Log file Viewer
#python2.8 logs for Thursday the 23rd of January, 2014
[00:48:18] <tomprince> regebro: Just because the core devs won't release a python 2.8 doesn't mean it won't happen. stackless has plans to release a 2.8 and pypy is open to the possibility.
[00:55:04] <dstufft> I hope either some implementation becomes the dominant and geneally accepted 2.8, or 2.8 never happens
[00:59:59] <tomprince> dstufft: Or both pypy and stackless, being developed by somewhat sane people can agree on what goes into it.
[01:01:20] <tomprince> I know the pypy are somewhat sane. At least ... in regards to writting software that people want to use.
[05:38:22] <regebro> And note that there is no "The core devs will not release Python 2.8". They can change heir minds.
[05:39:00] <regebro> My point is that until Python 3 is declared dead, a Python 2.8 will not have as it's aim to to help people stay on Python 2.
[14:01:00] <aclark> regebro: i don't buy that a 2.8 to bridge 3.x and 2.x will "keep people on 2.x" and prolong the discussion of moving to 3. I picture 2.8 as a bridge. 3's original design included breaking backwords compat, which worries some folks about "two communities". A 2.8 *might* lead toward a reunification of those communities. But I do agree that there is little incentive for: 1.) core devs to develop a 2.8 and 2.) 2.x users to m
[14:02:11] <aclark> regebro: a bridge is one or both of the following: 2.x features restored to 3.x and/or 3.x features backported to 2.x
[14:02:15] <regebro> Those people I argued with yesterday wanted a 2.8 that had all new features of 3.4, but nothing incompatible.
[14:02:49] <regebro> If that's a "bridge", that just makes the Python 3 benefit lower, while keeping the cost high, hence making less people move.
[14:02:53] <aclark> Yeah that's probably not going to happen. How could it? It's fair to break some stuff. Even important to do so I think.
[14:03:29] <aclark> That's how software works: you break shit, document it, and give people reasonable time to catch up.
[14:05:42] <regebro> But at least yesterday nobody wanted to discuss that, because they didn't want to upgrade.
[14:05:49] <aclark> I don't know the code that well I suspect that there are areas to meet in the middle to define what "makes sense" to break or include or whatever. I will also say this: I think drawing a line in the stand that 2.7 was the end of 2.x was a mistake. It's OK to say "this might be the last 2.x" or "we'd like it to be" but drawing the line just pisses people off unnecessarily
[14:06:44] <aclark> If you are a core dev, keep it open as a possiblity even if you never plan to write a line of code for it.
[14:06:45] <regebro> I think it's important to be clear. They intended 2.7 to be the last, then they need to say that.
[14:07:03] <regebro> Perhaps the PEP should have included a list of "things that will make us change our mind".
[14:07:14] <aclark> You can have your cake and eat it too if you know the slightest thing about human nature :-)
[14:07:54] <aclark> I know close to nothing about the situation and I'm a Python 3 fan, but hearing "no 2.8" just rubs me the wrong way. Imagine if I actually cared.
[14:08:18] <regebro> Yeah, well, a core-dev team of people, some of whom are even worse at dealing with others than I am are not expected to do things smoothly.
[14:10:18] <dstufft> I'd rather someone be upfront that they don't plan to putting effort into continuing a series then wait around and be unuaware and be more or less at the mercy of trying to guess if it's planned or not
[14:14:33] <aclark> dstufft: that's not the issue. the issue is: "there will never be a 2.8" vs. "i don't plan to work on 2.8". you can say the latter without getting in to a fist fight.
[14:15:15] <aclark> the former makes you sound like you are drawing a line and/or have some "moral authority". that's what pisses people off. if you don't mind fighting about, then you can say the former :-)
[14:17:24] <aclark> E.g. good: "I am interested in working on and/or developing Python 3" vs. Bad: "Python 3 is better than Python 2." The latter you will never convince some folks of so why try.
[15:21:10] <tos9> regebro: FWIW I find it really really really disingenuous that you keep saying htings like that we are not willing to discuss whatever you are saying.
[15:22:30] <regebro> tos9: If you are willing to discuss how to make it easier to upgrade to Python 3, why can't we discuss that?
[15:23:00] <regebro> Why do we end up discussing how Evil Python-dev is for not making a Python 2.8?
