ResultV2 update result contains Result (original), which provides
information in terms of files, objects and images, and also includes
file changes which provides information in terms of files, objects and
changes. The changes contain the old value, new value and the setter
that was involved in the update.
ResultV2 can be used to obtain detailed information about updates in
terms of the previous value and the new value, which could be beneficial
in commit messages to summarize the granular changes.
Signed-off-by: Sunny <github@darkowlzz.space>
The package setters2 has been removed from kustomize:
https://github.com/kubernetes-sigs/kustomize/pull/5291
This commit removes the need to import setters2, by reproducing the
last link with that code, which is a minor parsing helper.
I have not changed the comment explaining what was changed from the
original, since it's still accurate. The parsing func is a buried
detail.
Signed-off-by: Michael Bridgen <mikeb@squaremobius.net>
TrimRight was not correctly used. It takes a set of characters as the
second argument and any chars matching the set on the right side of the
string would be trimmed. TrimSuffix does exactly what I originally
intended.
Signed-off-by: Aurel Canciu <aurelcanciu@gmail.com>
This bumps the version of the image reflector types to v1beta1. This
doesn't technically make a difference for the minute, since the
Kubernetes API server would convert between versions, but is tidier.
Signed-off-by: Michael Bridgen <michael@weave.works>
This does the following:
- copies the type definitions from v1alpha2 to v1beta1
- changes the "stored" version to v1beta1
- gives the CRD a conversion strategy of None, meaning just rewrite the version
- switches the controller to use v1beta1
- moves the generated documentation to v1beta1
This effectively rebadges the v1alpha2 version of this part of the
image API to v1beta1. The v1alpha2 version is left in place; there are
no conversion issues, as with v1alpha1->v1alpha2. The CRD specifies
that converting between v1alpha2 and v1beta1 just means changing the
version (i.e., the schema and semantics are the same).
Signed-off-by: Michael Bridgen <michael@weave.works>
This commit finesses the use of the debug log a little, and introduces
a trace log. The trace log gets threaded through calls to utility
procedures -- it's a little awkward putting loggers into func
parameters and structs, but it always is.
Signed-off-by: Michael Bridgen <michael@weave.works>
The image name was incorrectly extracted from the given reference,
yielding a trimmed version of the string and breaking the resulting
replacement value.
Signed-off-by: Aurel Canciu <aurelcanciu@gmail.com>
The update procedure is obliged to return a Result struct with all the
objects that were changed, for filling in the commit template. At
present, the result is collated by running each setter on each object
and seeing if the setter is used. This uses the `Set` from kyaml, with
a small amount of glue.
It doesn't quite work, however, because a setter may be used for a
field without changing the value. The result gets an entry for each
policy _mentioned_, whether or not it had a new value. There is no way
to see whether a setter actually changed a field from the outside,
other than by comparing a copy of the object before using the setter
with the object after (which yaml.v3 does not make easy).
A better approach is to get the setter to record whether it changed
anything, since it is there doing the changing. This means
reimplementing kyaml's `Set`. I have stripped it down to the parts
needed for image updates -- so e.g., only field values are examined.
Signed-off-by: Michael Bridgen <michael@weave.works>
This explains the data available to the commit message template in the
API guide. While writing it, I realised it could be made more
convenient, so:
- mask external types by embedding them
- make the most useful parts of an image ref available using a
wrapper struct and interface
Signed-off-by: Michael Bridgen <michael@weave.works>
It's desirable (see #6) to be able to enumerate the updates that were
made by automation, in the commit message and perhaps in an event
announcing success.
Doing this is counter-intuitively difficult. A `kyaml.setters2.Set`
filter will keep a count of the times its used. Previously, one `Set`
was used with the `SetAll` flag set, which would replace any marker
that corresponded to an image, in one traversal. But to keep track of
images individually, you need to have a setter for _each_ image (and
its tag, and its name, since those can be used separately). This means
`3 x policies` traversals of each node! The saving grace, possibly, is
that only files with a marker in them are considered.
Since you might want to dice the results in different ways, the result
returned is a nested map of file->object->image.
Signed-off-by: Michael Bridgen <michael@weave.works>
This contains a number of fixes, in particular for an indentation bug
that seems to have been introduced not long ago, and which would
otherwise cause problems after updating controller-runtime (which
brings a slightly less recent kyaml with it).
This also comes with a nice enhancement: it's no longer necessary to
use the global schema for kio setters, you can just build a schema and
use that. No need to serialise access to the global schema. Yay!
Signed-off-by: Michael Bridgen <michael@weave.works>
In general a copyright notice takes the form:
Copyright year name
.. where the year is the year of first publication, to let people know
from when the copyright applies. It's fairly common in software to
affix additional years in which the software was modified and
released. I have chosen here to use `2020, 2021` for the new and
modified files; it is OK that not _all_ files are updated, since the
important bit is the _first_ year, which they already have.
Signed-off-by: Michael Bridgen <michael@weave.works>
This is intended to address two problems:
- LocalPackage{Reader,Writer} like to reformat the YAML that passes
through them; mostly this is harmless, but occasionally it will end
up fighting format tooling, e.g., prettier.
- It's possible that things like Helm chart templates are lying
around in the git repository to which automation is applied. Those
templates have extensions of ".yaml" but are not usually parseable
as YAML, so would result in errors from the file reader.
This commit changes how updates are run -- firstly, it screens files
by checking for a token (`"$imagepolicy"`) that will be present in
files that might need updating. This cheaply removes some nodes --
likely including Helm chart templates -- from consideration.
Secondly, it now only writes files that were actually updated by an
imagepolicy setter, rather than writing everything that was an
input. This means it's less likely to reformat something that doesn't
need to be touched at all.
Signed-off-by: Michael Bridgen <michael@weave.works>
It will be useful, for kustomizations e.g., to be able to set just the
tag or just the name (repository). This commit adds setters for those
to the schema -- they have the name of the image setter plus a suffix
of `:tag` or `:name`. For example:
newName: ubuntu # {"$imagepolicy": "ns:policy:name"}
newTag: 18.10 # {"$imagepolicy": "ns:policy:tag"}
This adds another means of updating files to the package pkg/update/,
in setters.go (and gives the existing file a better name).
In passing, I changed the test util for comparing before/after
updates, in pkg/files/, to give a little more context when comparing
file contents; and, since the comparison between actual and expected
is not symmetrical, I corrected the order of the args in the tests.
Matheus Pimenta
Sunny
Michael Bridgen
Sunny
Aurel Canciu
Michael Bridgen
Stefan Prodan