Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Pr test #6

Open
wants to merge 71 commits into
base: master
Choose a base branch
from
Open

Pr test #6

wants to merge 71 commits into from

Conversation

dperpeet
Copy link
Owner

@dperpeet dperpeet commented Nov 9, 2016

No description provided.

mvollmer and others added 30 commits November 3, 2016 11:55
Commit 29d0ca0 "networkmanager: Checkpoints" broke this by not taking
into account that `self.dev` is null for inactivate virtual interfaces
such as bridges, bonds, and vlans.

Closes cockpit-project#5293
Reviewed-by: Stef Walter <[email protected]>
The default in the D-Bus API is not the real default, and we now need
to set that flag explicitly.  See

    https://developer.gnome.org/NetworkManager/unstable/nm-settings.html#id-1.2.6.4.26

Setting the REORDER_HEADERS flag seems to be important in practice, see

    https://bugzilla.redhat.com/show_bug.cgi?id=1390605

Old versions NetworkManager used to always set the reorder flag by
accident, see

    https://bugzilla.gnome.org/show_bug.cgi?id=762626#c18

Closes cockpit-project#5294
Reviewed-by: Stef Walter <[email protected]>
Varying these make the generated files far less predictable than
they should be, including order of text, and HTML encoding.

Resulting patches between tarballs become far too large.

Closes cockpit-project#5296
Reviewed-by: Peter <[email protected]>
If the sections have randomized ids like 'idm47845730136032'
then they look like differences when patches are generated
for commits by release-source.

Closes cockpit-project#5307
Reviewed-by: Peter <[email protected]>
The various values in the arrays were not properly sorted, which
led to unneeded patches being created by release-source.

In addition I've tried to make the files more readable and
have cleaner patches when changes *are* needed.

Closes cockpit-project#5305
Reviewed-by: Peter <[email protected]>
Suggested by coverity

Closes cockpit-project#5310
Reviewed-by: Stef Walter <[email protected]>
This particularly happens on user journal files.

Closes cockpit-project#5303
Reviewed-by: Dominik Perpeet <[email protected]>
This is a development time dependency, not a build time
dependency.
Now that we're fully moved to webpack we can have a more
traditional model for our build time dependencies, and solve
several problems in the process.

 1. Having a bower_components/ directory makes the project
    behave in a more standard way for other javascript developers.

 2. Populating bower_components/ at autogen.sh time rather than
    committing the changes allows the project to behave as
    expected for packagers who want to understand dependencies.

 3. A separate bower_components/ directory lets us more simply
    build a cockpit-deps-122.tar.gz style tarball ... although
    that's for a later commit.

 4. Moving our own common code into pkg/lib allows us to run
    normal unit tests against it rather than having all such
    tests in the playground. Again implemented in a later commit.
The main tarball should be buildable without any of the dependencies.
If the dist/ directory remains intact.

The devel tarball is a replacement for 'npm install && bower update'
This way if you clear out the dist directory you'll be sure
to get a clean build. Goal everything pregenerated in the tarball
is distributed in the dist directory.
And fix up how the guide is distributed and built.
This is now possible because we store our shared stuff in pkg/lib
with all the other webpack stuff.
The semaphore script should fail on the first command that
fails. In addition it should run 'make distcheck' paralellized
because otherwise webpack takes too long.
Many released kubernetes implementations don't yet support
the 'Width' or 'Height' style stuff on the container terminal
websocket. So lets filter that stuff out for now.

The kubernetes-container-terminal project sends it, but we have
a chance to intercept it. It breaks tests.
Add an Ubuntu testing image for version 16.04 of Ubuntu
and update and disable tests that are applicable or need
tweaking to make them work.

Closes cockpit-project#5285
Reviewed-by: Marius Vollmer <[email protected]>
We don't yet build the cockpit-kubernetes stuff there.

Reviewed-by: Marius Vollmer <[email protected]>
This isn't used anywhere.

Reviewed-by: Marius Vollmer <[email protected]>
This makes debugging much easier, when we can see the output
either in the browser or in the test output.

Reviewed-by: Marius Vollmer <[email protected]>
Certain versions of useradd don't automatically make the home
directory when running useradd.

