Combination of three paprs #9

cevich · 2017-11-02T14:21:03Z

Very rough ATM, minimal cleanup done, but maybe could work?

rh-atomic-bot · 2017-11-02T14:21:28Z

💥 Invalid .papr.yml: failed to parse 1st testsuite: Schema validation failed:

Value 'golang' is not unique. Previous path: '/packages/3'. Path: '/packages/19'.
Value 'device-mapper-devel' is not unique. Previous path: '/packages/7'. Path: '/packages/14'.
Value 'gpgme-devel' is not unique. Previous path: '/packages/6'. Path: '/packages/20'.
Value 'git' is not unique. Previous path: '/packages/0'. Path: '/packages/16'.
Value 'glib2-devel' is not unique. Previous path: '/packages/10'. Path: '/packages/17'.
Value 'libassuan-devel' is not unique. Previous path: '/packages/8'. Path: '/packages/21'.
Value 'libseccomp-devel' is not unique. Previous path: '/packages/5'. Path: '/packages/22'..

rh-atomic-bot · 2017-11-02T14:26:45Z

💥 Invalid .papr.yml: failed to parse 1st testsuite: Schema validation failed:

Key 'ackages' was not defined. Path: ''..

cevich · 2017-11-02T14:42:28Z

ohh, forgot to add more memory.

jlebon · 2017-11-02T14:50:09Z

.papr.yml

+  - master
+
+host:
+  distro: fedora/26/cloud


Why not fedora/26/atomic here? Isn't that more aligned with what this tool is targeted at?

For an example of how this can work, see the pattern in projectatomic/atomic, where we just build and e.g. pylint in a container, but then install onto the host and run the integration tests natively. (And just like projectatomic/atomic, a second fedora/26/cloud testsuite can always be added).

@jlebon because IRC 😀

rh-atomic-bot · 2017-11-02T14:51:03Z

💥 Invalid .papr.yml: failed to parse 1st testsuite: Schema validation failed:

Key 'spec' was not defined. Path: '/host'..

cgwalters · 2017-11-02T14:56:04Z

.papr.yml

@@ -1,29 +1,34 @@
 branches:
  - master
-  - auto
-  - try


Drop this hunk?

jlebon · 2017-11-02T15:43:24Z

bot, retest this please

jlebon · 2017-11-02T15:57:09Z

bot, retest this please

jlebon · 2017-11-02T16:00:36Z

bot, retest this please

jlebon · 2017-11-02T16:09:33Z

bot, retest this please

cevich · 2017-11-02T16:23:18Z

Makefile


 testunit:
 	$(GO) test -tags "$(BUILDTAGS)" -cover $(PACKAGES)

-localintegration: clean binaries test-binaries
+localintegration: binaries test-binaries
 	./test/test_runner.sh ${TESTFLAGS}


@mheon I remove clean from this. That allowed tests to run (b/c clean removes bin/common). But maybe that was bad to do.

Binaries should be rebuilding conmon. It didn't look like it was. I'll look into it.

I am removing both the clean and the binaries. Currently we don't build anywhere but make all or make binaries. lets keep it that way.

kk, so make -C $GOSRC binaries install.tools all TAGS=${TAGS} and I commented out the original one.

cevich · 2017-11-02T16:28:57Z

@rhatdan @mheon and @baude helped me get this one building locally. It's running well all the way through executing the integration tests. So the papr stuff is probably working fine (when papr starts working again we'll know for sure) and it's code/makefile stuff that's breaking now. Heading to get lunch to give jlebon some time.

cevich · 2017-11-02T16:34:03Z

ahh there she goes! thx @jlebon !

rh-atomic-bot · 2017-11-02T19:05:58Z

💥 Invalid .papr.yml: failed to parse 1st testsuite: Schema validation failed:

Value 'libseccomp-devel' is not unique. Previous path: '/packages/5'. Path: '/packages/22'.
Value 'glib2-devel' is not unique. Previous path: '/packages/10'. Path: '/packages/17'.
Value 'golang' is not unique. Previous path: '/packages/3'. Path: '/packages/19'.
Value 'device-mapper-devel' is not unique. Previous path: '/packages/7'. Path: '/packages/14'.
Value 'gpgme-devel' is not unique. Previous path: '/packages/6'. Path: '/packages/20'.
Value 'libassuan-devel' is not unique. Previous path: '/packages/8'. Path: '/packages/21'.
Value 'git' is not unique. Previous path: '/packages/0'. Path: '/packages/16'..