[15:23:10] <tos9> regebro: We have. And I am. It's hard to have that discussion unless you understand the layout of what is on the table. But if we entirely ignore anyone who cares about Python 2.8 in particular, we can certainly have that conversation.
[15:23:39] <tos9> So, let's start from "forget Python 2.8 for people who want Python 2.8 for its own sake, and just talk about helping people port to Python 3"
[15:23:46] <regebro> tos9: Now you are implying that everyone who cares about Python 2.8 are people who don't want to move to Python 3.
[15:23:52] <faassen> I have a blog entry in the works with some ideas for Python 2.8, but I need to have time to finish it up.
[15:24:04] <regebro> I disagree with that, and I don't think that's a constructive path of discussion.
[15:24:11] <tos9> regebro: No? Although as I said yesterday, I don't think python 2.8 is such a viable idea for helping people port
[15:25:19] <regebro> tos9: I suspect you are right, Python 2.8 is not needed to help people port. But I don't know for sure, since nobody seems willing to discuss that.
[15:25:25] <tos9> I think that the only thing core-dev can do at this point is just keep adding things to Python 3 and wait it out
[15:26:14] <regebro> Well it's certainly *possible* to make a Python 2.8 that lowers the cost of upgrading.
[15:26:42] <tos9> Sure, I agree. It's not going to make things worse, but I don't see what it helps anymore.
[15:26:47] <faassen> I think if people aren't going to make a Python 2.8 that is helpful in upgrading, people are going to make one that doesn't. :)
[15:28:10] <tos9> Python 2.7 is a good enough language to hold out on for me. Although I guess core-dev has a slightly different definition of what *should* be backported to 2.7, which is highly unfortunate, but I can live with that
[15:28:11] <regebro> But I think the "Python 2 + all the compatible goodies" actually is such a big task that it won't happen without coredevs.
[15:29:10] <tos9> It's not too much work for some other team to tackle, I think there'd be more of a pushback/struggle kind of problem.
[19:49:59] <regebro> My position on things, http://regebro.wordpress.com/2014/01/23/the-potential-for-a-python-2-8/
[20:12:46] <richmoore2> interesting that you say " if you want “Psnakes 2.8″ what should you do? Well, stop complaining and do it. " but then seem to be sabotaging the discussions of people who are actually interested in doing it
[20:12:52] <tos9> but trying to get things on paper so that we all see how things lie is always good
[20:18:29] <regebro> There's nothing to discuss. If you want to make Psnakes 2.8, go ahead and do so.
[20:18:51] <richmoore2> and from the blog post that pointed me to this channel, this is the place to discuss it with like minded people
[20:20:53] <regebro> richmoore2: This channel is not only for discussing how to create Psnakes 2.8. It's for discussion about Python 2.8.
[20:21:31] <regebro> Specifically: It's potential for a Python. what could it contain, and how could it help moving people to Python 3.
[20:22:10] <regebro> richmoore2: I honestly don't understand the question "how". What kind of answers do you expect?
[20:22:41] <regebro> richmoore2: Yet again, you are not being constructive, but rude and whiney. Not useful.
[20:23:44] <richmoore2> but please stop saying "that will never happen" to everyone who disagrees with your vision
[20:24:32] <regebro> richmoore2: Is your vision that the Python core devs will start work on a Python 2.8 that contains new features?
[20:25:01] <richmoore2> i'm interested in discussion what we'd like to see in a python 2.8 as and end in itself, python 3.x is partly irrelevant to that (though certainly not completely)
[20:25:06] <regebro> Because that's what I said isn't going to happen unless Python 3 is declared a failure by those same core devs.
[20:25:52] <richmoore2> i'm much more interested in improvements to the library (including potentially the renaming stuff offered by six) than i am in any changes to the core language
[20:25:53] <regebro> richmoore2: The answer to that is pretty obvious: All news Python 3 features that can be included without breaking backwards compatibility.
[20:26:43] <regebro> I don't see what there is to discuss there. Possibly you can discuss in which order to add them.
[20:27:29] <richmoore2> and potentially (certainly for the library part which is much easier) this a place to get that. but not if they get shouted at because they don't care about python 3
[20:28:21] <richmoore2> note /don't care/ != /actively hate/ or trying to tell other people what they should code
[20:28:40] <regebro> richmoore2: This is honestly not a place to plan what does into your Python 2 fork. That should rather be planned with the team you cretae for that fork, in that forks channel.
