In a typical day, a FileMaker user is interested in storing information in a custom app and then reporting out on that data in many different ways. A record, at the very heart of FileMaker, accomplishes this need. Here, data is entered and stored, and from this object, information is retrieved. One way data is entered is through FileMaker Drag and Drop, and therein lies a problem.

The Life Cycle

A record has a life cycle. It is created, edited, and committed to the file. Sometimes it is deleted. This cycle happens over and over. A record may go through many edits and commits in its life.

FileMaker drag and drop

A record’s life cycle

A primary job of FileMaker developers is to control that record life cycle. We set up strict controls about when a record can be created, edited, and deleted. We build logic to ensure a record has the correct type of information, and we work to ensure that the record is committed with integrity.

Like a vampire or a walker, however, there’s a user experience piece of FileMaker that is completely outside this life cycle and disrupts how a record is manipulated and then committed. That is the Drag and Drop.

FileMaker Drag and Drop: the Gotchas

FileMaker drag and drop is a great feature for fast data entry. This feature is commonly used for container fields and PDFs or pictures, but it can also be used to drag text data to a field. And that action is the disruption. When dragging data from another source into FileMaker, a user is almost completely subverting the record’s carefully structured life cycle, and thus the unexpected can happen.

Here’s a possible scenario: The user drags a text block from a website or email. She wants to save the text of an article or the current price of an item. The user simply highlights the text on the web page and clicks on it. Holding the mouse down, she drags the text over to the open empty and inactive field on the layout. Finally, she lets the mouse go and the text fills the field.

To the user, all is good, but under the hood, all sorts of unexpected things are happening:

The record is never opened. Or is it?

Normally when I type in a field, the record opens. FileMaker holds this record open for all edits. This doesn’t happen in a FileMaker drag and drop scenario. The data viewer, when watching the function Get(RecordOpenState), does not update from zero. Practically it may have opened, but the problem is that we cannot get to the open state of the record.

The record is never committed.

Well, I take that back a bit. Yes. It is committed because the data is stored in the field and record. Like the open state, we cannot get to the moment before the record is committed.

The modification count goes up

The result of the function Get ( RecordModificationCount) in the data viewer, increments by one each time.

Object script triggers do not run at all or properly

Mislav Kos and Wim Decorte at Soliant Consulting have a great rundown of all the script triggers and when they happen. The script trigger sequence, however it breaks down in a drag/drop scenario:

  • onObjectModify does run, but it runs after the record has already been committed (or the record has been closed).
  • The rest of the triggers do not run at all, no matter if they were meant to be a pre- or a post- trigger.

These facts present many problems for the developer. We are not able to control the life cycle of the record. We don’t get to capture when a record is open or before it is committed. Our lack of control over these means we cannot rely on script triggers to check field data for validation or for anything else we might do.

The Solutions

We can’t allow users to enter data without going through the proper business logic. So I can think of two solutions: either disallow drag and drop or highly control when a user can enter data into a record by any means they wish.

Disallow it

The first method is easier said than done. FileMaker Drag and Drop a preference in the app on each user’s machine. It is NOT specific to the file. So a developer or IT person would have to turn this off for every user. And of course, it can be turned on again in the preferences menu by a savvy user.

The Drag & Drop functionality can be turned off via the app preferences.

It would be awful to have to update this for each user

Disable it

The second method holds more promise: As a developer I might want to control when a user can enter data. I would have a separate data-entry layout and scripting to open a record and then check to make sure the record can be locked. When finished, the user would click a button that runs a script that checks various validations and allows the record to be committed or causes it to be reverted. Controlling the life cycle of the record means I could allow drag and drop.

Disallow it (Again)

Going back to method one, there is a way to disallow FileMaker Drag and Drop for all fields in a table without resorting to visiting each machine to turn off the app preference. The technique consists of a nondescript little field (text or number type) in the table where we want to prevent inadvertent drag and drop. I named my field “zz_ValidationField” so it sorts to the bottom of the field list. I’ve learned this from wise people: sort fields that do not contain data to the bottom.

