You are browsing the archive for marc.

Profile photo of marc

by marc

“Code-First” Servers in RemObjects SDK 9 [BETA]

July 24, 2015 in .NET, Data Abstract, RemObjects SDK

We’ve just shipped the first betas of RemObjects SDK and Data Abstract 9.0, the next major new release of our remoting and multi-tier frameworks.

One of the big new features coming in RemObjects SDK 9, and the one I want to talk about today, is support for what we call “Code First” Servers.

What is “Code First”?

In the past, if you created a new RemObjects SDK server, you started with a RODL file, and would edit that in Service Builder to define all the services and the types that the services worked with. Behind the scenes, RO would then generate code and stub classes for your service implementation for you.

While this mode of creating servers will of course continue to be supported, “Code First” mode does away with RODL files completely — at least on the server — and lets you define everything purely in code.

You simply define your service classes and attach a couple of attributes to make them available remotely, and RemObjects SDK will take care of the rest and make sure they become callable. Similarly, any auxiliary types you define – enums, structs, etc. – you can simply define (and update) in code.

At runtime, RemObjects SDK will generate a RODL representation of your services based on the types it finds in your app and make that available to download for clients. So to the outside, a Code-First looks identical to a traditional RODL-driven server, and for the use on the client, you can still import the (now auto-generated) RODL from your IDE or the codegen tool to generate client Interface stubs.

In the current beta, Code-First servers are supported on .NET, but we plan to make the same functionality available in (newer versions of) Delphi, as well.

Let’s take a look at how this works in practice.

Converting an Existing Server

Since most of you will be starting out with existing RO servers already, let’s have a look at what is involved with converting one to the new model. Mind you, you don’t have to convert your servers to Code-First in order to use RO9. This is not a breaking change, just a new option. Still, I expect many of you will want to move to the new model, because it makes development so much easier going forward. I already converted two of my own projects.

There are really just four steps involved in making the switch, all of them simple, and two of them automated:

  1. Adjust your existing Service implementations
  2. Convert our auxiliary types, such as structs and enums
  3. Remove your RODL file and the auto-regenerating _Intf, _Invk and _Events files generated from it
  4. Add one statement of code to configure your server for Code-First

Adjusting Your Existing Services

While RO auto-generates service stubs for you from RODL, once you add code to those files, we consider them “yours”, and the RO toolchain won’t touch them. In the past, that was one of the annoyances when adding new methods to a service, as you had to edit both the RODL and the implementation.

For that same reason, we won’t auto-convert your service implementations for you, but making the switch is easy. For each of your services, you just do three things:

  • Remove the InvokerClass and ActivatorClass Parameters from the Service attribute. This is the main step, and it tells the RO infrastructure that this service is now using Code-First.
  • Remove the service interfac from the ancestor list.
  • To each of the methods you want to expose publicly, add the ServiceMethod attribute.

In essence, you’d go from something like

@RemObjects.SDK.Server.Service(Name = “LoginService”, InvokerClass = LoginService_Invoker.Type, ActivatorClass = LoginService_Activator.Type)
public class LoginService : ILoginService {

to just

@Service(Name = “LoginService”)
public class LoginService {

(That example is Silver/Swift, but the same idea applies to Oxygene, C# or even VB.)

Optionally you can also remove the StandardClassFactory attribute, as it is the default. (If you switched to a different class factory, you would want to leave the attribute there, and you can also safely leave StandardClassFactory there if you prefer.)

In future betas, the name parameter on the Service attribute will become optional too, requiring it only if you want to expose the service under a different name than the class itself.

Note that you don’t necessarily need to convert all services to be Code-First in one go. You can start by converting some of your services and deleting them manually from the RODL, and leaving others RODL-based. Of course in that case you will want to delay the next two steps until you’re ready to switch the last service over.

At runtime, RemObjects SDK will take care of merging what’s in your RODL and what’s defined “Code-First” and make sure it all works together.

Converting Auxiliary Types

Next, let’s look at all the additional types you may have defined in the RODL to help your services. Enums, Structs, Exceptions or Arrays (although Arrays defined in RODL don’t really reflect in generated code).

