* Refactor: use fs.WalkDir to copy from FS
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Refactor: rename and added doc comment
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Refactor: clean up templates
Make `fn.Template` interface not struct.
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Clean up: mask manifest.yaml from template FS
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Refactor: receiver is not pointer
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Refactor: use const instead of string literal
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: TemplatesPath defaulting at correct place
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: mask manifest.yaml in template code
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: remove unnecessary else
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: put docstring on interface
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: test calls better API
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: check return value
Signed-off-by: Matej Vasek <mvasek@redhat.com>
Use the latest Knative release in the hack/allocate.sh script.
In addition, there is a small change to how the munged yaml files
are piped through `yq`, adding a final pass through `yq` itself
essentially sanitizing the data.
Signed-off-by: Lance Ball <lball@redhat.com>
* chore: add version.txt to .gitattributes for style
The Knative style check uses linguist to determine if a given file is a
generated file. If so, it excludes that file from the check. This change
should prevent Knative's style check from complaining about no newline on
version.txt
Signed-off-by: Lance Ball <lball@redhat.com>
* fixup: also add zz_filessytem_generated.go
Signed-off-by: Lance Ball <lball@redhat.com>
* docs: restructure & add a language pack "contract"
The docs/guides/langugage-packs.md document does a good job of describing
what's required of a language pack in terms of directory structure and the
options for the `manifest.yaml` file. But it does a pretty crappy job of
describing the whole thing in context. I've tried to do that here.
Also some fixes to links and reorganization of the docs to make things a
little easier to find.
Signed-off-by: Lance Ball <lball@redhat.com>
* fixup: inline language pack referenc into contract
Signed-off-by: Lance Ball <lball@redhat.com>
* fixup: minor wording tweak
Signed-off-by: Lance Ball <lball@redhat.com>
* fixup: whitespace
Signed-off-by: Lance Ball <lball@redhat.com>
* fix: add func_darwin_arm64 to the release artifacts
Signed-off-by: Lance Ball <lball@redhat.com>
* chore: add ARM64 release target/propose process
I noticed in the most recent release, that a MacOS M1 binary was not
produced. This commit adds that target to `hack/release.sh`. Additionally
the CHANGELOG.md files and version.txt files have been updated to accurately
reflect the current status. Finally, I've proposed a release process in
the new file `docs/RELEASING.md`.
Signed-off-by: Lance Ball <lball@redhat.com>
* fixup: remove trailing whitespace
Signed-off-by: Lance Ball <lball@redhat.com>
* Improve test error output
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Template ZIP generator close file after use
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* update root and version structure and help text
* fix: limit openshift int test with tag
* refactor: commands to use simplifed, unified constructor
* fix ineffectual assignment lint error
* cleanup
* add repository to run command
* callout for forthcoming s2i builder impl
* lint errors
* re-add the deferred client factory
* remove setNamespaceFlag now that it is persistent
* avoid side-effect of global-mutating deploy tests
* reduce line-by-line difference for PR ease
* simplificaiton of tests and comment lines for further PR ease purposes
* reduce inconsequential differences for ease of PR
* tests to RootCommandConfig
* review comment updates
* fix lint errors
* replace stdlib Setenv in tests
Using t.Setenv will require an update to go1.17, which is out of scope
for this PR.
* pass ClientFactory throughout
* explicitly empty test command args
See https://github.com/spf13/cobra/pull/155
Errors can still be encountered when, for example, using precomiled
tests. Explicitly setting constructed command args to the empty slice
ensures we avoid hitting any futher edge cases.
* Support using 'func invoke' for springboot templates
- add README notes for using 'func invoke' to springboot templates
- update ce-type used for cloudevents template tests
* Re-generate zz_filesystem_generated.go after rebase
* Improve template ZIP generation
Use only 755 and 644 mod permission.
The exec bit is really the only one that we care about.
Also other bits may be set differently on developer machines depending
on its configuration, this would affect generation reproducibility.
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Add explanatory comment
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Recreate git repo from scratch
Recreated the repo by `git init --bare` and commited template there all
over again.
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Enable previously disabled tests for Windows
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* Use git HTTP protocol not file protocol in tests
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* fixup: cleanup
Signed-off-by: Matej Vasek <mvasek@redhat.com>
* doc: add documentation for using podman remotely
This commit adds more documentation for using podman, specifically in this
case, connecting to and using a remote podman service from a Mac computer.
Signed-off-by: Lance Ball <lball@redhat.com>
* fixup
Signed-off-by: Lance Ball <lball@redhat.com>
* updating to latest spring cloud release and refactoring code
* new health endpoints
* implementing simplified echo function in template
* adding banner
* adding NL at EOF
* updating native to release
* fixing trailing whitespace
* adding invocation back
* adding Newline
* whitespaces
* fix lint and regenerate
* regenerate zz filesystem after rebase
Matej Vasek