[12:20:58] <mgedmin> all I know from lurking in this channel/elsewhere is that https_proxy is supposed to work
[12:21:31] <gordonjcp> hm, I'm not sure how to make it work with the proxy serving https locally
[12:21:40] <gordonjcp> wouldn't it freak out over a different certificate?
[12:22:03] <mgedmin> I don't know how https_proxy works :/
[12:22:30] <gordonjcp> tbh I don't much care about https since if anyone's MITMing the 30 feet of wire between the server in the boiler cupboard and the PC in my workshop I've really got bigger problems on my hands
[12:22:42] <mgedmin> does it connect to the proxy server using HTTPS? or does it connect using HTTP and then use GET https://pypi..../ ? or does it try to use CONNECT pypi.python.org:443 ?
[12:22:56] <ionelmc> gordonjcp: you might need to specify a different index url (that's not on https)
[12:22:59] <gordonjcp> I don't know, I'll dig into it when I get back
[15:10:56] <borisj> Hi, Is it intended in Python 3 that namespace packages can contain regular packages?
[19:15:45] <dstufft> [Tritium]: a more realistic approach would be for PyPI to disallow .exe uploads I think. At some point I plan on going through and trying to disable uploads for formats that don't make sense to have on PyPI, like .rpm and such.
[19:16:11] <dstufft> gordonjcp: pip assumes a CONNECT based proxy I think
[19:17:08] <dstufft> gordonjcp: In any case, it's just requests under the cover, so you can eliminate pip and figure out the right options to get requests to do that and then just translate them to pip options. We do support http:// URLs, but you're required to use --trusted-host myhost.com to mark that host as allowed to be HTTP
[20:06:29] <jleclanche> hiyas, are there dumps of the pypi database available somewhere? I already found the /simple/ index, but wondering if I can crawl the packages without too much overhead for the site.
[20:06:42] <gordonjcp> agronholm: which version are you on?