In code, these all lie in the _Intf file that you normally don’t touch, because it gets updated from the RODL when that changes. Since the RODL is going away, this auto-updating won’t happen anymore, so you are, in a way, “taking ownership” of these classes now.

You could just keep the _Intf file around, but there’s a ton of extra crud in there that you no longer need, and that would make it hard to maintain/update in the future. So really, it makes sense to clean that up, and we have automated that for you with a small IDE wizard that takes care of this *and* the next step. We’ll get to that in a second.

Remove the RODL and Auto-Generated Code

Once we’ve take care of the auxiliary types, the RODL file and the three source files generated from it (but not the services!) can simply be removed from the server project and deleted.

Bye bye RODL, it was nice to have known you.

The “Convert to Code-First” Wizard

To automate the above two steps, we’ve added a small helper wizard to the IDE. In the first beta, this Wizard is in Fire only, but for the next beta we’ll be adding it to Visual Studio (and eventually Delphi) as well.

In Fire, you can always right-click the RODL file in your server project to see additional tasks specific to RODL files.

If your RODL contains services that don’t have stubs generated yet, you will want to invoke the “Generate Service Implementation(s)” item first to have those generated. In fact, the conversion wizard is grayed out until you do, because otherwise you would lose those definitions.

Normally, that won’t be the case, and all your services will already be implemented. In that case, the last menu option called “Convert to Code-First” becomes active. Simply invoke that to get things going:

When you do, the IDE will do two things:

  • It will generate a new code file in your projects’s language, named after the RODL file, but with a “_RodlTypes” suffix. This file will contain all the types you need – in essence, it will be the same as the previous _Intf file, but without all there extra crud.
  • It will remove the RODL file, as well as _Intf, _Invk and _Events files from the project, and move them to the Trash (where you can still restore them from, if needed. Although needless to say you should have a backup of your project, or better yet have it in version control before doing the conversion).

Once done, you can rename the _RodlTypes file to anything you like, or even move the types in it around into different source files (for example into the same file as their related service) as you see fit for the structure of your projects.

These are now “your” types. If you need to make changes to them for new versions of our server, you simply change them in code.

Configure Code-First

As a last step, there’s two lines of code you want to add to your server. This step probably will go away in subsequent betas and before release and use sensible defaults, but for now it is necessary.

Add these lines among the startup code of your server:

if !RemObjects.SDK.Server.Configuration.IsLoaded {
RemObjects.SDK.Server.Configuration.Load(“MyProject” , “MyProject.Server” )

The first string passed is the “name” of your server, as exposed externally (in the RODL that clients will download); the second string is the namespace that RO’s Code-First engine will use for the helper types it generates under the hood at runtime.

Press “Run”

And with that, your server is converted and ready to run. When you launch and go to http://localhost:8099/bin, you will still see the familiar RODL representation of your server, only now it is driven by the code in your project, not the other way around.

Creating new Services

As you can imagine, creating new services (or a whole new server) will be super easy with the new model, as well. We don’t have Code-First-based templates in the first beta drop, but we’ll be adding them really soon.

To add a new service to your existing server, simply add a new class, attach the Service attribute, and mark the methods you want published with ServiceMethod. Done. If you need new structs or enums, simply define them in code.

To create a brand new server, just add one or more services to the project in the same way, and configure/start up an instance of your existing NetworkServer class (the API for which we’ll be making even simpler later in the RO9 beta). And of course we’ll have templates to help you do this, soon.


We’ve only scratched the surface of Code-First services in this article, but I hope it gave you a good introduction and will help you get started with the feature.

Let us know what you think, how it’s working out for you, and what you would like to see improved!

Profile photo of marc

by marc

About our Build System Infrastructure, CI2

July 21, 2015 in non-tech, RemObjects, Tools

Over the past couple of weeks, we have been working to migrate our build servers to a fully AWS-hosted solution. Our build system has become pretty sophisticated over the last few years, so I thought I would talk about it a little bit, on the off chance that some of you might find it interesting, or get ideas for your own build chain out of it.

Our build system is based on a home-grown infrastructure we call “CI2”. CI2 is a set of tools that we have written over the years that coordinate product builds and – in short – handle Continuous Integration for our products. Essentially, that means whenever someone commits a change to a project to Git, the system goes out and creates a new build of the software and eventually tests it to make sure the new change integrates ok and does not break anything.

