Since the Installomator.sh script has grown to over 5000 lines, its management on git has become very unwieldy. The single file with all the logic and the data required to download and install the applications creates constant merge conflicts which add to the workload of the repo admins, especially when part of the team is working on the logic of the script while we still get PRs to add labels.
Because of that we have split the main script into multiple files which are easier to manage. Having multiple files results in less merge conflicts.
Nothing. When you just use the Installomator.sh, you still copy its contents from the Installomator.sh script at the root of the repository into your management service (don't forget to change the DEBUG
value). Or you install the script to the clients using the installer pkg from the Releases.
The changes will only affect you when you want to build your own application labels, modify existing labels or other wise modify the script.
To simplify git merges, the script is broken into several files or 'fragments.' The individual fragments on their own are not functional or executable. When you are modifying labels or script code, you assemble and run Installomator using the utils/assemble.sh
tool. For convenience, there is a symbolic link in the root of the repository.
If you want to build your own labels or modify existing ones, follow these steps (I will be using desktoppr as an example, even though the label already exists.)
First, create a directory somewhere for your labels. For example: ~/Documents/InstallomatorLabels
.
Inside this folder, create a new file named desktoppr.sh
. The name is not really relevant, but should match the label name to make it easier to locate.
The contents for the label file should be the case statement block, like this:
desktoppr)
name="desktoppr"
type="pkg"
packageID="com.scriptingosx.desktoppr"
downloadURL=$(downloadURLFromGit "scriptingosx" "desktoppr")
appNewVersion=$(versionFromGit "scriptingosx" "desktoppr")
expectedTeamID="JME5BW3F3R"
blockingProcesses=( NONE )
;;
If you want to modify an existing label, copy the label from fragements/labels
in the Installomator repository to your directory.
When you are building or modifying labels, you will need to run the assembled script. The assemble.sh
tool will do this for you. To test our new label, assemble and run the Installomator script like this: (make sure your working directory is the Installomator repository root)
> ./assemble.sh -l ~/Documents/InstallomatorLabels desktoppr
This will put together all the fragments, including your labels in your label folder. and run the script with the desktoppr
argument. Since custom label locations are merged into the script before the default location, custom label files will override existing labels. You can add more arguments to test the label further:
> sudo ./assemble.sh -l ~/Documents/InstallomatorLabels desktoppr NOTIFY=silent DEBUG=0
# NOTE: this _installs_ desktoppr
Once you are certain that your new custom label works, you can use the code from your custom Installomator.sh script in build/Installomator.sh
, or even build an installation pkg using the --pkg
or --notarize
options.
The Installomator.sh
script at the root of the repo does not really get involved in your building and testing. Similarly, if you want to apply, test, and contribute changes to the script's logic, you should modify the fragment file in question and test using the assemble script.
Pull requests against the Installomator.sh
script in the root of the repo will be rejected.
- If you haven't already, create a fork of the Installomator repo. Clone the for to your local Mac.
- Create a new branch in your local Installomator (fork) repo.
- Copy the new or modified label file to
fragments/labels
. (replacing the original, when necessary) - Test (push the change to your fork on GitHub. You can check that out on testing devices or vms.)
- Create a pull request against the Installomator
main
branch. - Don't use this branch for any other modifications, unless you need to update this particular PR. (Pull Requests are against a branch, not a particular commit.)
If you have multiple labels (or other changes) you want to contribute, please create a separate local branch and a separate pull request for each label. This allows us to accept, modify, or reject each label separately and simplifies the process.
Once your Pull Request is merged into the main repo, you can pull the change to your local repo, push it to your fork, and delete the branch, because it should be fully merged.
When you have multiple labels or changes, please create a separate issue for each label or change, unless they are closely related
We have a tutorial on How to create Pull Requests in GitHub.
These are the fragments in the order they are assembled. All files are in the fragments
directory
- header.sh
- version.sh
- functions.sh
- arguments.sh
- (optional) labels from locations given with the
--labels
argument - labels/*.sh
- main.txt
Even though the fragment files are not functional shell scripts, we decided to use the .sh
file extension, so that Finder opens the files in the proper application and text editors recognize their file type for code display.
header.sh
contains all the 'front matter' of the script. This means all the variables that users can change together with all the comment lines explaining them.
The version.sh
file is special in that is only contains the version string and nothing else. The assemble script will use this version to create two line of code that look like this:
VERSION="0.7.0b1"
VERSIONDATE="2021-08-17"
where the version is from the file and the date is generated dynamically from the current date.
The functions.sh
fragment contains all the functions used in the script.
The arguments.sh
fragment contains the argument parsing logic and some other logic that needs to happen before the label is evaluated. This includes the start of the large case statement that evaluates the label and the three 'built-in' labels version
, longversion
, and valuesfromarguments
.
All the contents of the label files in labels
(and any custom label locations you provide with the -l
/--label
option) will be inserted here.
Finally, the main.sh
fragment contains most of the main logic.
The assemble script does not check any of the files for syntax or completeness. You are responsible that everything fits together properly. (Pay special attention to remember the closing semi-colons ;;
and a final line break in the label files.)
assemble.sh
This will put together the fragments and labels from the default location (fragments
and fragments/labels
) and write it to build/Installomator.sh
and execute it. (When you run the script without any arguments, it will print all the labels.)
assemble.sh <label>
assemnle.sh <label> <VAR=value>...
This will put together the fragments and labels from the default location, create the script in build/Installomator.sh
and immediately run it with the given arguments.
Note: the assembled script will run in debug mode, unless you specifically override this with DEBUG=0
.
assemble.sh --labels path/to/labels_dir
Text files from this directory will be added in addition to the default labels directory fragments/labels
. The custom labels will be inserted in the script before the default labels, so custom labels will override default labels. You can add multiple --labels
arguments:
assemble.sh --labels ../my_labels/test --labels ../my_labels/production
In this case the labels from ../my_labels/test
will be inserted first, then the labels from ../my_labels/production
and then the labels from fragments/labels
assemble.sh --script
This will build the full script and place it in the root of the repo.
assemble.sh --pkg
Build the full script, disable Debug mode and build a signed installer pkg.
assemble.sh --notarize
Build the full script, disable Debug mode, build a signed pkg, and send it to notarization.
There are variables at the beginning of the assemble.sh
script which you need to modify to use your certificates and Developer ID instead of mine.