Re: Linux 2.0.36 in slink?
On Sun, Dec 27, 1998 at 12:13:46AM -0800, Joseph Carter wrote:
> On Sat, Dec 26, 1998 at 02:15:13PM -0800, Oscar Levi wrote:
> > It is clear that we don't understand each other. Let's take another
> > tack and identify what problems we are trying to solve.
> > 1) Just before release, it takes a long time for package updates to
> > make their way into the mirrored archives so that we can test
> > them. This is due to the appropriate caution of the release
> > manager(s).
> Yes. The proposal eliminates this by causing nothing to be uploaded to
> pre-release (now called frozen and not always actually even there)
> Packages are available the very next day after dinstall does its thing or
> for new packages after the archive managers (not the release manager as
> now, difference being there are half a dozen archive managers) make sure
> the packages dinstall does not know about are checked real fast and
> added as now---updates to unstable happen very fast because dinstall does
> most things and what dinstall doesn't the archive managers do quickly.
I've lost track of the naming scheme. Aside from that, it sounds like
the right answer.
> > 2) Performing incremental updates to a running system is desirable
> > for the sake of testing. Packages will be upgraded and
> > downgraded in order to isolate problems or return a broken system
> > to a working state.
[deletia explaining implementaion]
> pre-release _IS_ th "accepted" tree. It replaces frozen.
As I said, if it solves these problems, then it makes sense. My
critique has been based on the need for testability. Our decision to
move packages from one place to another does not make sense if there
is no way to upgrade/downgrade. Until apt has distribution tags and
can cope with a downgrade, I think a reorg provides only limited
improvement over the current setup.