CI2 is written in Oxygene (of course) and makes heavy use of RemObjects SDK and Data Abstract for communication.

Until a couple of weeks ago, the bulk of the system had been running on a custom Xserve server – a rack-mountable Mac server machine that Apple stopped making a few years back – that we owned, and co-hosted with a local hosting company here in Berlin. We’ve long been wanting to migrate this to a fully cloud-based hosting solution on Amazon Web Services (which we use for everything else), and when one of the disks in the sever started flaking out a couple of weeks ago, we decided that now was as good a time as any to tackle that.

With how CI2 is structured (I will talk more about that shortly), the migration was pretty smooth, in general. Most of the holdup was that we needed some more infrastructural changes. Because in the past all the builds happened on the same physical machine, we had quite a bit of logic in place that assumed the individual build VMs could access shared drives, and the like. Also, our local VMs had over the years gotten bloated, as we installed new tools, new versions of Delphi and Visual Studio required by the builds, and so on. We wanted to start with clean virtual machines in EC2, and that meant doing some well-deserved clean-up and simplifications to our build scripts – which actually makes the builds better, as well.

As of last Friday, we’re done, and have all our products build cleanly and successfully on the new infrastructure. As an added benefit, builds have gotten a lot faster, as well. It used to be that we waited a good 40 minutes for a full Elements build to finish – and that was without generating all the different set SKUs we build for releases we ship. On the new system, a complete build, with the large Shell-containing setup and everything, finishes in under 25 minutes.

Faster builds means a more productive team, as with complex products like ours, often one needs to wait for a new build from CI before a change can be fully tested.

How CI2 Works

But let’s take a quick look at how CI2 works.

As I mentioned before, CI2 is not a monolithic system, but a set off tools that work together, locally and across the network, to perform all the various tasks. It’s a very flexible and expandable system that has grown and gotten very sophisticated over the past 10 or so years (the first commit in our CI2 Git repository is from 2009. And mind you, the 2 stands for “version 2” ;)).


It all starts with CIStarter.exe. CIStarter is a very minimal tool that gets the whole CI system up and running.

CIStarter does only a couple of things:

  • Based on information read from either a local config file or (more likely, in production use) from EC2 instance data configured in AWS, it connects to an S3 bucket and downloads the actual CI2 system to the local server.
  • Optionally, it runs a tool called CIPrerequisiteInstaller to, well, install prerequisites needed for building our products. More on that below.
  • Finally, it passes things off to run one of the other CI tools or servers to do the actual work. These will usually be CIMainServer, CIBuildServer or CITestServer.

CIStarter has very little logic of its own, and the idea is that (a) only the exe (and optionally a config file) need to be deployed to a server, and more importantly that (b) the CIStarter itself will never really change, even as the core CI2 system gets upgraded.

This means CIStarter can be “burned” into a base image for a virtual machine (be it on EC2 like we use now, or any other virtualization system), and new instances can be booted up as needed. The core CI2 system – which does get regular updates and fixes, of course – can simply be updated on S3, and new instances will always use the latest system.

It’s worth noting that CI2 can run on Windows, Linux and Mac OS X – and we have it, in fact, in use on all three platforms, as you’ll see below.


The first thing CIStarted does after downloading the system from S3 is run CIPrerequisiteInstaller.exe, a small tool that installs and manages all the various packages of tools and files that might be needed to build our products. Like the system itself, these are obtained from S3, so publishing a new prereq is as simple as uploading it to the right bucket, and any new CI instances will pick it up on next run.

Of course CIPrerequisiteInstaller keeps track of which prereqs are installed, and only downloads and installs what’s missing. Prereqs are also incrementally versioned, so newer versions of a prereq can be uploaded, and get installed as needed.

Prereqs we have set up for our system include things such as Elements (we need Elements to build Elements ;), different Delphi versions, packs of third party libraries, and other tools.

One of the cool things is how we handle Delphi. As you might be aware, Delphi installers are slow and bloated, so installing Delphi is not really a good option. We need all versions of Delphi from 7 through the latest XE8 on our system, and if we were to run the installer for each, getting a new instance all prerequisited would, probably quite literally, take half a day (not to mention hundreds of gigabytes of disk space). Instead, our Delphi team figured out a way to package Delphi so that it can be xcopy-deployed with just the files we need to build with it. When a new version of Delphi (or an updated version of an existing version) comes out, we can get all the build machines updated to have it in a jiffy.