Reviewed-by: Marius Vollmer <[email protected]>
mvollmer and others added 29 commits November 7, 2016 16:11
libssh is too old.

Reviewed-by: Stef Walter <[email protected]>
UDisks2 doesn't have it.

Reviewed-by: Stef Walter <[email protected]>
This fixes an build issue with parallel build processes under load.

Closes cockpit-project#5319
Reviewed-by: Peter <[email protected]>
Now that we have multiple outputs from 'make dist' the
DIST_ARCHIVES should report that, so that tools such as
cockpituous can pull in the right stuff.

Closes cockpit-project#5320
Reviewed-by: Peter <[email protected]>
This is so we can include patches in the RPM.

In addition clean up some of the other aspects of the source
RPM. release-srpm now just sets the Version: and Release:
fields appropriately.

Lastly the comment after %changelog remains when merging the
changelog manually between various spec files and dist-git,
so lets move it.

Closes cockpit-project#5297
Reviewed-by: Peter <[email protected]>
This changes the networkmanager dialogs to display their errors
according to the patterns. Even though this just starts to use
the right CSS, and doesn't update the javascript to use common
dialog code.

Closes cockpit-project#5317
We no longer test or enable atomicapp related code. So
lets remove this known issue for now. It hasn't happened
in a long time.

Closes cockpit-project#3968
This was fixed two months ago. Lets remove it now.

Closes cockpit-project#5316
This needs to merge before an commit that starts adding the
image so it can be tested on our testing infrastructure.

Closes cockpit-project#5323
This allows us to copy the files into the same image and then
examine the core dump futher.

The motivation here is the storaged crashes on CentOS 7.

Reviewed-by: Dominik Perpeet <[email protected]>
This allows us to download just the core dumps on failure.

Updated various images. Skipped updating Fedora 25 itself since
SELinux completely breaks this functionality there and this
will need further work when known issue is fixed.

Reviewed-by: Dominik Perpeet <[email protected]>
This is catching way too many bugs across various operating
systems. The logs are inconsistent, as are the tests that
it matches.

In addition it has no upstream bug report, therefore the
GitHub issue is closed.
Closes cockpit-project#2938

Closes cockpit-project#5246
Reviewed-by: Dominik Perpeet <[email protected]>
Why are we skipping these on RHEL?
The wildcard skips cause us to ignore test failures long after
the functionality starts to work on the target operating system.

We should explicitly skip tests on specific operating systems.

In addition no tests should be skipped on canary images, such as
fedora-testing.

Closes cockpit-project#5322
So that we can import it and use it elsewhere. This fixes the
error case handling in check-multi-os.

Closes cockpit-project#5328
Reviewed-by: Dominik Perpeet <[email protected]>
The verbosity at this level is not needed. It doesn't happen
with any other part of 'make dist'

Reviewed-by: Dominik Perpeet <[email protected]>
This is because we already have a dist directory, and
in our wrapper makefile we have to mark 'DIST' as phony
target.

https://www.gnu.org/software/make/manual/html_node/Phony-Targets.html

Reviewed-by: Dominik Perpeet <[email protected]>
Several changes necessary here:
 * Stop doing a gc on temporary git repo we won't need anymore
 * Remove the temp git repo so tools/git-version-gen isn't confused.
 * Stop warning about whitespace in patches.

Reviewed-by: Dominik Perpeet <[email protected]>
This allows us to override a NODE_ENV=production from the environment
and have our minimization webpack logic react appropriately.

Reviewed-by: Dominik Perpeet <[email protected]>
Name the compiled webpack javascript files with a .min.js extension
when they are minimized. This allows the patching logic to identify
them appropriately. It also makes it more transparent what's going on.

Reviewed-by: Dominik Perpeet <[email protected]>
This is generated by webpack-make. It makes the deps file
be patched, distributed and updated when any of its dependencies
change.

This allows release-source generated patches to work correctly.

Reviewed-by: Dominik Perpeet <[email protected]>
Add a bit of documentation. Work better with the automake related
tools/missing code. And fix up the usage.

Closes cockpit-project#5330
Reviewed-by: Dominik Perpeet <[email protected]>
We symlink to this opportunistically instead.
@arilivigni
Copy link
Collaborator

aos-ci-test

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants