[opensuse-factory] Friendly warning: Python 2 is going to be removed from Factory on 2020-01-02
Hello, this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory. You should really really make sure to check all your tools have been already ported and tested on Python 3 right NOW. Best, Matěj -- https://matej.ceplovi.cz/blog/, Jabber: mcepl@ceplovi.cz GPG Finger: 3C76 A027 CA45 AD70 98B5 BC1D 7920 5802 880B C9D8 Less is more or less more. -- Y_Plentyn on #LinuxGER (from fortunes -- I cannot resist :-)
On 4/15/19 1:39 PM, Matěj Cepl wrote:
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
I really wonder why this is necessary. Given the fact that there are enterprise distros maintaining Python 2 for a much longer time there will be backport patches available for upcoming security issues. And I'm willing to take care of that. Ciao, Michael. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Michael Ströder píše v Po 15. 04. 2019 v 13:56 +0200:
I really wonder why this is necessary. Given the fact that there are enterprise distros maintaining Python 2 for a much longer time there will be backport patches available for upcoming security issues. And I'm willing to take care of that.
You are free to keep your own project on OBS with packages you want to support (or use Leap), but Factory is for “the newest stable software” and “bleeding edge software”, not the museum of computer archeology. Best, Matěj -- https://matej.ceplovi.cz/blog/, Jabber: mcepl@ceplovi.cz GPG Finger: 3C76 A027 CA45 AD70 98B5 BC1D 7920 5802 880B C9D8 Thou shalt not nede to be afrayed for any bugges by night. -- Coverdale's 1535 translation of Psalm 91 (or Christopher Higley's 1972 thesis explaining why programmers' are nocturnal beasts)
On 4/15/19 2:05 PM, Matěj Cepl wrote:
Michael Ströder píše v Po 15. 04. 2019 v 13:56 +0200:
I really wonder why this is necessary. Given the fact that there are enterprise distros maintaining Python 2 for a much longer time there will be backport patches available for upcoming security issues. And I'm willing to take care of that.
You are free to keep your own project on OBS with packages you want to support (or use Leap), but Factory is for “the newest stable software” and “bleeding edge software”, not the museum of computer archeology.
So this is simply your personal ego-trip not to care about what others need. Ciao, Michael. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, 15 Apr 2019 at 14:13, Michael Ströder <michael@stroeder.com> wrote:
On 4/15/19 2:05 PM, Matěj Cepl wrote:
Michael Ströder píše v Po 15. 04. 2019 v 13:56 +0200:
I really wonder why this is necessary. Given the fact that there are enterprise distros maintaining Python 2 for a much longer time there will be backport patches available for upcoming security issues. And I'm willing to take care of that.
You are free to keep your own project on OBS with packages you want to support (or use Leap), but Factory is for “the newest stable software” and “bleeding edge software”, not the museum of computer archeology.
So this is simply your personal ego-trip not to care about what others need.
Users want supported, secure, maintained software https://pythonclock.org/ That does not include Python 2 after the day upstream formally, absolutely, and unquestionably abandon it. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 2:18 PM, Richard Brown wrote:
On Mon, 15 Apr 2019 at 14:13, Michael Ströder <michael@stroeder.com> wrote:
On 4/15/19 2:05 PM, Matěj Cepl wrote:
Michael Ströder píše v Po 15. 04. 2019 v 13:56 +0200:
I really wonder why this is necessary. Given the fact that there are enterprise distros maintaining Python 2 for a much longer time there will be backport patches available for upcoming security issues. And I'm willing to take care of that.
You are free to keep your own project on OBS with packages you want to support (or use Leap), but Factory is for “the newest stable software” and “bleeding edge software”, not the museum of computer archeology.
So this is simply your personal ego-trip not to care about what others need.
Users want supported, secure, maintained software
That does not include Python 2 after the day upstream formally, absolutely, and unquestionably abandon it.
As said. There are various installation for which back-port patches will be available, including SLE. While migrating to Python 3 is of course necessary I see no reason to fighting a crusade against people who still have to maintain Python 2 after that date. And as said: I'm willing to do work. Ciao, Michael. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, Apr 15, Michael Ströder wrote:
As said. There are various installation for which back-port patches will be available, including SLE.
