pmxbot IRC Log Viewer

[18:13:53] <pradyunsg> techalchemy: https://github.com/pypa/pipenv/issues/4058

[18:14:13] <pradyunsg> ^ this has blown up a fair bit. You might want to chime in there.

[18:19:47] <techalchemy> thanks pradyunsg

[18:19:48] <techalchemy> answered

[18:25:26] <techalchemy> pradyunsg, is there anything that would prevent cutting over pip search to a web based backend if one existed?

[18:30:51] <cooperlees> techalchemy: If you come up with a hyper media CDN catchable replacement metadata for pip search EWDurbin + dstufft will love you

[18:31:20] <techalchemy> cooperlees, the one you were working on, how much agreement about it was there

[18:31:51] <cooperlees> techalchemy: Also, any interest in a PR for a GitHub action to build and push pipenv to help the next release :D

[18:31:53] <cooperlees> I can do that

[18:32:15] <techalchemy> cooperlees, it's not actually possible to do that

[18:32:29] <techalchemy> until i undo all of our insane patching of pip + piptools

[18:32:43] <techalchemy> which is the actual reason nobody else is able to cut releases atm

[18:32:56] <techalchemy> even i can barely cut releases

[18:33:04] <cooperlees> techalchemy: The new API stalled based on the unit test coverage requirement of Warehouse + no confirmation of acceptance after doing all that work

[18:33:21] <cooperlees> Myself and Austin (who's last name I forget) didn't want to invest that time with the uncertainty

[18:33:34] <techalchemy> cooperlees, is there a languishing pr somewhere?

[18:33:50] <cooperlees> Somewhere - I unfollowed it all etc. after trying @ 2 PyCons

[18:33:58] <techalchemy> i imagine we can make a discourse thread and peer pressure someone into responding

[18:33:59] <cooperlees> asmacdo might remember

[18:34:17] <techalchemy> no worries though i get the frustration

[18:34:22] <cooperlees> There was someone last PyCon who wanted to polish it up and land - I forget his name too sorry

[18:34:28] <techalchemy> hahaa

[18:34:42] <techalchemy> i want to do lots of stuff but then all my time vanishes

[18:35:19] <cooperlees> techalchemy: All this insane patching of pip etc. - This a Reitz decision?

[18:35:44] <techalchemy> cooperlees, that was how the patching came about yeah, we vendor and patch pip

[18:36:14] <techalchemy> now i do my insane pip related hackery in my own libraries and modern versions of pip don't require the kind of stuff we are doing anyway

[18:36:16] <cooperlees> Wow. that's always bound to be a ticking time bomb

[18:36:34] <techalchemy> i mean literally

[18:36:35] <techalchemy> it is

[18:37:39] <pradyunsg> ^ can confirm. I'd tried to poke at it and it blew up. It definitely isn't something I'd want to touch.

[18:37:39] <techalchemy> once pip + setuptools cut conflicting releases breaking editable installs there was no point in releasing an update anyway because revendoring pip would have broken dependency resolution

[18:38:10] <cooperlees> techalchemy: Can you split up what you have left? I could donate some cycles if you can do that. We should cut a Project and lots of small PRs and get people to help you!

[18:39:16] <techalchemy> then it seemed like nobody was going to take ownership of the issues involved so I spent like 3 months cleaning up code and such to handle editables and all the fallback cases from scratch

[18:39:49] <cooperlees> hasn't the latest pip restored all editable bugs?

[18:40:12] <techalchemy> cooperlees, yes everything works now, but i wasn't sure about what the timelines would be when i did that work

[18:40:18] <cooperlees> Can projects link to other pypa issues or PRs your dependent on. Just trying to get what's in your head into a project and PRs for everyone

[18:41:44] <techalchemy> then in may/june i started house hunting, half my team got laid off @ my last job, i wound up getting 'promoted' lol, moved, and had a lot of house related issues to resolve

[18:42:12] <techalchemy> like when it caught on fire that one time in august that was fun

[18:42:13] <cooperlees> techalchemy: Dude, no need to tell me why. It's all good and it happens :)

[18:42:32] <techalchemy> just explaining life has been a little crazy :p

[18:42:45] <cooperlees> yeah, sorry to hear man! Do you like the idea of a project and linking in the issues to get a new release for 2019?