On this field, I added validation in this manner:

  • The “Allow user to override during data entry” is unchecked.
  • The Validation can happen either “Always” or “Only during data entry”.
A validation of the field

Data Validation for my zz_ValidationField

  • Validation by calculation is used. Inside this is the function Get(RecordOpenState).
Validate using Get ( RecordOpenState)

The field’s calculation

  • Uncheck the “Validate only if field has been modified” checkbox at the bottom left.

Explanation

This setup simply prevents drag and drop from happening when the record is not previously open.  This field’s validation will fail if the record is not open ( Get ( RecordOpenState) = 0 ) and thus will not let the data be saved. The custom message doesn’t come up since the record wasn’t open in the first place.

Conversely, if the user opens the record  (Get (RecordOpenState) > 0 ) via scripting or clicking in any field and then drags and drops, the validation will allow the drag-drop to happen as normal.

As FileMaker developers, we have to control every aspect of our custom app so that we can ensure data integrity. The FileMaker Drag and Drop experience, dragging text blocks or PDFs gets around our control unless we take an extra step to ensure that users don’t get carte blanche in adding data to the file.

Download Demo File

It happens. You’re moving quickly, jammin’ on your favorite FileMaker custom app. Laying down some insane custom business logic in the form of one awesome, totally badass FileMaker Script. Then you notice that other script. The one you need to get rid of. No time like the present. So you pop over to it, hit “delete”, crush the popup confirmation before you have time to read it and… Whoops, something went very wrong!

NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!

You deleted the wrong script.  Your awesome, badass script with the insane logic is nothing but a memory. There is no undo. It’s gone. It’s not in the backup because you were just working on it.  It’s too new.  You are totally f#%@ed! You weep, you rage, you even try mediation. Nothing works. The script is still gone.

Then you remember. “Wait I just made a fresh DDR for FMPerception“. A smile creeps across your face. Yeah, you’re gonna be OK.  FMPerception is going to save your butt.

True Story

The above scenario happened to me this week. I deleted the wrong script. For a moment, it was bad. But then I remembered that FMPerception had the script. I could easily restore it, and reconnect it the other scripts that were using it. Just 2 minutes later I was done.

Phew!  Watch the video above for exactly how to recover a deleted script using FMPerception.

Need For Speed

FMPerception is so fast that it doesn’t get in the way of your workflow. As a result, you’ll use it all the time. You will be much more likely to have a recent analysis than if you are using an analysis tool that forces you to wait minutes or hours. Most FMPerception users know all about that, but what they might not know is that your analysis is also a backup, and can be used to restore code that you deleted or broken. And that can save your butt!

 

 

With the release of FileMaker 16 and native JSON functions, the ability to create a testable solution has been vastly improved. In this blog post, I will walk you through a simple example of what testing is, and how it can improve your solutions. Most developers have had situations where they need to alter some existing logic, and are concerned that they might alter something with unforeseen consequences. Wouldn’t it be nice if you could ensure that your changes didn’t break existing code? Welcome to FileMaker Testing.

Automated Software Testing

Software testing has a lot of different variations and buzzwords associated with it. You’ll hear all sorts of phrases and acronyms thrown about like TDD, BDD, Unit Testing, Regression Testing, etc. It’s worth a learning a little something about these at some point. But at the heart of all of them is a simple concept; write some code to automatically test some other code.  If you do, your code will be less fragile, and easier to maintain.

What is FileMaker Testing?

A test is a FileMaker just script used to ensure that other scripts are working properly. Tests ensure that scripts are not only accomplishing their goal but also producing the desired result when they fail. While there are many ways to perform FileMaker testing, we test using the following assumptions:

  1. Each script to be tested accepts & returns JSON as a parameter
  2. We will be using custom functions for handling errors, testing JSON objects, and validating a JSON payload

Writing Testable Scripts

To prepare your script that needs to be tested, it should always return JSON to the script that called it. This means that if the script fails, it should return an error JSON object. If it succeeds, it should return an object representing the data that was changed. See the example script called “Create a Project” in FileToTest. You will notice that it returns verbose errors on all failures, and returns a JSON object when successful.