Prereq packages themselves can be made up from an .exe installer that runs quietly, or can be a .zip with a Train script that does installation or deployment. CIPrerequisiteInstaller manages unique install locations that individual prerequisites can use, assuming they don’t have a well-defined place they get installed to (for example, our Elements prereq simply runs the Elements installer, while our Delphi packages drop their files where CIPrerequisiteInstaller tells them to).

The idea is that CIPrerequisiteInstaller can both take a cleanly set up and freshly booted build machine instance and get it set up with all the prereqs we need, and it can incrementally update an existing live instance with just the new prereqs that we added.

In our case, we have both a clean AMI that we can boot up to start from scratch (which is little more than a fresh Windows install with CIStarter), but also one with the current set of prereqs already installed. Of course, instances booted up from the latter are quicker to start and ready to run a build.


Once everything is set up, CIStarter passes execution on to one of (currently) three different server tools. Which one is determined by the config file or the EC2 Instance Data.


CIMainServer is the brains of the whole operation. It’s a very lightweight server that does not need its dedicated machine (ours runs on a small Linux EC2 instance that also hosts our email, some of our websites, and other infrastructure, for example).

What CIMainServer does is keep in touch with all the actual build servers via RemObjects SDK super channel connection, and monitor the known Git repositories (some of which we host ourselves, and others we have on GitHub) for changes. As new commits are detected, it will fire off build requests to a suitable build server (i.e. one that’s idle and has the right platform and setup to build the project in question).

Currently, we only distinguish between Windows and Mac build servers, and any server can take a build request for any product on its platform. In theory, one could set up dedicated build servers per product (for example, we could have a server running that builds only Elements and not Data Abstract, if we wanted).

In the long run, CIMainServer will also be able to boot up and shut down EC2 instances as needed, depending on demand. For example, we could have a skeleton crew of one server running over the weekend (if that), but come Monday morning when things get busy and people commit changes and want their builds quickly, it could boot up a few more instances. Currently, we don’t do that, and just run two “t2.large” instances for Windows and one Mac build machine.

CIMainServer uses a small database to keep track of known repositories, active and past builds, as well as their success and testing results. That database could be hosted in RDS, but we find the extra cost isn’t justified since we don’t need that level of fail-over and recovery support, so we just run a local Postgresql DB on the same Linux machine that hosts CIMainServer itself.

CIMainServer is also CI2’s connection to the outside world.

Of course it starts builds automatically as its sees changes in Git, but it also lets us start builds manually (for example if we want a build with a non-default set of options passed to the build scripts, or when the system is locked from doing automatic builds so we can start builds in a controlled order, which we usually do for RTM and public beta builds).

For that, CIMainServer interacts with Slack to listen to commands in the chat channels where our teams hang out, and to also report back statuses. For example, I can simple say “start build elements lockdown” to shoot off a new build of “Elements” from its “lockdown” branch. When a build finishes, a message gets posted to the right channel, along with the URL where the final binaries can be downloaded, or where the log file can be found, if there was a problem.

CIMainServer also has an RO/DA based API, and we have small apps for Mac and iOS that can be used to control builds, get a more detailed status overview, and receive push notifications for critical failures.


If CIMainServer is the heart of the operation, then CIBuildServer is the muscle.

On start, CIBuildServer just connects to the main server, and then waits for instructions. When it receives a build request, it goes out to Git to clone the repository and check out the right branch, and then fires off to the build script found in the repo to build the product.

Of course, it also does a lot of maintenance around that as well, including cleaning up after a build, managing the different folders where different branches are checked out for each repository (we keep the checkout around, because a simple pull is a lot faster for subsequent builds than cloning fresh every time would be), and so on.

CIBuildServer also takes care of publishing the finished binaries and the build logs (to S3, of course), as well as keeping CIMainServer up to date with the build status.

The actual build scripts we use are all based on Train, our JavaScript-based build script engine that is already open source, but in theory CIBuildServer can also pass off to other systems (in the past we used FinalBuilder, which CI2 still supports even though we no longer use it), and the infrastructure is pluggable so that other runners can be integrated as well.