SLES15 SP1 is already python3 only, all python2 code moved to a legacy module (which means limited support and a short lifecycle) which will be removed mid-term. As announced already two years ago, or even longer. Beside, how should SLE contain backports, if there isn't anything upstream anymore? You can only backport something if there is somebody active upstream. I wouldn't count on that. Thorsten -- Thorsten Kukuk, Distinguished Engineer, Senior Architect SLES & MicroOS SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nuernberg, Germany GF: Felix Imendoerffer, Mary Higgins, Sri Rasiah, HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 2:32 PM, Thorsten Kukuk wrote:
On Mon, Apr 15, Michael Ströder wrote:
As said. There are various installation for which back-port patches will be available, including SLE.
SLES15 SP1 is already python3 only, all python2 code moved to a legacy module (which means limited support and a short lifecycle) which will be removed mid-term. As announced already two years ago, or even longer.
So the python2 interpreter is removed too? I was under impression the interpreter remains for duration of SLE15.
Beside, how should SLE contain backports, if there isn't anything upstream anymore? You can only backport something if there is somebody active upstream.
I'm not sure. We distribute a lot of software that has no upstream anymore. For all intents and purposes, we are upstream for a lot of "enterprise" software. I'm quite certain that python2 will live on as some other name. - Adam -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, Apr 15, Adam Majer wrote:
On 4/15/19 2:32 PM, Thorsten Kukuk wrote:
On Mon, Apr 15, Michael Ströder wrote:
As said. There are various installation for which back-port patches will be available, including SLE.
SLES15 SP1 is already python3 only, all python2 code moved to a legacy module (which means limited support and a short lifecycle) which will be removed mid-term. As announced already two years ago, or even longer.
So the python2 interpreter is removed too?
Yes, that one has moved, too.
I was under impression the interpreter remains for duration of SLE15.
No, we always clearly stated, that this will not be the case. Thorsten -- Thorsten Kukuk, Distinguished Engineer, Senior Architect SLES & MicroOS SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nuernberg, Germany GF: Felix Imendoerffer, Mary Higgins, Sri Rasiah, HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 8:43 AM, Thorsten Kukuk wrote:
On Mon, Apr 15, Adam Majer wrote:
On 4/15/19 2:32 PM, Thorsten Kukuk wrote:
On Mon, Apr 15, Michael Ströder wrote:
As said. There are various installation for which back-port patches will be available, including SLE.
SLES15 SP1 is already python3 only, all python2 code moved to a legacy module (which means limited support and a short lifecycle) which will be removed mid-term. As announced already two years ago, or even longer.
So the python2 interpreter is removed too?
Yes, that one has moved, too.
I was under impression the interpreter remains for duration of SLE15.
No, we always clearly stated, that this will not be the case.
Well SLE15 is not the only game in town, we should not forget that. SLE12 has Python 2.7 and has a lifecycle until Oct 31st 2027, add 3 years of LTSS and Python 2.7 will exist in SLE 12 10 years past the upstream EOL. Unless of course SLE decides to break pretty much every rule we have and ditch Python 2.7. Not that I am advocating to keep Python 2.7 in Factory until EOL of SLE 12. However the argument about "it goes away upstream" is not very strong IMHO. Later, Robert -- Robert Schweikert MAY THE SOURCE BE WITH YOU Distinguished Architect LINUX Technical Team Lead Public Cloud rjschwei@suse.com IRC: robjo
On Mon, 15 Apr 2019 at 15:20, Robert Schweikert <rjschwei@suse.com> wrote:
No, we always clearly stated, that this will not be the case.
Well SLE15 is not the only game in town, we should not forget that. SLE12 has Python 2.7 and has a lifecycle until Oct 31st 2027, add 3 years of LTSS and Python 2.7 will exist in SLE 12 10 years past the upstream EOL.
Unless of course SLE decides to break pretty much every rule we have and ditch Python 2.7.
Not that I am advocating to keep Python 2.7 in Factory until EOL of SLE 12. However the argument about "it goes away upstream" is not very strong IMHO.
But considering that, unlike SLE 15, SLE 12 and Tumbleweed have had no practical, theoretical, or even tangential codebase relationship for years now, and Leap's last SLE 12-based release reaches end of life in June this year, I don't see how your point really bring anything to the table. I do not think anyone is suggesting that SUSE 'forward-port' to openSUSE any of their maintenance & fixes they will be conducting on their old SLE codebase. Meanwhile, the developers of openSUSE need to assess what they are able and willing to maintain themselves. That determination appears to have been made, the reasoning is sound, the intent has been made public. I think energies are best spent figuring out mitigation strategies for any disruption this may cause, but at it's heart, any suggestion that Python2 has any life left in it appears to be a lost cause to me. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 9:28 AM, Richard Brown wrote:
On Mon, 15 Apr 2019 at 15:20, Robert Schweikert <rjschwei@suse.com> wrote:
No, we always clearly stated, that this will not be the case.