Writing The Test Script

To prepare your test script, you need to do three things in karbon_tester:

  1. Create a JSON Object to pass the script using the layout “Edit Test Data”
  2. Create a script in the Folder “Project Testing” (Copy the existing scripts…you’ll get the idea!)
  3. Add the script you created to the script “Test – Project”

Do We Test Every Script?

There is no reason to test each & every script. We only write tests for those scripts that play a critical role in the solution. Once you are comfortable with testing, it will become clear what you should test. It’s up to you.

Benefits of FileMaker Testing

The benefits of creating test scripts are compelling. First, it provides any solution with a level of stability that is difficult to achieve in a timely fashion. If multiple developers are collaborating on the same solution, the test file enables everyone to run common tests on core business logic, and see if they have produced any unforeseen fatal errors.

Second, it provides a level of clarity for you, the developer. Once you begin writing tests, you will take on an alternate persona, where you are actively poking holes in your own code. This may seem like a small difference, but we believe you’ll be surprised once you try it. You become much more critical, which improves your focus & clarity.

Third, the readability of your code will improve. One thing that makes JSON difficult to handle is that there is not a repository for JSON objects within a FileMaker solution. When a script is designed to handle records in multiple tables, karbon_tester serves as a clear, contextual example of how to create & pass JSON objects to work within a solution.

Is it Worth It?

The first question many people seem to ask when introduced to testing is “is this worth it?” Our answer would be that it depends on the solution. If you have a complex system, multiple developers, and a need for stable solutions, it’s not just worth it, it should be the standard.

Download

To start, download our sample files here. Open “FileToTest”, and create a project. Review the script “Create a Project” to see how the file works. Then open the file “Karbon_Tester” and run the existing test. See how easy that was? Now you should create your own scripts to test creating a status.

What About Editing Data & Managing Test Data?

We will be releasing part 2 in this series in the future. We will walk through options for loading test data, & managing records that are used only for testing. The testing in this sample is intentionally simple to grasp and does not confront some of the nuanced issues involved in testing.

What about Karbon?

We know some have you have been waiting for us to release some of internal tools and frameworks that we group under something we call “Karbon”.  This is the piece piece of that puzzle.  More coming soon.

Looping through and filtering lists of values turns out to be a common task in FileMaker.  Having a few different techniques at your disposal is quite useful. In this post, we are going to look at a set of related techniques for processing lists.  First we look at how to loop through each item and perform some function or task on each item.  Then we will move on to a custom function for filtering lists by any valid FileMaker expression.  Finally we will use another Custom Function to easily parse Layout Object Names from a layout and use them in a dynamic slide navigation scheme.

Read more

FileMaker’s GetField() function has always been struck me as a curious beast.  Every since it’s release back in Filemaker 5.5 or there abouts I had assumed that it returned the contents of the field specified by the parameter.  But there were a couple of cases where it didn’t work that way and I’d get fooled every once in while.  I didn’t quite understand what was going on.  I think I finally stumbled on an easy way to keep it straight. Read more

Unit Testing is the practice of  testing individual chunks of code to make sure they are working as designed.  Ideally it is done with a single click of a button, so that you can do it all the time.  Traditional FileMaker coding doesn’t lend itself to this practice very well.  It’s just too tightly coupled.  But if you take some time to organize your code in a modular way you can start to incorporate something like Unit Tests into your development practice.

First you need to separate out the logic you want to test into well factored API like scripts.  These can take parameters and they usually return a result.  The main point is that you give these scripts the power to do what the need to do to your database, and you make sure that they report back if they fail or not.

Once you have your API scripts laid out you can the write yourself a bunch of test scripts that exercise those API script and make sure they are doing their job.  I put my Unit Tests in a separate file so I don’t clutter up my deliverable files. Then as I go along writing code, I periodically go back and run all the scripts. When I do something that cause my tests to fail, I can see it right away.  I know what test broke and hopefully I get some kind of error that might help tell me how it broke. From there, its relatively easy to go about fixing the code.

Here is a movie that shows what that ends up looking like.