Re: [Python-modules-team] Bug#806334: kombu: FTBFS: ImportError: cannot import name '_uuid_generate_random'
Chris Lamb writes: > from uuid import UUID, uuid4 as _uuid4, _uuid_generate_random > ImportError: cannot import name '_uuid_generate_random' I forwarded this upstream to https://github.com/celery/kombu/issues/543 Looks like Python 3.5.1rc1 dropped a symbol that was present in Python 3.5.0. Don't think we can complain at Python though, looks like a private symbol. CCing debian-python list in case anybody else has more information on why _uuid_generate_random disappeared. -- Brian May
Packaging Python that requires 3.5
Apparently this doesn't work just yet. In debian/control I have X-Python3-Version: >= 3.5 This gives in the Depends: python3.5:any, python3:any (>= 3.5~) This cannot be statisfied because python3 the version in unstable is 3.4.3-7 -- Brian May https://linuxpenguins.xyz/brian/
Bug#806450: RFP: python-inflect -- Python library for correctly generating plurals, singular nouns, ordinals, indefinite articles; convert numbers to words
Package: wnpp Severity: wishlist * Package name: python-inflect Version : 0.2.5 Upstream Author : Paul Dyson * URL : http://pypi.python.org/pypi/inflect * License : AGPLv3 Programming Lang: Python Description : Python library for correctly generating plurals, singular nouns, ordinals, indefinite articles; convert numbers to words This is a dependency of sqlacodegen.
Re: Pushing a bunch of packages to jessie-backports
On 11/27/2015 11:31 AM, Piotr Ożarowski wrote:
>> Will do. FYI, I'm still waiting for the removal of
>> python-repoze.{what,what-plugins,who-plugins}, which is blocking the
>> migration of Keystone, and which therefore makes my efforts useless
>> (there's no point of uploading OpenStack without keystone). So for the
>> moment, my uploads are stalled.
>
> what's blocking you with sphinx?
I don't want to upload it if I can't upload the rest at the same time.
> you realize some packages will have to go though NEW and sphinx is one of
> them?
Yes. Which is why I asked (on the backports list) the backports FTP
masters what they thought was best for the upload.
Though there's no way to make it happen correctly without having some
breakage due to NEW at some point. There's too many interactions and
inter-dependencies.
>> (build-)dependencies are satisfied, yes. Though I don't think I actually
>> need to upload in a specific order: I can push all at once if it is
>> within a single dak run (if I'm wrong here, let me know).
>
> if you upload A that doesn't have to go through NEW and new B that A build
> depends on, A is broken.
> (if all of them will end up in NEW, ftp-masters will take care of that)
What you give above as an example is too simplistic. It works for basic
cases, but not here, with all of what I want to upload, and already
existing packages in jessie-backports.
That's one of the multiple breakage that the NEW queue brings. There's
unfortunately no way to fix that for this kind of complex uploads with
inter-dependencies and updates of existing packages. The only way,
unfortunately, is to have the NEW queue processed fast enough by the FTP
masters.
Re: Pushing a bunch of packages to jessie-backports
> > python-sphinx is not backported and you didn't mention any patches
>
> I did backport python-sphinx too. It is present in these packages, and I
> also mentioned it in this list.
you did mention you will backport sphinx. Yes, I missed that and I'm sorry for
that. I did check if it's backported and it still isn't.
> Will do. FYI, I'm still waiting for the removal of
> python-repoze.{what,what-plugins,who-plugins}, which is blocking the
> migration of Keystone, and which therefore makes my efforts useless
> (there's no point of uploading OpenStack without keystone). So for the
> moment, my uploads are stalled.
what's blocking you with sphinx?
> Unless you want to maintain the backports yourself: no. This is not a
> hijack anyway, you're the only one calling it this way.
I don't have time for it now and I don't want additional work caused by
you (sorry, I still don't trust you, not anymore).
Note that even if it "officially" is somewhere else (or is it not?) -
I will get all the DAK emails and bugs. Ubuntu fixed the fist part by
changing Maintainer to a mailing list. It doesn't solve the bugs issue
in our BTS (which is more annoying).
