Bug#193566: still confused about the root of the problem
On Sun, Jul 06, 2003 at 04:07:00PM -0400, Matt Zimmerman wrote:
> On Fri, Jul 04, 2003 at 03:09:25PM -0700, Ross Boylan wrote:
> > 3) However, the version in unstable would work, if it were in testing,
> > wouldn't it? So the question is why the package hasn't migrated down.
> > The reason appears to have nothing to do with the issues discussed in this
> > thread
> That is incorrect; it is simply that there is an additional issue beyond
> those discussed in this thread.
I'm not completely sure what the "that" refers to in "that is
incorrect." Let me make a few statements, and maybe someone can tell
me if they are accurate:
1) python-apt depends on apt, but not the reverse (for some reason I
had this backward in my mind)
2) Building python-apt from sid source in testing will fail, because
it requires sid apt (this is the one I'm least sure of).
3) Building sid apt in testing is asking for trouble, since sid apt
has RC bugs.
So, getting back to one of my earlier questions (not shown above),
there is no good work-around for testing.
> The primary issue is that apt has
> release-critical bugs which are not being fixed:
> The new python-apt cannot enter testing until the new apt does.
> If you wish to complain about this situation, please contact firstname.lastname@example.org,
> since I assume that is who broke python-apt in testing by forcing in a new
> python (if not, he would know who did).
I might make him aware of it if he doesn't know. I assume it was "the
greatest good for the greatest number." ajt = Anthony Towns?
> > Oops, I realized arm may just be the first reported failure and there
> > could be others. But it does look as if it built OK for i386.
> It built fine everywhere except arm. As far as I can tell from looking at
> that log, this is either a problem with python or with the buildd
> environment. I do not think there is anything wrong with python-apt. The
> fact that no porter has filed a bug regarding this failure makes it seem
> even less likely that python-apt is at fault.
Is there anyone who should be aware of the problem who isn't? It does
certainly look as if it could be some general issue with the build
system (on that platform).
Or will attention just naturally focus on the build problem if it's
still there when the other RC bugs go away?