Log file Viewer

[04:04:21] <lifeless> sheese

[04:04:28] <lifeless> I'm over these JSON-LD and RDFa posts.

[05:47:39] <dstufft> lifeless: have you accepted JSON-LD as your lord and savior

[09:41:27] <lifeless> dstufft: no, not yet

[09:42:34] <ronny> dstufft: is there a church, doesit come with mental health benefits?

[20:50:24] <dstufft> lifeless: https://www.reddit.com/r/Python/comments/3k41zf/pip_71_constraints/

[20:50:46] <dstufft> Yasumoto: you know things about pex right?

[20:50:57] <Yasumoto> yep

[20:51:16] <lifeless> dstufft: :)

[20:51:17] <Yasumoto> wickman is the man with the plan, but I'm a decent starting point

[20:51:46] <dstufft> Yasumoto: if I pex up something that depends on say, psycopg2. Will pex bundle libpq.so as well, or will it still depend on libpq.so to be installed on the system

[20:52:00] <Yasumoto> it will still need that native lib on the system

[20:52:08] <dstufft> sad

[20:52:11] <dstufft> I AM SAD

[20:52:13] <dstufft> Okay

[20:54:00] <Yasumoto> hm... could've sworn

[20:54:01] <Yasumoto> https://github.com/pantsbuild/pex/issues?q=is%3Aopen+is%3Aissue

[20:54:03] <Yasumoto> had something about it

[20:54:29] <Yasumoto> fwiw I'm not too stoked about it either, I'd like to pull in all platform specific libs into a pex

[20:54:33] <Yasumoto> vs. just a compiled dist

[20:54:39] <Yasumoto> so I'd support that fo sho

[20:54:53] <Yasumoto> (my hunch is there's a reason it doesn't happen already, but I don't recall what it may be)

[20:56:32] <dstufft> probably because everything is terrible

[20:56:47] <dstufft> that's my defacto guess for why nothing works how it should!

[20:56:51] <dstufft> also computers.

[21:01:31] <lifeless> dstufft: so I have this draft pep to move forward on getting that lib installed

[21:08:32] <dstufft> lifeless: I haven't even begun to read that PEP :( I probably should at some point

[21:08:49] <lifeless> dstufft: I have a raft of edits planned

[21:08:56] <lifeless> dstufft: been so fragmented these last weeks

[21:09:05] <dstufft> sounds like I should wait downstream then

[21:20:25] <ErikRose> dstufft: Any better ideas for a name for the hash-checking mode? "Secure Mode" has too much overlap with other security features, like HTTPS cert checking. "Hash-checking mode" is more accurate but long.

[21:22:34] <dstufft> does it need a short name (or a name at all really)? Do we expect people to interact with the name at all? AFAIK the proposal is that if you hash a hash then all must be hashed, but thre's no flags or anything like that?

[21:22:42] <dstufft> I mean there's like --sha256 or whatever

[21:22:52] <dstufft> but there's no flag to explicitly enable it

[21:23:29] <ErikRose> I'm just naming vars for now. Just wondered if you had an obvious one.

[21:23:50] <ErikRose> The only exposure I see in the future would be a CLI flag, and --ignore-hashes is fine for that.

[21:34:35] <xafer> dstufft, https://github.com/pypa/pip/pull/3085 seems ok to me but I'm wondering about the tests: we were testing "pip install --download..." and with this PR we will only test "pip download ...". Shouldnt we test both calls during the deprecation period ?

[21:40:48] <dstufft> xafer: um, IDK. We generally don't test deprecated stuff but maybe we should here

[21:46:34] <xafer> I think we should

[21:46:43] <dstufft> works for me

[21:47:49] <xafer> k, if the PR is green tomorrow I think I'll merge it if you dont see any blocker ?

[21:48:13] <dstufft> I haven't looked very closely, I don't have any problem in theory

[21:48:27] <dstufft> it might need to duplicate the deprecation warnings for --allow-external

[21:48:33] <dstufft> (or make sure it doesn't accept them)

[21:48:41] <dstufft> it probably doesn't need to accept them at all

[21:49:52] <xafer> well it's a new command so it shouldn't accept them I think

[22:17:58] <lifeless> ErikRose: HCM ?

[22:18:52] <ErikRose> lifeless: I myself just googled that to see what you were talking about. ;-)

[22:19:25] <ErikRose> I did end up calling hash-checking mode in the code. It turns out to show up in some exception messages.

[22:19:30] <ErikRose> ^calling it