cgwalters · 2017-11-02T19:20:20Z

.papr.yml

+    - gpgme-devel
+    - libassuan-devel
+    - libseccomp-devel
+    - libselinux-devel


It looks like papr doens't like the duplicates.

no, makes it very angry 😀 s'okay, my own damn fault (again).

cevich · 2017-11-02T19:35:31Z

OMG! There are no quotes around ${TAGS} so it's treating multiple SPACE-SEPARATED tags as multiple arguments, not multiple tags...adding J$%*%&&388#! quotes...

cevich · 2017-11-02T19:44:31Z

bot, retest this please

cevich · 2017-11-02T20:14:32Z

bot, retest this please

rh-atomic-bot · 2017-11-02T23:35:45Z

☔ The latest upstream changes (presumably #6) made this pull request unmergeable. Please resolve the merge conflicts.

rhatdan · 2017-11-02T23:37:42Z

Needs a rebase.

cevich · 2017-11-03T13:48:49Z

@rhatdan @mheon said there's no reason other than "sexiness" to do the build inside a container. I'm not opposed to it in practice...except for...your way introduces a tight-coupling between the host-setup, and the testing. IMHO, best to have a gap there, b/c it makes it easier to swap one host environment for another. And/Or, having multiple VMs running the same testing steps (supported by papr). I think @baude had in mind to do something like this

All that said, are you married to the container-way? (I'm happy to amend this PR to only realize the division of labor part.)

cgwalters · 2017-11-03T13:56:29Z

It's not sexiness - you should always do builds in containers, and don't have that stuff on your host. Even for workstations. There are many reasons (security), and you'll be a lot less likely to hit package management disasters. And further once you do this, it's just as easy to spin up e.g. CentOS/RHEL/whatever containers and use those.

Signed-off-by: Chris Evich <[email protected]>

cevich · 2017-11-03T14:27:26Z

@cgwalters or...we do it both ways :D

cgwalters · 2017-11-03T14:29:26Z

FWIW I think "Fedora Atomic" is Fedora too. So I'd title it "Test in Fedora Cloud and Fedora Atomic".

cevich · 2017-11-03T14:30:41Z

FWIW I think "Fedora Atomic" is Fedora too. So I'd title it "Test in Fedora Cloud and Fedora Atomic".

True true, an easy fix, thx.

Signed-off-by: Chris Evich <[email protected]>

rhatdan · 2017-11-03T15:28:58Z

I would prefer if the tests would happen outside of the container, since things like SELinux will not be tested if running inside of a container.

cevich · 2017-11-03T15:34:48Z

@rhatdan latest push does both, in and out of a container. However @baude wants to take this over, so closing...

cgwalters · 2017-11-03T15:48:31Z

In bubblewrap's papr we build in a container, then install to the host and test there.

cgwalters · 2017-11-03T15:48:55Z

Note specifically the use of e.g. ostree admin unlock etc.

jlebon · 2017-11-03T15:56:02Z

Another great example of this pattern that I mentioned earlier is projectatomic/atomic, which @baude is intimately familiar with :). Specifically, notice how the PAPR YAML simply calls out the same script for all three platforms.

cevich · 2017-11-03T16:40:09Z

@jlebon @cgwalters I think we're all in general agreement. Thanks for the links. @baude is taking this over.

Move libraries from cni only used by plugins.

