Saturday, February 16, 2019

Easy Enterprise Org SFDX and CLI

Auntie Pat Tern started taking Step Dancing classes.  "Break it down", is all she said.

Your new tools can coexist with existing User Interface options.
So I decided to break down my last post: Easy SFDX for Enterprise and Others to prove just how easy it is with 6 simple steps for application lifecycle management with SFDX for the Enterprise.

The Command Line Interface (CLI) for SFDX can be accessed via menu options in Visual Studio Code, or through the terminal, or through a Windows Command Prompt. So I'm going to use the CLI commands in the steps for the most clarity.  And I'm assuming you've enabled the Dev Hub from your Production Org via the Setup options in the Salesforce UI.

Step 1

SFDX force:org:list
Shows which orgs you are connected. You want to connect to your sandbox orgs to manage your development lifecycle, for example: developer, integration, staging.

Step 2

SFDX force:auth:web:login -r https://test.salesforce.com -a orgAlias -s
Connect a sandbox where you will find and update source code and configuration. 
Text in italics will need to be specified by you when you run, and should apply to your work.
-r is followed by the login URL for the org, using test subdomain for sandboxes and login subdomain for your Production org. Be very careful when connecting to Production, test changes before you deploy them, of course.
-a is followed by an alias you are giving to this org connection.  Something like 'Dev' for a developer sandbox maybe? 'Int' for an integration sandbox. You see where I'm going here.
-s indicates that this should be the default org and username for your current work and is important to remember. You can also check the force:org:list to make sure you know which org you are working in at all times.

Step 3

SFDX force:mdapi:retrieve -p somePackage -r r### -u orgAlias -s
Bring code and configuration from your org down to your local computer in a package.xml file. Text in italics will need to be specified by you when you run, and should apply to your work.
-p is followed by a package name, either outbound change set or package defined inside a Salesforce org. 
-r is followed by the directory name on your local computer where the package will be saved.
-u is followed by the alias for the connected org where your package/outbound change set was defined.
 -s indicates a single package structure for the xml file.

The code and configuration should be defined in your existing org using either Setup>Package Manager to automatically include all dependencies and get a ton of related metadata or using Setup>Outbound Change Set just to get the specific code and/or metadata that is changing.  

Packages are super helpful and can be pre-defined in Production so that it is easy to see the full extent of code and customization dependencies related to your custom features. These pre-defined packages carry over to sandboxes you create for super easy org management. Don't upload it, we just want the definition of what it contains, and we want to keep that open and changeable. 

Creating and retrieving an outbound change set in the sandbox where you are starting development gives you a more manageable set of metadata and code, just the bits you need right now. Keep the change set open, don't define any deployment connections, we just want the definition of what it contains and we want to be able to add and remove if necessary.

And of course you can create code locally with commands like force:lightning:component:create and save or deploy those creations to your development sandbox for testing.

I should mention that these outbound change sets and in-org package definitions are all available with Workbench for retrieve and deploy as well, just in case your admins and devs love a familiar UI.

Step 4

Unzip
Using your favorite tools, unzip the "unpackaged.zip" file you just saved to your specified local directory.  Maybe you use a favorite application for zipping and unzipping, maybe jar xf from the command prompt. 

You can now work with this code and configuration in any local app you choose, maybe VS Code. Make sure to save your changes back to your sandbox org. Run automated tests in the org using your favorite tools, check out Setup>Custom Code>Execute Tests.

Step 5

Backup with version control
Save to your favorite repository for backup and version control. You can run git commands from the command prompt or you can literally just drag and drop with cloud-based interfaces for GitHub and BitBucket. You can also connect tools like VS Code to your repository to help you do this automatically.

Step 6

SFDX force:mdapi:deploy -c -d r### -u orgAlias -w 100 -l RunSpecifiedTests -r testa, testb
Deploy changes to the next sandbox as you move your work through the various lifecycle stages. Text in italics will need to be specified by you when you run, and should apply to your work.

