This directory contains support files that are used when working on the website. Files here are not published on the website.
We have configuration in static/.htaccess that occasionally we need to update. Our preview environment is hosted on the build server (Jenkins) and for the local preview Hugo is acting as a web server. So with that there is no way to test changes to static/.htaccess file before it is applied to the production website. For this we have a httpd.conf file here that mimics the production environment locally. In conjunction with running HTTP server in a Linux container it can be used to test changes to static/.htaccess.
Use like this from the root of the git repository after building the website (yarn build):
$ docker run --rm -p 80:80 -p 443:443 \ -v "$PWD/public":/usr/local/apache2/htdocs/:Z \ -v "$PWD/support/http":/support:Z \ httpd:2.4 /bin/bash -c "cp /support/* /usr/local/apache2/conf/ && httpd-foreground"
When running the build within a container for some checks we might need to run more containers. For example the tests/redirect.sh script when run with -s will start Apache HTTP in a container. A common solution to that is a “Docker in Docker”, i.e. mounting /var/run/docker.socket and /usr/bin/docker in the container. This has two drawbacks, one is that the libc version of the build container might not be compatible with the libc version of the docker CLI installed on host, now mounted within the build container; and another is that when running with additional restrictions, e.g. user-mapping or SELinux, accessing /var/run/docker.socket from the build container's namespace could lead to permission errors.
To circumvent that the docker-pipe.sh script creates several named FIFO special files (see mkfifo(1)) that can be mounted to the build container, along with the docker stub script that uses those named pipes to communicate with the docker-pipe.sh running on the host.
Typical usage scenario is something like:
$ support/docker-pipe.sh tmp & $ docker run -v "$PWD/tmp":"$PWD/tmp" \ -v "$PWD/tmp/docker":/usr/bin/docker" \ ... root@62ace315af17:...# docker run ... root@62ace315af17:...# exit $ tmp/teardown.sh
Some notes on the usage: the argument provided to docker-pipe.sh is a directory where named pipes docker stub script and teardown.sh scripts will be placed. This directory must be mounted to the same path within the container, as in the above example. One can use docker stub script from that directory or, as in the above example, mount it to /usr/bin/docker so it's immediately available on $PATH. docker-pipe.sh should be run in the background, and to terminate it use teardown.sh, one could use job control commands as well, but that might be a bit more involved in CI environments (hence teardown.sh). Upon exiting docker-pipe.sh will delete the files it created from the provided directory.