Well SLE15 is not the only game in town, we should not forget that. SLE12 has Python 2.7 and has a lifecycle until Oct 31st 2027, add 3 years of LTSS and Python 2.7 will exist in SLE 12 10 years past the upstream EOL.
Unless of course SLE decides to break pretty much every rule we have and ditch Python 2.7.
Not that I am advocating to keep Python 2.7 in Factory until EOL of SLE 12. However the argument about "it goes away upstream" is not very strong IMHO.
But considering that, unlike SLE 15, SLE 12 and Tumbleweed have had no practical, theoretical, or even tangential codebase relationship for years now, and Leap's last SLE 12-based release reaches end of life in June this year, I don't see how your point really bring anything to the table.
Well you stated that: "Users want supported, secure, maintained software" In a context where Michael offered to do the necessary maintenance work for Python 2. Maybe you are implying that Michael is not capable of delivering according to your definition? Further, based on your statement one could easily draw the conclusion that you are either deliberately, inadvertently, or uncaringly implying that SLE will be shipping unsupported insecure software. This ML does not exist in a vacuum and everyone is well aware of the relationship between openSUSE and SLE. Therefore, IMHO, such statements probably have funny and strange ways of making people think things that may or may not have been intended. To this you will of course state that this is an openSUSE mailing list and what SLE does has no bearing to the discussion. I will then reply that SLE simply becomes the new upstream of Python 2.7 and that yes, what SLE does has bearing on openSUSE; and round and round we go. So lets be clear, SLE 12 will have a supported Python 2.7 version until 2030, at least as long as current guidelines and policies apply. This also implies that there will be fixes to Python 2.7 in case there are security issues. Given that we have someone volunteering to take such fixes and put them back into factory I fail to see why the institution of a grace period would be such a horrible thing as it appears it is made out to be. The next logical argument from those that can't wait will be that there has been a 10 year or so grace period already and that waiting any longer will not make a difference. Certainly that is a fair argument and one that I have made as well on occasion. For better or worse people have a funny way of deciding for themselves when the time is right, may it be finishing the osc port or the Firefox build system or "my_favorite_application". And yes it is a matter of tradeoffs, are we more interested in having only Python 3 or are we interested in having Firefox? Or maybe someone here will step up and fix up the Firefox build system, I don't know. But the primary Firefox maintainer already stated that it appears at this point unlikely that Firefox will be ready for days where Python 2.7 does not exist. Anyway the post that started all of this did not appear to consider such tradeoffs, and maybe that was done in previous posts, I do not know. The point is, collectively we find better solutions. The process to get there is made more difficult when lines are drawn in the sand.
I do not think anyone is suggesting that SUSE 'forward-port' to openSUSE any of their maintenance & fixes they will be conducting on their old SLE codebase.
Meanwhile, the developers of openSUSE need to assess what they are able and willing to maintain themselves.
So are you saying that Michael Stroeder is not considered part of this "developers of openSUSE" group? His comments were simply dismissed out of hand, that does not appear to be an assessment to me.
That determination appears to have been made, the reasoning is sound, the intent has been made public. I think energies are best spent figuring out mitigation strategies for any disruption this may cause, but at it's heart, any suggestion that Python2 has any life left in it appears to be a lost cause to me.
Well it has life left as long as it is needed by "my_favorite_application" and if Firefox no longer builds after 2020-01-02 then I guess for all those that consider Firefox as "my_favorite_application" it's lights out. That is certainly a path that can be chosen. Whether that is a good decision or not is up to others to decide. Later, Robert -- Robert Schweikert MAY THE SOURCE BE WITH YOU Distinguished Architect LINUX Technical Team Lead Public Cloud rjschwei@suse.com IRC: robjo
Michael Ströder píše v Po 15. 04. 2019 v 14:13 +0200:
So this is simply your personal ego-trip not to care about what others need.