Add a bunch of endpoints

cevich force-pushed the pr_no_7 branch from 5a3889b to fce6010 Compare November 2, 2017 14:26

cevich force-pushed the pr_no_7 branch from fce6010 to cc7f6fb Compare November 2, 2017 14:27

cevich force-pushed the pr_no_7 branch from cc7f6fb to 5dfaf64 Compare November 2, 2017 14:44

jlebon reviewed Nov 2, 2017

View reviewed changes

cevich force-pushed the pr_no_7 branch from 5dfaf64 to 96bc98d Compare November 2, 2017 14:54

cgwalters reviewed Nov 2, 2017

View reviewed changes

.papr.yml Outdated

@@ -1,29 +1,34 @@

branches:

- master

- auto

- try

Copy link

Contributor

cgwalters Nov 2, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Drop this hunk?

cevich force-pushed the pr_no_7 branch 2 times, most recently from 06f960c to d64221f Compare November 2, 2017 15:07

cevich changed the title ~~Merge PR 6 and PR 7~~ Combination of three paprs Nov 2, 2017

cevich force-pushed the pr_no_7 branch from d64221f to 141aaec Compare November 2, 2017 15:13

cevich force-pushed the pr_no_7 branch from 141aaec to 5dde576 Compare November 2, 2017 16:02

cevich force-pushed the pr_no_7 branch from 5dde576 to e1b15bb Compare November 2, 2017 16:20

cevich commented Nov 2, 2017

View reviewed changes

cevich force-pushed the pr_no_7 branch 3 times, most recently from 1e2abb9 to e52ed1e Compare November 2, 2017 18:55

cgwalters reviewed Nov 2, 2017

View reviewed changes

cevich force-pushed the pr_no_7 branch from e52ed1e to 3eb050c Compare November 2, 2017 19:22

cevich force-pushed the pr_no_7 branch from 3eb050c to 01d6f11 Compare November 2, 2017 19:42

cevich force-pushed the pr_no_7 branch from 01d6f11 to 91b7527 Compare November 2, 2017 20:12

cevich force-pushed the pr_no_7 branch from 91b7527 to ea6eead Compare November 2, 2017 21:04

cevich force-pushed the pr_no_7 branch from 9a3386f to e486ee3 Compare November 3, 2017 13:39

Decouple host-setup from test-setup

ce537ff

Signed-off-by: Chris Evich <[email protected]>

cevich force-pushed the pr_no_7 branch from e486ee3 to ce537ff Compare November 3, 2017 14:06

Test in both Fedora and Fedora Atomic

a121d3e

Signed-off-by: Chris Evich <[email protected]>

cevich force-pushed the pr_no_7 branch from 65c8b81 to a121d3e Compare November 3, 2017 14:59

cevich closed this Nov 3, 2017

cevich deleted the pr_no_7 branch September 17, 2018 21:26

baude referenced this pull request in baude/podman Aug 31, 2019

Merge pull request #9 from squeed/move-pkg

8816992

Move libraries from cni only used by plugins.

jwhonce referenced this pull request in jwhonce/podman Nov 5, 2019

Merge pull request #9 from baude/apiv2rmi

f68b139

Add a bunch of endpoints

tyler92 mentioned this pull request Jul 20, 2022

libpod: do not lock all containers on pod rm #14976

Merged

github-actions bot added the locked - please file new issue/PR Assist humans wanting to comment on an old issue or PR with locked comments. label Sep 27, 2023

github-actions bot locked as resolved and limited conversation to collaborators Sep 27, 2023

Combination of three paprs #9

Combination of three paprs #9

Conversation

cevich commented Nov 2, 2017

rh-atomic-bot commented Nov 2, 2017

rh-atomic-bot commented Nov 2, 2017

cevich commented Nov 2, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rh-atomic-bot commented Nov 2, 2017

Choose a reason for hiding this comment

jlebon commented Nov 2, 2017

jlebon commented Nov 2, 2017

jlebon commented Nov 2, 2017

jlebon commented Nov 2, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cevich commented Nov 2, 2017

cevich commented Nov 2, 2017

rh-atomic-bot commented Nov 2, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cevich commented Nov 2, 2017

cevich commented Nov 2, 2017

cevich commented Nov 2, 2017

rh-atomic-bot commented Nov 2, 2017

rhatdan commented Nov 2, 2017

cevich commented Nov 3, 2017

cgwalters commented Nov 3, 2017

cevich commented Nov 3, 2017

cgwalters commented Nov 3, 2017

cevich commented Nov 3, 2017

rhatdan commented Nov 3, 2017

cevich commented Nov 3, 2017

cgwalters commented Nov 3, 2017

cgwalters commented Nov 3, 2017

jlebon commented Nov 3, 2017

cevich commented Nov 3, 2017