* Fix locking in new Alpine release container
Seems like the new Alpine container (3.15) has some issue when doing apk upgrade first which ends up erroring out the process, so moving it to the last part of the process (after installing the packages) makes the trick.
You can test it out by using the current process (errors with a "Unable to lock database: temporary error (try again later)") and the one in this branch which will finish successfully

* Switch to Temurin official image

This image is a pure musl alpine image

NODE_ENV interferes with other tools (e.g. `yarn` won't pull
devDependencies). Since what we need is only a flag indicating Webpack
to bundle for production vs developement, invent our own env.

This also solves the mystery of unable to call `yarn build` in the build
script `bin/build-mb/src/build.clj`.

* Separate steps for faster build

* Fix typo

* Fix EE Dockerhub for once and for all

* Try to overload the ARG attribute

* Kick the CI

* Port code for building i18n resources to Clojure build scripts; remove dep on gettext

* Use Metabase classloader

* Test fixes 



* Bump adoptopenjdk version

* Delete gettext dependency from Dockerfile

* Update developers-guide.md

* Address PR feedback

* Fix frontend singular msgstr format

Co-authored-by: Luis Paolini <paoliniluis@gmail.com>

* Add container build for metabase-enterprise-head

* Add build argument

* Delete script to build images

* Update Dockerfile for parallel and 20% faster builds

* Changes proposed by @ariya

* Trim packages and update

- Removed the make and wget dependency
- Made the release container to auto-update
- Adds are now a curl in a single command to reduce intermediate images
- consecutive run commands into 1 line
- ENVs to 1 line to reduce intermediate images

* Update the docker file for the releases

* Dockerfile Improvements

* Fix

* Mega build script improvements

* Look for project.clj to determine root directory since we delete .git on CI for speed

* If .git folder is missing (i.e., in CI) fail gracefully

* Don't need to generate version file to run Cypress tests, right?

* Build uberjar step can cache version.properties
(so the Cypress test steps don't need to generate it again)

* Flush caches

* Update uberjar cache key

* Change font registration to run on first Pulse PNG render rather than at launch.
    * This will improve startup times slightly . 
    * This was causing DockerHub builds for `metabase-head` to fail 
because the stage 1 "builder" image did not have `ttf-dejavu` and `fontconfig` installed (the builder was running in to #12223).
     * This will resolve issues discussed in #12223 where Metabase fails to launch when running on a "headless" JVM Docker image (i.e., one without AWT libraries) or when we run into the infamous font issues. We still won't be able to render Pulses, but Metabase without Pulses is slightly preferable to no Metabase at all
*  Improved error messages & logging when Pulse rendering fails. Unfortunately, if the AWT classes are unavailable/broken in the JVM there is no way we can render Pulses given our current reliance on it. The best we can do is explain the situation and point people in the right direction. The new error message explains the situation and points people to #7986 for more details and workarounds.  As such, I am going to close #7986 because we cannot fix the underlying issue itself so the best we do is handle it better.
*  Refactor tests in `metabase.api.pulse-test`
*  Add tests for `GET /api/pulse/preview_card/:id`
*  Include entire Exception chain in 500 responses
*  Fix font registration logic not cleaning up `InputStream`s when done with them, wasting memory
*  Minor Dockerfile tweaks

* Change font registration to run on first Pulse PNG render rather than at launch.
    * This will improve startup times slightly . 
    * This was causing DockerHub builds for `metabase-head` to fail 
because the stage 1 "builder" image did not have `ttf-dejavu` and `fontconfig` installed (the builder was running in to #12223).
     * This will resolve issues discussed in #12223 where Metabase fails to launch when running on a "headless" JVM Docker image (i.e., one without AWT libraries) or when we run into the infamous font issues. We still won't be able to render Pulses, but Metabase without Pulses is slightly preferable to no Metabase at all
*  Improved error messages & logging when Pulse rendering fails. Unfortunately, if the AWT classes are unavailable/broken in the JVM there is no way we can render Pulses given our current reliance on it. The best we can do is explain the situation and point people in the right direction. The new error message explains the situation and points people to #7986 for more details and workarounds.  As such, I am going to close #7986 because we cannot fix the underlying issue itself so the best we do is handle it better.
*  Refactor tests in `metabase.api.pulse-test`
*  Add tests for `GET /api/pulse/preview_card/:id`
*  Include entire Exception chain in 500 responses
*  Fix font registration logic not cleaning up `InputStream`s when done with them, wasting memory
*  Minor Dockerfile tweaks

* Bump ring-core to 1.8.0

* Fix Docker CI build

https://github.com/technomancy/leiningen/issues/2658#issuecomment-593716154

Using version 220 of `jest-localstorage-mock` required that we upgrade
our version of NodeJS from our previous minimum of 6.11.3. To support
that, this commit switches from the deprecated
`java:openjdk-8-jdk-alpine` docker image to the newer
`openjdk:8-jdk-alpine` image. This image is running Alpine version 3.8 which
has NodeJS 8.11.4 in it's repository. This also bumps the minimum
version of node in the packages.json file.