Personal attack without any relevant argument to the issue -> *plonk* -- https://matej.ceplovi.cz/blog/, Jabber: mcepl@ceplovi.cz GPG Finger: 3C76 A027 CA45 AD70 98B5 BC1D 7920 5802 880B C9D8 A man once asked Mozart how to write a symphony. Mozart told him to study at the conservatory for six or eight years, then apprentice with a composer for four or five more years, then begin writing a few sonatas, pieces for string quartets, piano concertos, etc. and in another four or five years he would be ready to try a full symphony. The man said, "But Mozart, didn't you write a symphony at age eight?" Mozart replied, "Yes, but I didn't have to ask how." -- ripped from another sig
Hello, Am Montag, 15. April 2019, 14:26:59 CEST schrieb Matěj Cepl:
Michael Ströder píše v Po 15. 04. 2019 v 14:13 +0200:
So this is simply your personal ego-trip not to care about what others need.
Personal attack without any relevant argument to the issue -> *plonk*
You are of course free to block whoever you want, but if I look at the texts from the previous mails, I'd say Michael's previous answer was more technical than yours ;-) Let me bring that text back:
On 4/15/19 2:05 PM, Matěj Cepl wrote:
Michael Ströder píše v Po 15. 04. 2019 v 13:56 +0200:
I really wonder why this is necessary. Given the fact that there are enterprise distros maintaining Python 2 for a much longer time there will be backport patches available for upcoming security issues. And I'm willing to take care of that.
So there are patches (even if I feel with the poor SLE 12 maintainers) and someone is willing to do the work for openSUSE. What else do we need? ;-)
You are free to keep your own project on OBS with packages you want to support (or use Leap), but Factory is for “the newest stable software” and “bleeding edge software”, not the museum of computer archeology.
That sounds like "do whatever you want, but I'll forbid you to keep python2 in Factory" [1] - and the "museum of computer archeology" also doesn't sound like a technical argument to me. Hey, with that argument, we'd have to remove all those command-line dinosaurs like vim, emacs and mutt because nowadays everybody wants GUI programs anyway! That said: I'm not surprised that Michael answered in the way he did ;-) I have to agree that your "argument" sounds more like a personal opinion and not like a technical argument, therefore you shouldn't be too surprised about the (not really diplomatic, but IMHO correct) summary you got. And finally - in case you don't like my answer, feel free to block me too ;-) Regards, Christian Boltz [1] Last time I checked our rules, the main reason to forbid a package in Tumbleweed were legal reasons, and I'm sure that won't apply to python2 ;-) -- I preferred to do it here so that everybody can think about it, refuse it, burn it, stone me and condemn me for sorcery or be cheerful and happy (the unlikely case). [takanov AT freenet.de in opensuse-project]
Am 15.04.19 um 13:39 schrieb Matěj Cepl:
Hello,
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
You should really really make sure to check all your tools have been already ported and tested on Python 3 right NOW.
I do not think that the Mozilla buildsystem (speaking of Firefox, Thunderbird and seamonkey at least) is Python3-only ready yet. The likeliness that this happens in time before 2020 is rather low. So this is just a friendly warning that I will not and also cannot make sure that Firefox et al will build without Python 2.7 in time. I will keep trying it though. In worst case I hope that we still can have it for building without publishing in the "normal distro"? Contributors are free to help resolving all depending issues from https://bugzilla.mozilla.org/show_bug.cgi?id=1388447 Wolfgang -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 1:39 PM, Matěj Cepl wrote:
You should really really make sure to check all your tools have been already ported and tested on Python 3 right NOW.
Hopefully node-gyp will be finally be ported to Python 3 as otherwise no NodeJS in Factory because of this. The problem has to do with chromium/v8/node build system that affects sections of the npm ecosystem - who wants to touch that can of worms? https://github.com/nodejs/node-gyp/issues/1337 the issue has been going since at least 2013, https://github.com/nodejs/node-gyp/issues/193 If there is a lesson here - "don't invent your own build system!" - Adam -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Monday 2019-04-15 14:04, Adam Majer wrote:
On 4/15/19 1:39 PM, Matěj Cepl wrote:
You should really really make sure to check all your tools have been already ported and tested on Python 3 right NOW.
Hopefully node-gyp will be finally be ported to Python 3 [...] the issue has been going since at least 2013,
https://github.com/nodejs/node-gyp/issues/193
If there is a lesson here - "don't invent your own build system!"
