Is the rawhide repo broken?

[George R Goffe]

Russel,

Thanks for your response.

Removing python3-docs did not appear to change the problem.

Any other ideas? Just wait for whomever to fix the packaging problems? I could write a bug if you think that might be a good thing to do.

Regards,

George...

Date: Sat, 29 Jul 2017 13:44:23 +0000 (UTC)

From: George R Goffe

Subject: Is the rawhide repo broken?

To: "[e-mail]"

Message-ID:

Content-Type: text/plain; charset=UTF-8

Hi,

I am trying to keep my Fedora 27 (Rawhide) system up to date by running "dnf upgrade" at least daily. For the past week I have seen a list of packages with broken dependencies. 165 in number. dnf --best reports 118 or so "problems".

I'm not sure what to do about this. Could someone give me a hint please?

Regards,

George...

Transaction Summary

==================================================

Skip  165 Packages

Nothing to do.

------------------------------

Date: Sat, 29 Jul 2017 17:52:47 +0100

From: Russel Winder

Subject: Re: Is the rawhide repo broken?

To: [e-mail]

Message-ID:

Content-Type: multipart/signed; micalg="pgp-sha512";

protocol="application/pgp-signature";

boundary="=-A3eF6XdO7hD3qicTS/OP"

--=-A3eF6XdO7hD3qicTS/OP

Content-Type: text/plain; charset="UTF-8"

Content-Transfer-Encoding: quoted-printable

Hi,

I believe there is currently a problem between Boost and LibreOffice

staopping quite a few upgrades but not 165/118, far fewer than that.

I found though I had to remove python3-docs as that package has not

been updated where all the other python3-* packages appear to have

been.

On Sat, 2017-07-29 at 13:44 +0000, George R Goffe wrote:
> Hi,
> =20
> I am trying to keep my Fedora 27 (Rawhide) system up to date by
> running "dnf upgrade" at least daily. For the past week I have seen a
> list of packages with broken dependencies. 165 in number. dnf --best
> reports 118 or so "problems".
> =20
> I'm not sure what to do about this. Could someone give me a hint
> please?
> =20
> Regards,
> =20
> George...
> =20
> =20
> Transaction Summary
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=

=3D
> Skip  165 Packages
> =20
> Nothing to do.
>
> test mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]

--=20

Russel.

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=

=3D=3D

Dr Russel Winder      t: [phone]   voip: sip:russel.winder@ekiga.n=

et

41 Buckmaster Road    m: [phone]   xmpp: [e-mail]

London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

--=-A3eF6XdO7hD3qicTS/OP

Content-Type: application/pgp-signature; name="signature.asc"

Content-Description: This is a digitally signed message part

Content-Transfer-Encoding: 7bit
> -----BEGIN PGP SIGNATURE-----
>
>
> iQIzBAABCgAdFiEETwDs1X+Beyaiaer41L3V0s7Np7gFAll8vV8ACgkQ1L3V0s7N
>
> p7hpVA//bPS3KireZ5K8a6J19obd1caUxWjefoA6xlD+qwxgBjkr0pted4oeEONY
>
> /bfXu18AcqCnnePAZVt+bJhnH3Ac/8A/bYBL+HEMJUoctspqliZyxuVVkpA8maHR
>
> ClNg0M8rBN/eKZdCYIrZPhRkQccR0B5A00e9iqUR6q1JkbpUng5sZtmVPKyZTpeZ
>
> qK9cgQNpGBpTbBUHAklCzCCBcRko3L8LIw00k8w0bRK8GFj0w2341rKm/jvXRWKC
>
> xC4phIBYD1PWuWKgygWFKsXRq7lpMPcXohAZMmCjf7zG2HmyOkq4+PiBr8qaAe+P
>
> ZcG8NFzs+2EN3HY7XqqQizSTWStT9Wgi7+mWv1y1KxLfpDafbotNfJe9cUD1EpeF
>
> +s0cFmlMTQZQdXgsIED0ZM7IK6s6QYDeU/OlvngVsglJo3rV02El1dPkfXjev8G8
>
> Fh3b4bv9/9362/jgPqX3KSj39fJZZTDLGXvrWmlXiwsO7NQcWrI5c5rmbPi4CZvT
>
> hRjWwxoMyC426SAziHW5CxF0gXUqaupxMi9WJiRSwKAbrHTxwGRFDYf0xC0AXYRN
>
> GfNWx0Q8KE7LIATlb0hAUZRfxdXHIF78C1rO70LEewNxFIyRKn/+b5elYX5w6cbk
>
> jj0PXQWSguFVmPHEsircZCe3wb2atQ6Gpq1IbBmA9PiVivd6tPI=
>
> =VS8R
>
> -----END PGP SIGNATURE-----

--=-A3eF6XdO7hD3qicTS/OP--

------------------------------

Subject: Digest Footer

_______________________________________________

test mailing list -- [e-mail]

To unsubscribe send an email to [e-mail]

------------------------------

End of test Digest, Vol 161, Issue 99

*************************************

[Russel Winder]

On Sun, 2017-07-30 at 05:00 +0000, George R Goffe wrote:
> Russel,
>
> Thanks for your response.
>
> Removing python3-docs did not appear to change the problem.

I am surprised you didn't get some change since a number of Fedora
system packages depend on the python3 version.
> Any other ideas? Just wait for whomever to fix the packaging
> problems? I could write a bug if you think that might be a good thing
> to do.