[20:29:02] <regebro> Asking for people to help is certainly something you should do here. (and other places).
[20:30:32] <regebro> But you are welcome to discuss what should go in it here as well. But stop trying to discuss it with *me*.
[20:30:37] <richmoore2> http://blog.startifact.com/posts/python28-discussion-channel-on-freenode.html doesn't seem to agree with you
[20:31:43] <eriknw> huh? this is absolutely the place to plan what may go into a (forked) Python 2.8
[20:32:10] <eriknw> if you're not interested in that discussion, please stop trying to stop others from having it
[20:32:48] <regebro> It seems that many people have a hard time understanding that when then talk to me they will get an answer.
[20:33:57] <regebro> eriknw: What I'm saying is that if you want a Python 2.8 which backports the backwards compatible new feature from Python 3 (and that seems to be what you want) there really isn't much to discuss about *what* goes into that Python 2.8.
[20:36:02] <regebro> eriknw: Oh, well, OK, so you don't know what the new features in Python 3 are? So you don't know what you want, really.
[20:37:03] <eriknw> I use primarily python 2, so, no, I don't know *all* the improvement in Python 3. I am sure there are many like me. I do know many features, but have not actually used them in production, *because I don't use Python 3*, and that's not my choice
[20:37:20] <regebro> What should go in is the easy question for you. The question you *should* ask yourself is how to make it happen.
[20:39:01] <regebro> Well, OK, since I'm not interested in your Python 2.8, I'm not going to spend more time helping you make it happen. That's your problem.
[20:39:51] <eriknw> also, don't presuppose that the community would *ask* most of the core devs to develop python 2.8, but to allow for its existence (however it may be developed)
[20:40:54] <regebro> eriknw: They already do that. It's open source. The are not in a position to disallow it.
[20:41:03] <richmoore2> i don't think anyone needs to ask the core developers for anything (though hosting things on python.oth would make stuff easier)
[20:41:53] <regebro> eriknw: You don't have to doubt it, it can not use it. This is not a significant problem.
[20:42:26] <eriknw> it would reduce the adoption of the proposed python 2.8 if it didn't have "Python" in the name
[20:42:36] <regebro> (This is why I call this fork "Psnakes 2.8". It's shorter that "A Python 2.8 with backported features")
[20:42:40] <richmoore2> we had someone wanted to carry on maintaining kde3, and we provided their hosting, bugtracker etc. even though we thought they were mad. it wasn't a problem
[20:43:08] <richmoore2> i'm sure the python core developers etc. would be just as accomodating even if they think the same thing.
[20:43:27] <richmoore2> actually i suspect they'd be interested in using it as a way to define requirements for future python3 releases
[20:43:48] <regebro> richmoore2: Should a fork become viable and popular, I'm sure it can be embraced. But in the start that's not going to be an option. But as mentioned, that's not a problem.
[20:45:21] <regebro> To me you are still discussing things that are obvious and easy. The problem still is finding people that will put down all the massive effort involved.
[20:46:11] <richmoore2> step one is getting the details of the requirements - "All news Python 3 features that can be included without breaking backwards compatibility." really doesn't cut it
[20:47:08] <richmoore2> even including six and a few other modules would be a step forward, potentially with a from future import hook to load them
[20:47:35] <richmoore2> or from potential future if we don't want to collapse the space time continuum :-)
[20:48:17] <regebro> Unless you have a team to do it, the details of the requirements are irrelevant. I promise this is the last time I say this, though, good advice is hard to come by and even harder to get people to follow. :-)
[20:50:54] <eriknw> the details of the requirement are not irrelevant. knowing what needs done, and having the effort coordinated (as evident by a detailed list) would encourage a team to be formed
[20:51:21] <richmoore2> my oldest active open source project has been running for 18 years, so it's not like it's the first time :-)
[20:52:52] <richmoore2> i actually think that backporting the lib improvements would make porting to python3 easier, even though that's no an aim for me personally
[20:57:32] <regebro> But that can also be done with six, etc, so you don't really *need* Python 2.8 for that.
[20:58:52] <richmoore2> one of the first things on my list is that six should be part of the whateverwecallit 2.8 std library
[20:59:28] <richmoore2> not to ease migration but simply because the organisation is better. the differing goals are not incompatible