> On Fri, 01 Oct 2010, Tollef Fog Heen wrote:
> > Package: huge-dataset
> > Depends: huge-dataset-base-2010-09-01, huge-dataset-patch-2010-09-10,
> > huge-dataset-patch-2010-09-20, huge-dataset-patch-2010-09-30
>
> > And then have huge-dataset set up the data set and patch it as new
> > versions
On Fri, 01 Oct 2010, Tollef Fog Heen wrote:
> | Sounds like you want a dpkg mode where you can pre-depend on a prior
> | version of your own package, and _not_ delete files that are shipped by
> | the base package that are not shipped by the 'delta deb'. Because of
> | this pre-dependency, you wo
On Fri, 01 Oct 2010 at 09:32:02 +0200, Tollef Fog Heen wrote:
> Something like:
>
> Package: huge-dataset
> Depends: huge-dataset-base-2010-09-01, huge-dataset-patch-2010-09-10,
> huge-dataset-patch-2010-09-20, huge-dataset-patch-2010-09-30
>
> And then have huge-dataset set up the data set and
]] Peter Samuelson
| [Yaroslav Halchenko]
| > for HUGE DBs with frequent updates -- may be we could come up with data
| > package chains -- i.e. with some base package version + sequence of
| > dependent packages with binary deltas which get applied upon
| > installation? e.g.
| >
| > data-blob-
[Yaroslav Halchenko]
> for HUGE DBs with frequent updates -- may be we could come up with data
> package chains -- i.e. with some base package version + sequence of
> dependent packages with binary deltas which get applied upon
> installation? e.g.
>
> data-blob-1.0.0
> data-blob-1.0.0+1
> data-b
> Also, I do wonder what the highest update frequency should be for data
> packages.
> Some databases for instance release weekly or even daily snapshots, but
> updating a 10g debian package each week probably is not what we want.
I guess my follow-up is the result of being smoke deprived today:
c
On Thu, Sep 30, 2010 at 06:01:38PM +0100, Simon McVittie wrote:
> On Wed, 29 Sep 2010 at 13:09:03 -0400, Michael Hanke wrote:
> > * compression of content
> >
> > Any policy to enforce compressed file formats regarding the stuff
> > installed
> > by a data package?
>
> I think this ought to
On Wed, 29 Sep 2010 at 13:09:03 -0400, Michael Hanke wrote:
> * compression of content
>
> Any policy to enforce compressed file formats regarding the stuff installed
> by a data package?
I think this ought to be case-by-case: some users of large data blobs
(e.g. Quake 3/Openarena PK3 files)
On Wed, 29 Sep 2010, Michael Hanke wrote:
> I know that data.debian.org is not yet there, but I wonder whether there
> is already a concept, i.e. a set of requirements that packages have to
> fulfil to qualify for this archive?
Also, I do wonder what the highest update frequency should be for dat
Hi,
On Wed, Sep 29, 2010 at 01:09:03PM -0400, Michael Hanke wrote:
> * compression of content
> Any policy to enforce compressed file formats regarding the stuff installed
> by a data package?
Also, how would we deal with data that is to be accessed using a query
engine such as an SQL datab
Hi,
I know that data.debian.org is not yet there, but I wonder whether there
is already a concept, i.e. a set of requirements that packages have to
fulfil to qualify for this archive?
I assume packages would have to be architecture 'all', but other than
that I wonder whether there will be any sig
11 matches
Mail list logo