The easiest way is to simply play nice with current maintainer. I asked
you for changes you want to upload and you didn't provide them until
yesterday (and there's a difference between "just a rebuild" and "just a
rebuild with backported other build dependencies") - I could easily read
that from f.e. build logs or resulting binaries (which you didn't
originally provide).
> On 11/26/2015 09:42 PM, Piotr Ożarowski wrote:
> > Building and uploading in the right order was always the plan,
> > right?
>
> Building in the right order is necessary for checking that
you realize some packages will have to go though NEW and sphinx is one of them?
> (build-)dependencies are satisfied, yes. Though I don't think I actually
> need to upload in a specific order: I can push all at once if it is
> within a single dak run (if I'm wrong here, let me know).
if you upload A that doesn't have to go through NEW and new B that A build
depends on, A is broken.
(if all of them will end up in NEW, ftp-masters will take care of that)
--
Piotr Ożarowski Debian GNU/Linux Developer
www.ozarowski.pl www.griffith.cc www.debian.org
GPG Fingerprint: 1D2F A898 58DA AF62 1786 2DF7 AEF6 F1A2 A745 7645
Re: Pushing a bunch of packages to jessie-backports
On 11/26/2015 04:02 PM, Piotr Ożarowski wrote:
> that's why I don't believe you built it in jessie
I did. I'm running with sbuild on Jessie + Jessie backports, and of
course, I'm also managing a local repository with the packages already
rebuilt for jessie-backports, which will be part of the upload.
Now, feel free to not believe me, I don't mind. But please don't write
public accusations, when you don't actually know what I'm doing, and
haven't checked your facts.
> python-sphinx is not backported and you didn't mention any patches
I did backport python-sphinx too. It is present in these packages, and I
also mentioned it in this list.
> solutions:
> 1) backport python-sphinx
Which is what I did. Routes is by the way not the only package that
needed this build dependency.
> After upload, please subscribe routes' bugs.
Will do. FYI, I'm still waiting for the removal of
python-repoze.{what,what-plugins,who-plugins}, which is blocking the
migration of Keystone, and which therefore makes my efforts useless
(there's no point of uploading OpenStack without keystone). So for the
moment, my uploads are stalled.
On 11/26/2015 09:33 PM, Piotr Ożarowski wrote:
> I actually can tell zigo to not hijack my packages?
Unless you want to maintain the backports yourself: no. This is not a
hijack anyway, you're the only one calling it this way.
On 11/26/2015 09:42 PM, Piotr Ożarowski wrote:
> Building and uploading in the right order was always the plan,
> right?
Building in the right order is necessary for checking that
(build-)dependencies are satisfied, yes. Though I don't think I actually
need to upload in a specific order: I can push all at once if it is
within a single dak run (if I'm wrong here, let me know).
On 11/26/2015 03:02 PM, Piotr Ożarowski wrote:
> If he decides to upload broken package, I cannot do anything.
Seriously, change your attitude.
On 11/26/2015 03:02 PM, Piotr Ożarowski wrote:
> and even if he's able to generate
> debdiff (out of broken packages) then there's something wrong with his
> CI/sbuild setup and we'll figure that out. He will also subscribe bugs
> from all these packages just to shut up this asshole (and that's a
> good outcome, even if I had to be rude; being nice doesn't work with
> him, I tried that for few years, he just ignores you)
And again. I'm getting really tired of these gratuitous accusations even
before anything happens.
Cheers,
Thomas Goirand (zigo)
Re: Pushing a bunch of packages to jessie-backports
❦ 26 novembre 2015 21:33 +0100, Piotr Ożarowski : >> No, that's not necessary as long as the orig tarball is already >> present in the archive in another suite (which would be true for >> backports, since the package is supposed to be derived from the >> current version in testing). > > any pointers to docs/announcements? That's news to me... https://lists.debian.org/debian-devel-announce/2013/03/msg7.html -- There is a great discovery still to be made in Literature: that of paying literary men by the quantity they do NOT write. signature.asc Description: PGP signature