"People who don't understand $x are bound to reinvent $x, poorly." ;-) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hello, On Mon, 2019-04-15 at 13:39 +0200, Matěj Cepl wrote:
Hello,
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
I'm not sure if you cna word it this drastic - you can set it as a GOAL to remove python2 from Factory. Even a target date is a nice to have. And a lot of effort has already happened. But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?! So, please, let's tone this thread down a notch: Maej formulated a goal to remove python2 by 2020 - and as a distribution I hope we can all see the benefit of not having to maintain something that is abandoned upstream. But just jumping off the cliff because upstream declares something unmaintained is not going to happen.
You should really really make sure to check all your tools have been already ported and tested on Python 3 right NOW.
This is something I can undersign though! Please help to make the goal reachable as much as you can (a lot of effort has come from SLE15 already) Cheers, Dominique
On Mon, Apr 15, 2019 at 8:35 AM Dominique Leuenberger / DimStar <dimstar@opensuse.org> wrote:
Hello,
On Mon, 2019-04-15 at 13:39 +0200, Matěj Cepl wrote:
Hello,
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
I'm not sure if you cna word it this drastic - you can set it as a GOAL to remove python2 from Factory. Even a target date is a nice to have. And a lot of effort has already happened.
But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?!
At least on the subject of osc, the PRs to fix it for Python 3 have stalled out. Nobody with write access is reviewing any of them. At least in Fedora, due to the pressure of getting the dep chains for Python 2 reduced[1], I imported the patches and switched osc to Python 3 in Fedora 30[2]. It works great. Too bad the upstream project maintainer seems to not have time for the project... So I hope Matěj's post pushes people to prioritize getting the ports done, if nothing else. [1]: https://fedoraproject.org/wiki/Changes/Mass_Python_2_Package_Removal [2]: https://src.fedoraproject.org/rpms/osc/c/31982d5946fcd3b916bc77dc1708017bd80... -- 真実はいつも一つ！/ Always, there's only one truth! -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On pon, Apr 15, 2019 at 2:40 PM, Neal Gompa <ngompa13@gmail.com> wrote:
On Mon, Apr 15, 2019 at 8:35 AM Dominique Leuenberger / DimStar <dimstar@opensuse.org> wrote:
Hello,
On Mon, 2019-04-15 at 13:39 +0200, Matěj Cepl wrote:
Hello,
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
I'm not sure if you cna word it this drastic - you can set it as a GOAL to remove python2 from Factory. Even a target date is a nice to have. And a lot of effort has already happened.
But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?!
At least on the subject of osc, the PRs to fix it for Python 3 have stalled out. Nobody with write access is reviewing any of them. At least in Fedora, due to the pressure of getting the dep chains for Python 2 reduced[1], I imported the patches and switched osc to Python 3 in Fedora 30[2]. It works great. Too bad the upstream project maintainer seems to not have time for the project...
Eh, who needs the primary tool to develop openSUSE on developer focused openSUSE distribution? LCP [Stasiek] https://lcp.world -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 2:34 PM, Dominique Leuenberger / DimStar wrote:
But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?!
So, please, let's tone this thread down a notch: Maej formulated a goal to remove python2 by 2020 - and as a distribution I hope we can all see the benefit of not having to maintain something that is abandoned upstream. But just jumping off the cliff because upstream declares something unmaintained is not going to happen.
Let me describe my personal situation: 1. I'm supporting Linux distros like CentOS7 where the standard distro does not ship with Python3 yet. So there is some constraint there. 2. I do want to migrate to Python3. But when touching all the code I'd like make sure to really move to Python3 (especially with type hints). So I won't maintain modules compatible to both 2 and 3 to avoid all the ugly code mess you can find in various module packages. But with a Python maintainer moving stuff around, breaking existing config workflows I'm forced to implement many work-arounds wasting my time. I'd really prefer to spend this time on migrating the upstream code base! So Matěj's pushing towards Python3 effectively slows down Python3 progress on my side. And that's what really makes me angry. Ciao, Michael.
On Mon, Apr 15, 2019 at 9:25 AM Michael Ströder <michael@stroeder.com> wrote:
On 4/15/19 2:34 PM, Dominique Leuenberger / DimStar wrote:
But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?!
So, please, let's tone this thread down a notch: Maej formulated a goal to remove python2 by 2020 - and as a distribution I hope we can all see the benefit of not having to maintain something that is abandoned upstream. But just jumping off the cliff because upstream declares something unmaintained is not going to happen.
Let me describe my personal situation:
1. I'm supporting Linux distros like CentOS7 where the standard distro does not ship with Python3 yet. So there is some constraint there.
