| ilasc | am I the only not able to run Turnip API layer? Getting this from a fresh Master clone: | 09:50 |
|---|---|---|
| ilasc | pkg_resources.ContextualVersionConflict: (pycparser 2.19 (/home/ilasc/turnip/env/lib/python2.7/site-packages), Requirement.parse('pycparser<2.18'), set(['pygit2'])) | 09:50 |
| cjwatson | ilasc: What container setup steps have you done? | 09:50 |
| cjwatson | Also Thiago just landed some dependency changes which might be relevant | 09:51 |
| cjwatson | Worth checking that your dependencies clone is up to date | 09:52 |
| ilasc | cjwatson: went with fresh container and followed the Turnip Readme file for a completely isolated env. | 09:52 |
| ilasc | will do, thanks! | 09:52 |
| cjwatson | I remember this area being delicate | 09:52 |
| cjwatson | Hmm | 09:54 |
| cjwatson | I wonder how pappacena managed to land this | 09:55 |
| cjwatson | requirements.txt has pycparser==2.19 and pygit2==0.27.4. But pygit2 0.27.4 depends on pycparser<2.18 | 09:55 |
| cjwatson | Which is indeed what pip is telling you | 09:55 |
| cjwatson | I don't understand how Jenkins passed given that | 09:57 |
| cjwatson | pappacena: ^- could you have a look at this when you get in? | 09:57 |
| ilasc | agreed - thank you Colin! | 09:58 |
| SpecialK|Canon | tests pass on master | 09:59 |
| cjwatson | pappacena: pygit2 upgrades are delicate and require careful coordination because of the libgit2 dependency; it may be simplest to back off a little to pycparser 2.17 if that's not too problematic for other reasons | 09:59 |
| cjwatson | perhaps whether it passes depends on what version of pip you're starting with | 09:59 |
| cjwatson | I wonder if ilasc has a different version of pip on $PATH or something | 10:00 |
| ilasc | I'm trying to follow your recommendation and eliminate all the "on the host" setup items. Started by get my setup as similar to Tom's as possible - completely isolated LXD (as I should've done from the begging) and VSCode remotely attached for debug to each LXD | 10:00 |
| ilasc | SpecialK|Canon: tests pass in my container too | 10:01 |
| ilasc | it's the "make run-api" that fails | 10:02 |
| cjwatson | Ah right, so tests pass but then something later complains, I see | 10:02 |
| ilasc | yes, sorry! wasn't clear enough at first | 10:02 |
| cjwatson | So I guess pserve calls the usual entry point machinery in pkg_resources and that notices the unsatisfied dependencies | 10:03 |
| cjwatson | Wouldn't hurt to spend a little time writing a failing test for this while we have the convenient opportunity to do so | 10:03 |
| cjwatson | And then we can make sure it passes once pappacena fixes up the dependencies | 10:03 |
| cjwatson | (And yes, I see the same "make run-api" failure) | 10:04 |
| cjwatson | ilasc: You can run "env/bin/pip install pycparser==2.10" (in the container of course) to temporarily fix up your environment so that "make run-api" works | 10:06 |
| cjwatson | Perhaps we should have a unit test that runs "env/bin/pip check" | 10:07 |
| cjwatson | Or even just run that in the Makefile after installing dependencies | 10:07 |
| cjwatson | Probably the latter makes more sense | 10:07 |
| cjwatson | Could also possibly just land an MP that backs out to pycparser==2.10 in requirements.txt. But pappacena should be aware in case that disrupts their py3 work | 10:11 |
| ilasc | cjwatson: noted and thank you, 1:1 with Kristian, reading now, agreed and I will add the Unit Test after lunch | 10:29 |
| cjwatson | Possibly build system change rather than unit test, but sure | 10:40 |
| pappacena | Hey, folks. I'll take a look on this problem new. I'll probably downgrade this dep. AFAIR, it was probably just raising some DeprecationWarning on py3. Shouldn't be a big problem to downgrade it. I'll add the check on Makefile too. Sorry about that, and thanks for the help, ilasc! | 11:50 |
| tomwardill | https://code.launchpad.net/~twom/launchpad/+git/launchpad/+merge/380416 is updated with a shortcut as I couldn't find a bettter way that was reliable | 11:51 |
| cjwatson | tomwardill: Yep, I think that'll do | 11:59 |
| tomwardill | ta, landing | 12:03 |
| ilasc | pappacena: I'm just back from lunch, great timing :) I'll leave you to it in that case - working on the Unit Tests for my "show MP URL" in Turnip, I can stay away from "run make-api" for now | 12:12 |
| pappacena | Thanks, ilasc. It shouldn't take long. I'll open the MP in a couple of minutes. | 12:12 |
| ilasc | pappacena: excellent, thank you! | 12:13 |
| pappacena | It should be good to be reviewed now: | 12:21 |
| pappacena | Deps MP: https://code.launchpad.net/~pappacena/turnip/+git/turnip-dependencies/+merge/380545 | 12:21 |
| pappacena | Code MP: https://code.launchpad.net/~pappacena/turnip/+git/turnip/+merge/380546 | 12:21 |
| cjwatson | pappacena: r=me on both, thanks | 12:24 |
| pappacena | Thanks, everyone. | 12:24 |
| ilasc | pappacena: +1 | 12:25 |
| cjwatson | pappacena: also I've just updated https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+merge/380391 per your comments | 12:25 |
| pappacena | Awesome. I'll take a look! | 12:25 |
| cjwatson | Cheers | 12:25 |
| cjwatson | pappacena: Great, thanks. You have a bunch of review comments in exchange ;-) | 13:16 |
| cjwatson | (Not sure to what extent you'll consider this a fair exchange) | 13:16 |
| SpecialK|Canon | cjwatson: Thank you for handling Q689260 (mirror 307) | 13:29 |
| cjwatson | NP, I had some useful mental state about that | 13:37 |
| pappacena | cjwatson, well... it's a change. And feedback is always good hahaha | 13:43 |
| pappacena | *"an exchange", I mean | 13:55 |
| tomwardill | Did the thing: https://dev.launchpad.net/Soyuz/HowToDevelopWithBuildd | 15:39 |
| tomwardill | Linked under 'Set up for development' on https://dev.launchpad.net/Soyuz/HowToUseSoyuzLocally | 15:39 |
| cjwatson | Great, thanks | 15:42 |
| tomwardill | I'll try and keep on top of lxd changes there, when we can make it simpler | 15:44 |
| SpecialK|Canon | <3 | 15:44 |
| tomwardill | the drive mounting not doing uid mapping is a bit awkward atm | 15:44 |
| tomwardill | cjwatson: can you point me at your oci registry branch? | 16:12 |
| tomwardill | I've lost it somewhere | 16:12 |
| cjwatson | tomwardill: https://code.launchpad.net/~cjwatson/launchpad/+git/launchpad/+ref/db-oci-push-rule I think? | 16:14 |
| tomwardill | that's the one | 16:14 |
| tomwardill | I was searching for registry | 16:14 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!