-c optionally check the deployment before removing this flag and deploying for keeps.
-d is followed by the local directory name where you have your work and package file (hey, it was '-r' before, but now it is not). You can even do a quick retrieve from sandbox to make sure you are deploying the latest version you tested in the sandbox org.
-u is followed by the alias you have given to the destination org you set up previously.  Not the org you made changes in and tested in, but the org you are deploying changes to for further testing.
-w is followed by a wait time and is used for especially long running deployments.
-l RunSpecifiedTests optionally lets you exclude long running or unrelated tests, otherwise expect all your tests to run in the destination org, even bulk tests. Combine this with the -r option below:
-r is followed by a list of test classes.  Keep your bulk tests in their own classes and separate from unit tests so you don't have to run them during production deployments if you don't have to, for example.


Done. You can look at the Deployment Status right in the destination org, using Setup options you already know.

Four super helpful SFDX CLI commands can really get help you break down your development lifecycle and use new tools where they are most helpful while keeping all the familiar Saleforce UI options to support your orgs and your users and to maintain those valuable audit records for changes in your Enterprise org.




Saturday, January 19, 2019

Easy SFDX for Enterprise and Others

When cousin Tim Toady took the doorknob and lock apart on the back door after his key broke, he had a lot of trouble figuring out how to put it back together.  With a hundred tiny pieces laid out across the kitchen, Auntie Pat Tern sifted out a dozen of the most important, handed it to cousin Tim and told him to just use that for now.  Unsatisfied, Tim tried to put a few more parts back in, triggered the spring and the entire thing exploded across the kitchen again.That's when Auntie Pat Tern reminded him of two guiding principles: 1) Keep It So Simple and 2) Keep track of all your changes.  He could have taped each piece to a sheet of paper, in order, as he dismantled the lock, for example.

That reminded me of the guiding principles behind implementing Salesforce Developer Experience for the Enterprise Org.  Staying Agile means finding the Minimum Viable Product for tools you implement, not just products you build, so here is my guide to easy SFDX for the Enterprise. (See part 2 of this post for the actual commands.)

SFDX MVP


1. Discover:  A little reading reveals that SFDX is intended to "enable source-driven development, team collaboration with governance, and new levels of agility".  To accomplish this, it offers the important benefit of using "developer tools you already know." Use what you know, that's easy.

One thing we know in our Enterprise orgs is that configuration changes may happen in Production to meet immediate release requirements.  So the Production org itself is always the absolute source of truth for the org.  You can't just implement source control and claim that the source of truth is there because the bugs your users report will be found in Production and hopefully easily reproduced in Sandbox, that easily created org copy that is a developer tool you already know.

So the first step is to use the developer tools you already know: Production is source of truth for code and metadata and Sandboxes provide great individual and team test environments.  Keep them all. Use Dev Hub in Production to enable command line logins to these orgs.

2. Develop: SFDX includes a Command Line Interface (CLI) that lets you connect to your Enterprise Production org as well as your Sandboxes using simple and batchable commands. That gives you the power to make use of local tools, including local code editing and version control by connecting the CLI to your orgs and pulling down the code to edit locally.

Different Developer Sandboxes and different bits of code on the local computer make keeping track of all the versions and making sure they are merged and integrated into the larger system for testing means that version control is more important than ever.

For version control, we use Git on the local computer and cloud-based systems like GitHub for team backups and version control. Our local git directory can be connected to our own personal branch (or maybe to a feature branch) of the org repository in GitHub. Following tech review, developer branches can be merged into the main code branch. And this can be automated so that a developer's work gets backed up from their local computers to the cloud repo on a regular basis.

You already know about backups and version control if you use something like Microsoft Word.  It periodically auto-saves the working version of your documents. If your computer crashes, you have a choice between loading a version you saved or a version that Word automatically saved.  Version control to the rescue!  That can definitely be useful when you are doing Test Driven Development and your test fails because of your last set of changes and you decide to roll those back to the previous successful version.