Python 3.6 is coming to the base EL7 distribution with the 7.7 release: https://bugzilla.redhat.com/show_bug.cgi?id=1639030 Prior to this, Python 3 has been available through Software Collections in both EL6 and EL7. Python 3 is also available in EPEL for EL6 and EL7. So I don't consider this as a valid excuse, as there are ways to have a usable Python 3 environment.
2. I do want to migrate to Python3. But when touching all the code I'd like make sure to really move to Python3 (especially with type hints). So I won't maintain modules compatible to both 2 and 3 to avoid all the ugly code mess you can find in various module packages.
But with a Python maintainer moving stuff around, breaking existing config workflows I'm forced to implement many work-arounds wasting my time. I'd really prefer to spend this time on migrating the upstream code base!
Use a Tumbleweed snapshot for that. It's an effective strategy if you consider Factory churn to hurt your development process. But frankly, when I worked on getting Pagure to Python 3, I worked from Tumbleweed almost exclusively. It worked out well, even with the churn, because I could keep validating against new stuff. -- 真実はいつも一つ！/ Always, there's only one truth! -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 3:44 PM, Neal Gompa wrote:
On Mon, Apr 15, 2019 at 9:25 AM Michael Ströder <michael@stroeder.com> wrote:
On 4/15/19 2:34 PM, Dominique Leuenberger / DimStar wrote:
But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?!
So, please, let's tone this thread down a notch: Maej formulated a goal to remove python2 by 2020 - and as a distribution I hope we can all see the benefit of not having to maintain something that is abandoned upstream. But just jumping off the cliff because upstream declares something unmaintained is not going to happen.
Let me describe my personal situation:
1. I'm supporting Linux distros like CentOS7 where the standard distro does not ship with Python3 yet. So there is some constraint there.
Python 3.6 is coming to the base EL7 distribution with the 7.7 release: https://bugzilla.redhat.com/show_bug.cgi?id=1639030
Prior to this, Python 3 has been available through Software Collections in both EL6 and EL7.
Python 3 is also available in EPEL for EL6 and EL7.
Extra repos cannot be always used in every environment. I'm pretty sure you know such installations.
So I don't consider this as a valid excuse, as there are ways to have a usable Python 3 environment.
This is not meant as general "excuse" to not work on the migration. It's an explanation that I have to keep things running in the mean-time.
2. I do want to migrate to Python3. But when touching all the code I'd like make sure to really move to Python3 (especially with type hints). So I won't maintain modules compatible to both 2 and 3 to avoid all the ugly code mess you can find in various module packages.
But with a Python maintainer moving stuff around, breaking existing config workflows I'm forced to implement many work-arounds wasting my time. I'd really prefer to spend this time on migrating the upstream code base!
Use a Tumbleweed snapshot for that. It's an effective strategy if you consider Factory churn to hurt your development process. But frankly, when I worked on getting Pagure to Python 3, I worked from Tumbleweed almost exclusively. It worked out well, even with the churn, because I could keep validating against new stuff.
Be assured I know how to use Tumbleweed - not only for development. Ciao, Michael. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 4/15/19 8:34 AM, Dominique Leuenberger / DimStar wrote:
Hello,
On Mon, 2019-04-15 at 13:39 +0200, Matěj Cepl wrote:
Hello,
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
I'm not sure if you cna word it this drastic - you can set it as a GOAL to remove python2 from Factory. Even a target date is a nice to have. And a lot of effort has already happened.
But, as a release manager I claim to have a say in this topic too: Dropping py2 is a noble goal, it is a high target. But if this means losing things like firefox, osc, "insertfavoriteapphere" - then you all can be sure that we will weigh the pros and cons of keeping/dropping py2. In plus, with a capable maintainer who is willing to further work on possibly reported CVEs: how can we claim it is not maintained?!
So, please, let's tone this thread down a notch: Maej formulated a goal to remove python2 by 2020
Well, from my perspective formulating a goal would be more along the lines of: "...the goal is to..." "... we would like to..." "... the plan is to... if possible..." However it was stated that "...we are going to remove python2 package ..." This is a declaration and an absolute statement as to what is going to happen. One should not really be surprised that such absolutes create waves. Later, Robert -- Robert Schweikert MAY THE SOURCE BE WITH YOU Distinguished Architect LINUX Technical Team Lead Public Cloud rjschwei@suse.com IRC: robjo
On Monday 2019-04-15 15:31, Robert Schweikert wrote:
So, please, let's tone this thread down a notch: Maej formulated a goal to remove python2 by 2020
Well, from my perspective formulating a goal would be more along the lines of:
"...the goal is to..." "... we would like to..." "... the plan is to... if possible..."
