|author||Jens Geyer <email@example.com>||Tue Mar 09 23:39:20 2021 +0100|
|committer||Jens Geyer <firstname.lastname@example.org>||Tue Mar 09 23:39:20 2021 +0100|
trigger build to reflect THRIFT-5332 Question: list all the reserved words in thrift doc
If you are using an OS packaged version of Ruby, you may also need to install the ruby-dev (Ubuntu) or ruby-devel (Fedora) package as well to build any native code for gems that are installed later. Installing these will also ensure your system's RubyGems package is installed. Depending on your OS, you may also need other packages to install/build gems, such as ruby-full, make, gcc, nodejs, build-essentials, or patch.
Once Ruby, RubyGems, and any necessary native tools are installed, you are ready to install Bundler to manage the remaining RubyGem dependencies. Bundler is included in Ruby 2.6 and later as a default gem, so installing it may not be needed.
Because we use Bundler to install specific versions of gems, it is not recommended to use an OS packaged version of gems other than what comes built-in. If you are using an OS packaged version of Ruby, it is strongly recommended to avoid
sudo when installing additional gems, in order to avoid conflicting with your system's package-managed installation. Instead, you can specify a
GEM_HOME directory for installing gems locally in your home directory. You can do this in your
$HOME/.bashrc file or other appropriate place for your environment:
# in .bashrc export GEM_HOME=$HOME/.gem/ruby
With Ruby installed on your machine, you can install Bundler using the command below:
# not necessary in Ruby >2.6, since it is a default gem since 2.6 gem install bundler
Next, use Bundler to install Jekyll and other dependencies needed to run the website (this command assumes your current working directory is your clone of this repository with the
main branch checked out, because that's where the Gemfile dependency list exists).
The command to serve the site contents using Jekyll‘s built-in webserver is as follows (this webserver may behave differently than apache.org’s servers).
bundle exec jekyll serve -w
You do NOT need to execute a
bundle exec jekyll build command first, as the
serve command is sufficient to both build the site and serve its contents. By default, it will also try to re-build any pages you change while running the webserver, which can be quite useful if trying to get some CSS or HTML styled “just right”.
Jekyll will print a local URL where the site can be viewed (usually, http://0.0.0.0:4000/).
Changes pushed to our
main branch will automatically trigger Jekyll to build our site from that branch and push the result to our
asf-staging branch, where they will be served on our default staging site.
First, add our repository as a remote in your local clone, if you haven't already done so (these commands assume the name of that remote is ‘upstream’).
git clone https://github.com/<yourusername>/thrift-website cd thrift-website git remote add upstream https://github.com/apache/thrift-website
Next, publish the staging site to production by updating the
asf-site branch to match the contents in the
# Step 0: stay in main branch; you never need to switch git checkout main # Step 1: update your upstream remote git remote update upstream # Step 2: push upstream/asf-staging to upstream/asf-site # run next command with --dry-run first to see what it will do without making changes git push upstream upstream/asf-staging:asf-site
A convenience script can be found that performs these steps for you, after asking which remote you want to use. It is located in the
main branch at
Note that Step 2 should always be a fast-forward merge. That is, there should never be any reason to force-push it if everything is done correctly. If extra commits are ever added to
asf-site that are not present in
asf-staging, then those branches will need to be sync'd back up in order to continue avoiding force pushes.
The final site can be viewed here.