads of smaller teams doing 
> all parts of everything, making it more efficient no matter how you look at it.

I can agree with you on this point.

> Anyhow. Yes. This is part of the reasoning [1]. Kubuntu watches over some 50 
> source packages' bugs, making up for probably > 150 applications, and when I 
> say Kubuntu in that context I mean Jonathan Thomas, because he is about the 
> only person doing bug triage at the rate necessary to prevent all reports for 
> those 150 apps from rotting away.

I could totally accept it 100% if...Of course there is a if...Or maybe it
should be a but...Anyways, when was the last Kubuntu Bug Hug Day? I can't
remember to be honest. I know we did a couple a few release back in unison
with the regular Ubuntu Bug Hug Day. It seems that Jonathan was doing his
own bug days, and that's just one part of the reason why he rocks. I know
triaging bugs suck, but it is part of every developer community their is.
We should have been holding the hug days to not only triage our bugs, but
also attempt to get more contributors this way. Ubuntu does their hug day,
and they always get a bunch of new contributors. If you watch the BugSquad
membership at all, you will see it growing pretty much every day.

> Of course KDE does not have unlimited resources. But let's assume a user finds 
> a crash in kwrite, an app we do not ship by default, so the need of instant 
> triage is lower than usual, yet it gets reported against kdebase.
> At this point it all depends on whether Jonathan or someone else decides that 
> it is triage worthy or not. If they do not (which is probably the case since 
> kwrite is not part of the default install), then the report does
> a) clot the kdebase reports
> b) prevent upstream from fixing a bug
> c) appears as if Kubuntu did not care

I understand the aggravation that (a) causes, and in a lot of cases (b) is
already known upstream when it is filed in LP. As for (c), I would kind of
think we may not have cared. I know we do, but we never attempted to make
it seem like we cared.

> This is the case up to the point where a triage squad comes by and happens to 
> triage all of kdebase so the bug can move along, by which time it will either 
> be fixed or already reported upstream anyway.

Correct.

> The ratio of Kubuntu-only bugs is so horribly low that it simply does not 
> justify letting hundreds of bugs rot in malone, which from a user perspecitve 
> does not put good light on Kubuntu nor KDE.

Well, with the "Report a bug" feature now going to bugs.kde.org, how is a
new user to know if the crash is related to Kubuntu or not? What I am kind
of afraid of is this, a crash occurs, gets filed in b.k.o, it is because of
either packaging or a booged patch of ours/Debians. Now we have to rely on
upstream to triage it, in which case they say "This is Kubuntu, not us". So
now what does the new user do? More than likely, they may not be hip on
filing a bug in LP. We can document the process until our face turns blue,
hell we could even make it the default wallpaper. In a lot of cases, it is
a topic that will not even be read.

> Ultimately all reports would go to launchpad, but that implies that we have 
> the resources to pipe them through or get them triaged in a sensible time 
> frame, which is not the case.

If we would do Hug Days, then I think we could find the resources. KDE has
been doing amazing with their Hug Days, Ubuntu, openSUSE and Fedora as
well, who both have just as small of a team as we do.

> So we can continue arguing about how it should be done, but unless we all get 
> a triage rate to compete with Jonathan's, I do not think that we have really 
> any say.

See, I feel we took the easy way out, and I feel it may not be the correct
way. We never attempted to hold a triage day at all. I am definitely not
blaming anyone for not attempting this, as I am just as guilty if not more.
I know I could have taken the initiative to do so, and I wish I would have.
I have no problems with bugs going upstream at all. I kind of worry about
the instances where DrKonqi doesn't get enough information and doesn't
allow one to file a bug report. This happens plenty of times, especially
when one doesn't have the necessary packages installed. I just made
KEuroCalc crash, which isn't in KDE. Yet the KDE Crash Handler comes up,
provides you the "Report a bug" feature. If you go through the wizard, in
the end it doesn't allow you file a bug report. Of course because KEuroCalc
isn't in KDE. Now what does the user do? If the user isn't aware of filing
the bug report in LP instead, then this is a bug that doesn't get reported.
Kind of like the popups on Windows where you can send a report. People just
got used to hitting cancel, and I am afraid we could end up being like that
as well.

So with all of that rambling, I am still not convinced, even with the use
cases and everything reported in the triage spec.

-- 
 Name|  Richard JOHNSON
Title|  Developer
  WWW|  http://www.ubuntu.com
Email|  nixternal at ubuntu.com
GnuPG|  3578 0981 A21D D662 2A96  7623 F4C1 838C D8C4 4738
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: Digital signature
Url : https://lists.ubuntu.com/archives/kubuntu-devel/attachments/20100123/99c233d7/attachment.pgp 