However it was stated that
"...we are going to remove python2 package ..."
This is a declaration and an absolute statement
“ The Rules [of Acquisition] are nothing but … guideposts, suggestions. Would you buy a book called "Suggestions of acquisition"? Doesn't quite have the same ring to it! ” -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Dne pondělí 15. dubna 2019 13:39:10 CEST, Matěj Cepl napsal(a):
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
Running 'sudo zypper rm python2*' would now remove bleachbit, calibre, duplicity, grass, inkscape and osc. Trying to investigate Python 3 transition https://github.com/bleachbit/bleachbit is actively maintained and there is some Python 3 progress. At https://github.com/kovidgoyal/calibre and https://calibre-ebook.com/ download_linux I don't see anything about Python 3. Duplicity https://bugs.launchpad.net/duplicity/+bug/1440372 seems to be Python 3 ready soon. GRASS GIS https://svn.osgeo.org/grass/grass/branches/releasebranch_7_6/ REQUIREMENTS.html should be Python 3 in 7.8, so also hopefully soon. Regarding https://inkscape.org/ I didn't find anything. -- Vojtěch Zeisek Komunita openSUSE GNU/Linuxu Community of the openSUSE GNU/Linux https://www.opensuse.org/ https://trapa.cz/
понеділок, 15 квітня 2019 р. 16:01:09 EEST Vojtěch Zeisek написано:
At https://github.com/kovidgoyal/calibre and https://calibre-ebook.com/ download_linux I don't see anything about Python 3. https://github.com/kovidgoyal/calibre/commits/master A lot of py3 commits.
-- Kind regards, Mykola Krachkovsky -- Найкращі побажання, Микола Крачковський
Am Montag, 15. April 2019, 13:39:10 CEST schrieb Matej Cepl:
Hello,
this is just your periodic remainder (which could be expected for the last 10+ years or so, so there shouldn't be any surprise) that on the bright fresh chilly morning of 2020-01-02 we are going to remove python2 package (that's Python 2.7.* interpreter) from Factory.
You should really really make sure to check all your tools have been already ported and tested on Python 3 right NOW.
Last time, I looked, the whole Openstack eco system was primary based on Py2. Not that this is relevant for Factory that much, but highly relevant to the SUSE cloud strategy, if I'm not mistaken... https://www.suse.com/products/suse-openstack-cloud/ Pete -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, Apr 15, Hans-Peter Jansen wrote:
Last time, I looked, the whole Openstack eco system was primary based on Py2.
Must be many years ago if I read the yearly progress reports of porting Openstack to python3 correctly.
Not that this is relevant for Factory that much, but highly relevant to the SUSE cloud strategy, if I'm not mistaken...
The python3 openstack wiki is nearly completly green, only a few orange and red entries. Thorsten -- Thorsten Kukuk, Distinguished Engineer, Senior Architect SLES & MicroOS SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nuernberg, Germany GF: Felix Imendoerffer, Mary Higgins, Sri Rasiah, HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, Apr 15, 2019 at 4:50 PM Thorsten Kukuk <kukuk@suse.de> wrote:
On Mon, Apr 15, Hans-Peter Jansen wrote:
Last time, I looked, the whole Openstack eco system was primary based on Py2.
Must be many years ago if I read the yearly progress reports of porting Openstack to python3 correctly.
Not that this is relevant for Factory that much, but highly relevant to the SUSE cloud strategy, if I'm not mistaken...
The python3 openstack wiki is nearly completly green, only a few orange and red entries.
There's some stupidity in how the SUSE OpenStack packages are done, but on the balance, they're _supposed_ to be Python 3 compatible now. I suspect SUSE will fix those in due time... -- 真実はいつも一つ！/ Always, there's only one truth! -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (15)
-
Adam Majer -
Christian Boltz -
Dominique Leuenberger / DimStar -
Hans-Peter Jansen -
Jan Engelhardt -
Matěj Cepl -
Michael Ströder -
Mykola Krachkovsky -
Neal Gompa -
Richard Brown -
Robert Schweikert -
Stasiek Michalski -
Thorsten Kukuk -
Vojtěch Zeisek -
Wolfgang Rosenauer