[18:42:55] <cooperlees> Or do you think jan 2020 is a better goal?

[18:43:03] <techalchemy> yeah for sure, i honestly dont think there is too much to do

[18:43:39] <techalchemy> i just merged a PR in requirementslib which handles editable resolution and converts formats from pipfile/pip formatted requirements

[18:43:53] <cooperlees> When I get in this situation I generally spit it out in a Doc. Get other contributors to review. Cut a project and link away the smaller issues

[18:43:57] <cooperlees> This is all using FB tools usually

[18:44:07] <cooperlees> But I tried to github-a-fy my lingo

[18:44:17] <techalchemy> yeah good idea

[18:44:32] <techalchemy> i'll start by getting organized about what is left

[18:44:48] <cooperlees> Split things up into really small shiut and then call on that Issue for help linking to the project and watch the PRs fly in

[18:44:49] <techalchemy> maybe i'll take some time today and just work on that

[18:44:57] <cooperlees> Look @ how many people commented on that thread. People want the pipenv!

[18:45:07] <cooperlees> Capitalize :)

[18:45:28] <cooperlees> Once you do it I might do a PR or two so I can be a l33t pipenv contributor

[18:45:36] <techalchemy> i didn't read the HN thread but i saw the # of reactions on the issue

[18:45:39] <cooperlees> Haven't got dam Internet points yet

[18:45:47] <cooperlees> yeah, I only looked at the Issue

[18:45:56] <cooperlees> s/dam/dem/g

[18:46:21] <techalchemy> we should make some kind of way to track internet points

[18:46:28] <techalchemy> i know we should call it a 'like' button

[18:49:32] <cooperlees> That may be trademarked by some bastardos

[18:49:39] <cooperlees> With a large legal team

[19:02:06] <techalchemy> i'll go argue my own case surely the judge will agree iwth me

[19:37:14] <asmacdo> techalchemy: cooperlees: i closed it but if this is helpful https://github.com/pypa/warehouse/pull/4078

[19:52:19] <toad_polo> I hope no one minds, but I have locked this thread: https://github.com/pypa/packaging-problems/issues/307

[19:54:07] <toad_polo> I may have been a bit uncharitable in my assumptions about the mindset of the person creating the issue, but this whole thing feels like an HN pile-on and I'd rather just cut it short than indulge it.

[19:55:48] <cooperlees> thanks asmacdo

[21:11:54] <ronny> is there any way to make pip build isolation not destroy the folders?

[21:12:31] <ronny> i want to investigate https://github.com/pypa/setuptools_scm/issues/381 - to find the exact breakage point, for that i need the tmpdirs pip makes intact

[21:12:33] <ronny> cc pradyunsg

[21:20:25] <ronny> i found --no-clean

[21:40:12] <techalchemy> thanks toad_polo

[21:40:48] <techalchemy> toad_polo, I literally never made the connection about who you are

[21:41:06] <techalchemy> i am so slow sometimes

[21:41:41] <toad_polo> That is probably for the best. I am mainly only using this `toad_polo` name because I use Matrix, which doesn't have good separation of identities.

[21:42:01] <ronny> hmm

[21:42:43] <toad_polo> At some point I will abandon the `toad_polo` name like a Hermit crab discarding a shell.

[21:42:49] <ronny> does pip have any way to force a local folder in as replacement for a pep517/8 build env - i want to do a regression test for setuptools_scm and it seems i cant force it in

[21:43:15] <techalchemy> ronny, you can just patch pip like i do

[21:43:25] <techalchemy> see above to find out how great of an idea it is ;)

[21:43:52] <techalchemy> but seriously you can comment out the line where it deletes build directories

[21:44:34] <techalchemy> or to avoid reading code you can comment out the cleanup functionality in temp_dir.TempDirectory or whatever the class name is called

[21:45:12] <techalchemy> I don't think you have options to manipulate these things from the command line

[21:51:08] <ronny> --no-clean already does the non cleanup, i already wrote that

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Friday the 13th of December, 2019