If you use Git for version control of your work on your computer, you can keep track of versions.  And you can post those versions to GitHub, BitBucket or similar.  These tools can make team collaboration easier with everyone checking their latest working version into the repository in the cloud. It's like the difference between Word on your local computer and Quip Documents in the cloud.  Collaboration can be simplified with cloud solutions. And you can create a process for branches that get merged into a single main repository based on your own team's organization.

We have different repositories for different platforms we develop on.  And developers are free to create their own repositories for their own organization as well.  One developer keeps every feature in separate branches and another developer keeps just a local repo and a cloud repo, no branches. For our org-wide repository, we have branches based on releases so that short-term and long-term projects are kept separate.

The CLI adds flexibility to our Development Cycle by letting individuals choose tools they are familiar with and develop code locally or continue to develop in the Dev Console even, either way we can create a single shared backup and version control repository of everyone's work thanks to simple CLI commands.

3. Retrieve and Deploy: Following our SFDX process of using what you already know, we can easily coordinate the efforts of admins who make configuration changes in Sandbox into our version control as well as developers who want to edit code with apps they run on their own computer.  Salesforce provides two easy tools to work with the CLI: Packages and Outbound Change Sets.  If you haven't tried both, take a look in Setup>Home for them.

Packages give you the chance to define related sets of custom code and custom configuration metadata. Choose a bit of code and dependency data like custom objects and fields are automatically added. You can create Packages to look for overlap of functional dependencies in your org. And to help you document feature specifications.

Outbound Change Sets give you a bit more control to stick to only code and configuration that has immediately changed, and not include more stable code and configuration from prior development that is already in Production. You can use these to get the most up-to-date configuration changes on the fly from production for anything you need to work on.

Both Packages and Outbound Change Sets can be pulled out of the org and onto the local computer using the CLI retrieve command. These can also be deployed to any other org (yes, including Production) using the CLI deploy command.  Retrieve pulls down a Zip file including a package.xml defining what has been retrieved and folders of metadata, all the same for packages or change sets. You can unzip this into your local Git directory to update your version and you can use the CLI to deploy back to any org after you've made changes. In between, backup ongoing work to shared repositories to merge changes for the next stage of the development cycle.

Sticking with these familiar tools, we not only keep it simple, but we ensure that Admins and Devs can work well together because the org remains the source of truth and work efforts can be tracked, changes can be audited all in the org as usual.

4. Iterate: Since we have the ability to connect to our familiar Enterprise orgs as well as the ability to work on changes locally and use tools for searching and comparing code and configuration versions, we create a stronger collaboration.  Since we can use Git commands on our local computer to manage the backup and version control, we have even better insights along the development cycle, regardless of Sandbox refreshes.

With changes happening in a variety of orgs, local repositories may need to rebase if a branch gets too far our of sync. For Enterprise orgs, the source of truth is what your users see when they report bugs and praise new features. So bringing your local repository back into sync with upstream changes is simplified.

If you have separate repositories for each development stage from Integration to UAT, Staging, and Production, developers can retrieve changes at each point in the process. This makes it even easier to compare versions when a feature breaks after an upstream deployment.  Verifying whether the correct version was deployed with all of its component parts can be done locally, in shared repositories, and by auditing recent deployments right in the orgs.  And reverting to another version of a feature is even easier when you have backups in version control at each step.

Benefits of Change


SFDX helps Enterprise orgs follow Auntie Pat Tern's advice: Keep It So Simple and keep track of all your changes.  By allowing admins and devs both to work with tools most familiar to them, at the same time, SFDX offers greater insight and transparency into the development cycle. Because the CLI opens up opportunities to make use of tools like VS Code and online repositories, debugging throughout the release processes may be simplified as well.  And the CLI gives the power to make everything more automated for a more continuous integration of changes into each new environment at every stage of Enterprise development.

Ultimately, the best implementation of the Salesforce Developer Experience is to make use of what you know and implement it your way, just like the SFDX Developer Guide recommends. Give yourself even more power with command line deployment and retrieval of packages pre-defined in your orgs through easy and familiar Salesforce Setup options. Get even more peace of mind with backups at each stage of the release process thanks to the CLI and Git local repositories connected to shared repositories online.