Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-06 Thread Nick Coghlan
On 7 August 2017 at 12:19, Randy Syring wrote: > Thanks Donald for the clear reply from which it was easy to create these > separate issues. I didn't realize you had so little paid time to work on > packaging issues. I'm surprised with all the large companies that use > Python that the PSF can't

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-06 Thread Randy Syring
Issues created: * Ignore dup uploads instead of error: https://github.com/pypa/warehouse/issues/2284 * Rename "/legacy" API URL: https://github.com/pypa/warehouse/issues/2285 * Propose "packaging announce" mailing list: https://github.com/pypa/packaging-problems/issues/98 o Previous

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-06 Thread Nick Coghlan
On 5 August 2017 at 14:15, Lucas Boppre Niehues wrote: > Also, I agree with Nathaniel Smith that numbered versioning of the API > (/v0/, /v1/, /v/2) would be preferable to /legacy/. I don't have a lot to add to Donald's other responses (Thanks Donald!), but wanted to register a +1 on the idea of

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-05 Thread Alex Walters
> Also -- I know that the various tools involved are maintained by different > people, but it sure would be nice if, for instance, the latest versions of > setuptools and twine would not try to upload non longer supported file types! The tools should have better reporting on the errors pypi give

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-05 Thread Chris Barker
On Fri, Aug 4, 2017 at 5:42 PM, Lucas Boppre Niehues wrote: > > The long description was originally Markdown, and converted to RST by > pandoc. I would 100% understand if this conversion triggered some bug. > It's a good idea to run docutils on it yourself -- but in any case, broken RST has alwa

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-05 Thread Chris Barker
one point: I probably should have named it something other than /legacy/, > yes, it should have -- names matter! and having a "legacy" in teh name when there is not "modern" or "current", or, indeed, anything else is very confusing. pypi.org, might as well get it all done at once. It might be

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-05 Thread Paul Moore
On 5 August 2017 at 03:12, Donald Stufft wrote: > I’ll see what I can do here. It might just make sense for these related set > of tools to just ditch the Github issue trackers and share an issue tracker > that can divide things up by project somehow, but where we can readily move > issues between

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Lucas Boppre Niehues
I created an issue on setuptools tracker about Windows newlines ( https://github.com/pypa/setuptools/issues/1126), and another on Warehouse about the "Invalid file extension" message and PEP 517 ( https://github.com/pypa/warehouse/issues/2281). I'm not creating one for ignoring duplicated files in

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Lucas Boppre Niehues
> So it appears that the wheel isn’t being properly generated. Ah, this helps a lot. I didn't realize it's the first upload that defines the description for the package. Testing different configurations I found that the wheel generation chokes and ignores everything after the first Windows-style

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Donald Stufft
> On Aug 4, 2017, at 8:42 PM, Lucas Boppre Niehues > wrote: > > > >> I can't even specify the description of a package, not even during > >> registration or upload. > > Tell me your package name privately or publicly and I’ll figure out what > > went wrong. > > https://pypi.python.org/pypi/

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Lucas Boppre Niehues
Thank you for the kind responses. I hope I haven't offended anybody, and please know that I'm incredibly grateful for all the sweat and blood that has been poured into these projects. I'm complaining about the current situation to show the largest pain points from my perspective. That's also why I

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Eric Brunson
Donald, thanks for the thoughtful and complete reply to the original poster. On Fri, Aug 4, 2017 at 3:52 PM Donald Stufft wrote: > Going through your text, there’s a mixture of things that sort of combined > here, I’m kind of just going to stream of conscious reply to them here, > because I can

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Nathaniel Smith
On Fri, Aug 4, 2017 at 2:51 PM, Donald Stufft wrote: > This API is also considered “legacy” (it just doesn’t have a > replacement yet). So the legacy in https://upload.pypi.org/legacy/ has to do > with the API rather than the location/deployment. > > I probably should have named it something other

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Donald Stufft
> On Aug 4, 2017, at 4:43 PM, Lucas Boppre Niehues > wrote: > > I just had a dreadful day dealing with PyPI and the migrations that are > happening. I'm posting this rant here in hopes of reaching the people who can > fix it (apologies if I'm in the wrong list), or at least offer an explanati

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread Chris Jerdonek
On Fri, Aug 4, 2017 at 2:29 PM, James Bennett wrote: > I'm sorry that you're frustrated. You're not the only person who's > frustrated, and in fact I'd bet a lot of the people who are trying to get > this stuff to work seamlessly and perfectly for you are at least as > frustrated as you are. They

Re: [Distutils] new PyPI: a rant from a package maintainer

2017-08-04 Thread James Bennett
Since I have less direct personal involvement, I'll say this: Yes, there are problems. A lot of them are around documentation, and timing of when things switch over (the whole Python packaging ecosystem has been in a state of flux/rewrite for a while now; this is unfortunate, but also necessary).