Idea: review + massive sync of Debian changes not involving new upstream versions?

[Stefan Potyra]

Hi,

Am Thursday 18 March 2010 13:14:06 schrieb Lucas Nussbaum:
> On 18/03/10 at 22:55 +1100, William Grant wrote:
> > On Thu, 2010-03-18 at 12:33 +0100, Reinhard Tartler wrote:
> > > On Thu, Mar 18, 2010 at 11:36:26 (CET), Stefan Potyra wrote:
> > > > Am Wednesday 17 March 2010 19:43:10 schrieb Lucas Nussbaum:
> > > >> Hi,
> > > >>
> > > >> 5 lines summary:
> > > >>   For universe packages that don't have anybody looking at them, I
> > > >>   propose that, instead of letting them rot after Debian Import
> > > >> Freeze, we instead review the Debian changes that don't introduce
> > > >> new upstream versions, and continue to sync them (manually) to
> > > >> Ubuntu to get all the bug fixes that happen in Debian.
> > > >
> > > > While I think that's a noble goal, I fear we lack manpower to do
> > > > this.
> > >
> > > Well, AFAIUI, Lucas suggests that creating an annotated and properly
> > > sorted (by e.g., priority, #bugs, etc.) was not terribly complicated,
> > > given that all that data is already stored in UDD.
> > >
> > > I think such a list would be a great opportunity for fixing low hanging
> > > fruits, if it can be
> > >
> > >  a) easily updated
> > >  b) easily commented by contributors, to quickly triage false positives
> >
> > http://qa.ubuntuwire.org/bugs/rcbugs/lucid/ already lists all the Debian
> > RC bugs that are unfixed here. It has comment and dismissal
> > functionality.
>
> My idea was less focused than just "packages with RC bugs". For example,
> I could quite efficiently review the changes for all packages matching
> '.*ruby.*', and request syncs where applicable. This is likely to bring
> quite a lot of useful fixes (though < RC severity) to Ubuntu.

My argument is mainly that we should focus on bugs with higher severity first. 
My main point there is that I don't think it's easier (at least for me) to 
review a set of particular packages, then to review anything from the  RC bug 
list. However you provided quite a good counter-example, and thinking about 
it, I guess we do have several developers who have expertise on a given set 
of packages, so yes, this seems indeed like a good idea.

>
> But currently:
> - I'm not sure if just using requestsync is appropriate for that, as it
>   will flood archive admins with requests

I think one bug report with the list of packages + versions to be synced might 
prove more efficient? Maybe an archive admin could elaborate?

> - I'm not sure if this is welcomed (seems like it's OK)

Getting any bug fixes in is always good :).

> - We don't have any infrastructure (but that could easily be changed).

Are your plans along the lines that Reinhard suggested? Anything else you have 
in mind?

Cheers,
    Stefan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
Url : https://lists.ubuntu.com/archives/ubuntu-devel/attachments/20100319/943f4fe2/attachment.pgp