History indicates there are two scenarios and all too often both apply:

1. You have one or two deprecated and/or removed packages that are
causing a version conflict and stopping upgrade.

2. There is a version conflict because of lack of update by packagers.

The python3-doc thing is an instance of 2. As is my current
Boost/LibreOffice situation. I am suffering without python3-doc in the
short term, and waiting on the packagers to sort out the
Boost/LiberOffice thing as I need both.

Whenever I have instances of 1, I use "dnf upgrade --best" and "dnf
upgrade --best --allowerasing" (always saying no to actioning the
upgrade) to investigate where the package conflict is. In many cases
the package is redundant and removing it (or them) clears the blockage.
Sometimes the problem is the dnf system itself, it gets the graph of
dependencies over specifying things and thus causing blockages. In this
case I allow removal of packages to clear the blockage and then
reinstall. Occasionally you have to live without a package or two for a
while to get everything else up to date. cf. python3-doc.

It is a tad difficult to explain things in general. Might it be worth
submitting the result of:

dnf check-update --refresh

dnf upgrade --best

here (as attachments if the output is very lengthy) so as to get more
focused ideas bubbling?

[Matthew Miller]

On Sun, Jul 30, 2017 at 05:00:02AM +0000, George R Goffe wrote:
> Any other ideas? Just wait for whomever to fix the packaging
> problems? I could write a bug if you think that might be a good thing
> to do.

We're in the middle of doing a mass rebuild, and it's not surprising
that rawhide is temporarily broken.

[Peter Robinson]

On Sun, Jul 30, 2017 at 4:03 PM, Matthew Miller
 wrote:
> On Sun, Jul 30, 2017 at 05:00:02AM +0000, George R Goffe wrote:
> > Any other ideas? Just wait for whomever to fix the packaging
> > problems? I could write a bug if you think that might be a good thing
> > to do.
>
> We're in the middle of doing a mass rebuild, and it's not surprising
> that rawhide is temporarily broken.

Actually the mass rebuild happens on a side tag, but the rawhide being
broken is likely due to the ppc64le binutils/glibc mess

[Adam Williamson]

On Mon, 2017-07-31 at 09:16 +0100, Peter Robinson wrote:
> On Sun, Jul 30, 2017 at 4:03 PM, Matthew Miller
>  wrote:
> > On Sun, Jul 30, 2017 at 05:00:02AM +0000, George R Goffe wrote:
> > > Any other ideas? Just wait for whomever to fix the packaging
> > > problems? I could write a bug if you think that might be a good thing
> > > to do.
> >
> > We're in the middle of doing a mass rebuild, and it's not surprising
> > that rawhide is temporarily broken.
>
> Actually the mass rebuild happens on a side tag, but the rawhide being
> broken is likely due to the ppc64le binutils/glibc mess

Well, one specific problem mentioned was actually due to the Boost
version bump. A lot of things depend on Boost, including LibreOffice,
and rebuilding LibreOffice is never much fun (it takes something like
20 hours to build, I believe).

It looks like both the first attempt to rebuild LO for the Boost bump,
and the mass rebuild attempt to build it, failed:

https://koji.fedoraproject.org/koji/buildinfo?buildID=921574
https://koji.fedoraproject.org/koji/buildinfo?buildID=928814

so LO will certainly still have broken deps in current Rawhide, and
that will remain the case until a successful rebuild is done. Other
things may also not yet have been successfully rebuilt against the new
Boost.

The first LO build failed on ppc64le (which may have been due to the
known issue in ppc64le builds that was recently fixed), and s390x
(which failure I can't immediately tell the cause of - it failed
because some file 'does not exist'). The mass rebuild build looks to
have failed on the same arches in much the same way.

[Peter Robinson]

On Mon, Jul 31, 2017 at 6:44 PM, Adam Williamson
 wrote:
> On Mon, 2017-07-31 at 09:16 +0100, Peter Robinson wrote:
> > On Sun, Jul 30, 2017 at 4:03 PM, Matthew Miller
> >  wrote:
> > > On Sun, Jul 30, 2017 at 05:00:02AM +0000, George R Goffe wrote:
> > > > Any other ideas? Just wait for whomever to fix the packaging
> > > > problems? I could write a bug if you think that might be a good thing
> > > > to do.
> > >
> > > We're in the middle of doing a mass rebuild, and it's not surprising
> > > that rawhide is temporarily broken.
> >
> > Actually the mass rebuild happens on a side tag, but the rawhide being
> > broken is likely due to the ppc64le binutils/glibc mess
>
> Well, one specific problem mentioned was actually due to the Boost
> version bump. A lot of things depend on Boost, including LibreOffice,
> and rebuilding LibreOffice is never much fun (it takes something like
> 20 hours to build, I believe).
>
> It looks like both the first attempt to rebuild LO for the Boost bump,
> and the mass rebuild attempt to build it, failed:
>
> https://koji.fedoraproject.org/koji/buildinfo?buildID=921574
> https://koji.fedoraproject.org/koji/buildinfo?buildID=928814
>
> so LO will certainly still have broken deps in current Rawhide, and
> that will remain the case until a successful rebuild is done. Other
> things may also not yet have been successfully rebuilt against the new
> Boost.
>
> The first LO build failed on ppc64le (which may have been due to the
> known issue in ppc64le builds that was recently fixed), and s390x
> (which failure I can't immediately tell the cause of - it failed
> because some file 'does not exist'). The mass rebuild build looks to
> have failed on the same arches in much the same way.

Yes, there's other problems too, like the bump of a graphics library
to fix CVEs, but the accidental miss of the soname bump, there's a
bunch of issues converging unfortunately.