feat!: remove gherkin, add events support

Signed-off-by: Todd Baert <[email protected]>

.gitignore

@@ -1,2 +1,3 @@
 .idea
-node_modules
+node_modules
+flags/changing-flag.json

Dockerfile

@@ -1,6 +1,10 @@
-FROM ghcr.io/open-feature/flagd:v0.5.2
+FROM ghcr.io/open-feature/flagd:v0.6.3 as flagd
 
-COPY testing-flags.json testing-flags.json
+FROM busybox:1.36
+
+COPY --from=flagd /flagd-build /flagd
+COPY flags/* .
+COPY wrapper.sh .
 LABEL org.opencontainers.image.source = "https://github.com/open-feature/test-harness"
 
-ENTRYPOINT ["/flagd", "start", "-f", "file:testing-flags.json"]
+ENTRYPOINT ["sh", "wrapper.sh"]

README.md

@@ -1,18 +1,9 @@
 # OpenFeature Test Harness
 
-This repository contains the makings of an end-to-end test suite for OpenFeature SDKs.
+This repository contains a docker image to support the [gherkin suites](https://github.com/open-feature/spec/blob/main/specification/appendix-b-gherkin-suites.md) in the OpenFeature specification.
 
 ## flagd-testbed container
 
 The _flagd-testbed_ container is a docker image built on flagd, which essentially just adds a simple set of flags for the purposes of testing OpenFeature SDKs.
-
-## Gherkin test suite
-
-The test suite is a set of [_gherkin_](https://cucumber.io/docs/gherkin/) tests that define expected behavior associated with the flags defined in the flagd-testbed. Combined with the _flagd-provider_ for that SDK and the flagd-testbed, these comprise a complete end-to-end tests suite.
-
-### Lint Gherkin files
-
-The Gherkin files structure can be linted using [gherkin-lint](https://github.com/vsiakka/gherkin-lint). The following commands require Node.js 10 or later.
-
-1. npm install
-1. npm run gherkin-lint
+`testing-flags.json` contains a set of flags consistent with the [evaluation feature](https://github.com/open-feature/spec/blob/main/specification/assets/gherkin/evaluation.feature).
+`change-flag.sh` runs in the test container generates a file `changing-flag.json`, which contains a flag that changes once every seconds, allowing to easily test change events.

flags/change-flag.sh

@@ -0,0 +1,15 @@
+#!/bin/sh
+
+# this simple script toggles a flag value for a easy way of testing events
+cat ./changing-flag-foo.json > ./changing-flag.json
+
+while true
+do
+    sleep 1
+    cat ./changing-flag-foo.json > ./changing-flag.json
+    echo 'updated flag changing-flag value to "foo"'
+
+    sleep 1
+    cat ./changing-flag-bar.json > ./changing-flag.json
+    echo 'updated flag changing-flag value to "bar"'
+done

flags/changing-flag-bar.json

@@ -0,0 +1,12 @@
+{
+  "flags": {
+    "changing-flag": {
+      "state": "ENABLED",
+      "variants": {
+        "foo": "foo",
+        "bar": "bar"
+      },
+      "defaultVariant": "bar"
+    }
+  }
+}

flags/changing-flag-foo.json

@@ -0,0 +1,12 @@
+{
+  "flags": {
+    "changing-flag": {
+      "state": "ENABLED",
+      "variants": {
+        "foo": "foo",
+        "bar": "bar"
+      },
+      "defaultVariant": "foo"
+    }
+  }
+}

wrapper.sh

@@ -0,0 +1,22 @@
+#!/bin/sh
+# wrapper script to start change-flag.sh and flagd, and forward signals to the child process
+
+# handle SIGINTs and SIGTERMs so we can kill the container
+handle_term() { 
+  kill -TERM "$child" 2>/dev/null
+}
+
+handle_int() { 
+  kill -INT "$child" 2>/dev/null
+}
+
+trap handle_term SIGTERM
+trap handle_int SIGINT
+
+# start change scriptm and flagd
+sh ./change-flag.sh &
+./flagd start -f 'file:testing-flags.json' -f 'file:changing-flag.json' &
+
+# wait on flagd
+child=$! 
+wait "$child"