Workflow
Under construction.
Install and configure an Java IDE
Jenkins use Java Properties and Jelly files to implemented internationalization, and all translations mus be stored in files that uses ISO-8859-1 encoding and any characters that are not supported by it must be converted to Java entities.
Use your preferred IDE configured to receive UTF-8 as input and automatically convert to ISO-8859-1 with Java entities: that you save you from a lot of headaches, leaving you to worry about normally editing the files.
Instructions to setup that will vary depending on the IDE, but here are some references:
The IDEA has a good chance to include a spell checking tool, which is a great addition to the process.
Run jtt for an overview
Executed with only the --lang option, jtt will only check all available
translation files and compare those in English with those available (or not) in
the selected language.
This will give you an overview what needs to be done.
My suggestion is to always started with the removal of deprecated keys with the
--remove option, so you start cleaning up instead of having to mess with
unnecessary keys text.
Use Git
Check with Git (git status) what changes are being proposed and follow up
from there.
Once a translated file is done, add it with git add. I strongly suggest to
work in each file individually so you can get a better control of progress.
New files
For new files, you will need to open them in the IDE and start the translation.
Deprecated keys
Review each file that had keys removed, comparing them with the original. It is
possible that you will find lines that were removed because jtt identified
them as a empty key (they probably were left overs, without being a well formed
property).
Removed file
Nothing really to be done in this case, except accept the change with git add.
Compile the Jenkins source code
It's a good idea to recompile the Jenkins source code after changing the properties, just to be sure everything is fine.
While is always possible to make it with your preferred IDE, using the Maven CLI might be the easiest way.
First download Maven from https://maven.apache.org/download.cgi.
Then unpack it and follow the instructions detailed in the README.txt file.
Once you finish the setup, just move to the Jenkins checkout out Git repository and execute:
mvn -am -pl war,bom -DskipTests -Dspotbugs.skip -Dspotless.check.skip clean install
Issues with NodeJS
You might get errors from not having NodeJS (and yarn) available to compile Jenkins frontend. You should install those manually before trying to compile.
NodeJS version 18 or higher is required for that.
If you're using older Linux distribution versions, you might need to compile it yourself, as described here.
Using a virtual machine to compile Jenkins
Giving the bad management of versions and packages for NodeJS, I decided that using a virtual machine to build Jenkins was easier than fixing the mess that NodeJS environment is.
You can find in the root of this repository a Vagrantfile that can be used
with Vagrant.
In order to use it, it is required to have both Vagrant and Virtualbox installed.
Once that is done, it is just a matter to enter with vagrant up in a shell
and once the VM is ready, enter with vagrant ssh, git clone the Jenkins
repository and start compiling.
Tips
Working in a shell
If you're using a UNIX shell (Bash, Korn, etc), you can use programs like find
and grep to search files that requires some fixes.
Note
Since version 0.03, jtt includes the --search option to do something like
that.
For example, let's suppose I know there are remaining "build" English words spread over Brazilian Portuguese translation properties, so I can do:
$ find core/src -type f -name '*pt_BR.properties' | xargs grep -F build -l -i > tmp.txt
That creates the text file tmp.txt (that I won't commit to the repository)
which I can use as a "working queue", since now I will need to review each file.
The next sequence of commands will "pop out" the first filename to STDOUT
from the file, allowing me to open the file in the IDE, edit it, then move
to the next file in the "queue".
head -1 tmp.txt && sed -i -e '1d' tmp.txt