Re: Bug tracking system

[Nivedita Singhvi <niv@us.ibm.com>]

Ian Pratt wrote:
>>>There was some discussion some time ago about a bug tracking 
>>
>>system for 
>>
>>>Xen.  Is there any plan on when that will be available?
>>
>>I like to vote for the bug tracking system to come sooner 
>>versus later.  I have notice that there are several bugs that 
>>keep coming back up on the mailing list and if we had a bug 
>>track tracking system, we would be able to put it out once.  
>>Then as people have time, they could pick different bugs and 
>>supply fixes for them by assinging the bug to themselves.
> 
> 
> We discussed this a bit at the summit. We have 'trac' set up and ready
> to go, but we weren't planning on announcing it until we fork off
> 3.0-testing around the end of May/ beginning of June. If there was
> strong demand, we'd obviously revist this, but my impression is that the
> mailing lists are working effectively and bugs generally get fixed
> pretty quickly. (Last week and this have been a bit tough because of the
> summit and now Usenix TC)

I'd like to vote for earlier rather than later - because there
seem to be a few things that we are running into, and not everything
is getting posted to list, which is a bit of a concern, especially
the less severe and the once-off stuff. It would be easier to search
thru a buglist rather than mailarchives to find a report of a problem.

I started keeping a list this week:

1. Spurious oops (I just posted this earlier)
2. Dom1 silently hung (only ran a man cmd)
3. Dom1 hung on a large ping send (approx 16700 bytes)
4. raw netperf pkts seem to be not working (possible app bug)
    (orig run into by jonm)
    from domU to dom0
5. xend failed to start (various causes)
6. dom0 network script fails to set up bridge/interfaces
    (originally hit by paull) - this one just needs error handling,
    have fix, not waiting long enough sometimes.
7. xm create -c failed to start up console (last night on 4/12 build
    but it was 2:00AM and I didn't investigate further)
8. fails to boot (kernel config options)

Stopping to repro and investigate is slowing down the 
implementation/testing fronts.


thanks,
Nivedita

6.

> The other issue that was raised was whether we need to go with bugzilla
> to be able to integrate in to distro bug databases. We certainly prefer
> trac, but if we need the ability to assign bugs between databases, I
> guess we'll have to go with bugzilla.
> Any views?
> 
> Thanks,
> Ian
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
> 
>