We currently run CIBuildServer for Windows in EC2 virtual machines (these are new since we switched from VMware machines that ran on our Xserve). Since Amazon doesn’t offer OS X-based virtual machines (unfortunately), we also currently run CIBuildServer on a local Mac mini. In the long run, we might switch to something like MacMiniColo. Only the “DA/Cocoa” build needs the Mac, everything else builds on the Windows machines.

The nice thing with how the whole build system is distributed now is that the location of the build servers is immaterial; the only difference we see from the Mac builds running locally is that the upload and download to and from S3 is a bit slower that it is from the EC2 instances (where S3 access is so blazingly fast we were blown away. In fact, a lot of the caching of S3 content that we built into Train and CIStarter feels unnecessary when running on EC2).

The Shared Bucket

One thing that required the most rethinking before we could move to EC2 was what we refer to as “shared” files.

While our products build independently, we do have a lot of inter-dependence between them. For example, Everwood builds as its own repository, but both Elements and RO/DA need it to build. The DA/Cocoa build needs binaries from the DA/Windows build (such as Relativity Server), and vice versa.

In the past this was handled with a shared folder that all build VMs (and the Mac) had access to, but for the move to EC2, this infrastructure had to be switched over to leverage S3.

The way we handle this is that every build that runs gets passed a dedicated S3 location where it can publish generated binaries that other projects might need access to. This location is unique for the project (the ”repository”, really), and the particular branch that builds.

For example, the Everwood build zips up the binaries needed by the other projects, and publishes them to the Shared bucket. When an Elements build runs next, it can use a predetermined logic to find the version of Everwood in S3 that most closely matches the right branch, and grab the bits from there. I say most closely matches, because there will not always be a 1:1 match, of course. We might have a “develop-feature-X” branch in our Elements repo, but no such branch in Everwood – so it will fall back to “Everwood/develop” instead.


The third and last server (currently) is CITestServer. Like CIBuildServer, it runs on its own instance (or instances, although we only run a single one currently). What CITestServer does is wait for builds to finish, and then grab the “dedicated” installer off S3, run it, and apply our suite of tests to it. As with the builds itself, statuses are reported back to CIMainServer, which can post to Slack, send push notifications, and keep track of test results and their evolution over time in the database.

By default, we run tests for the latest new build of every single-word branch (i.e. we’ll test “develop”, but not “develop-feature-x”) that’s new when the test server becomes idle. Tests can take a long time, in many cases longer than the actual product builds, so it’s not uncommon for – say – three consecutive builds to run, but only the first and the last to get tested. Testing times are also what lead us to compromise and test sub-branches such as “develop-feature-x” on demand only.

Our tests once again leverage Train, and a testing infrastructure called “Afterhours” that lets us run test suits created using various different testing frameworks (such as NUnit/XUnit on .NET, DUnit on Delphi, JUnit on Java, as well as of course our own EUNit), as well as some domain-specific test runners we have for Elements, and collect all their results in a standardized manner.


Final binaries, as well as log files from builds, tests and even prerequisite installs are all published to S3. The beauty of S3 is that it’s fairly fail-over resistant, and very accessible.

For giving our staff access to the builds as they get done, we have a few pages on the Staff portal on our website that monitor the respective S3 buckets and offer the files in them for download. This works based on the same infrastructure we use for the public and licensed downloads that are available from our website for you, the only difference being that we skip CloudFront, and just chose a S3 bucket location that’s closest to the bulk of our team, to give the fastest possible download speeds.

Publishing files for our customers (for releases, beta builds, and sometimes one-on-one to your Personal Downloads folder) is as easy as moving the files from one bucket to another (which we currently do manually as needed, but could automate in the future, if we wanted to). We can also give select customers “Fire Hose” access to all the builds that CI2 emits.

S3 offers automatic cleanup/maintenance functionality that we use to move older builds off to Glacier, and eventually delete them (except for the release builds we of course archive, we don’t really need each of the 57 Elements builds that ran yesterday for more than a couple of weeks).


So that’s our CI2 system, as it’s currently in use. We’re fairly proud of it, and very happy with how it works and what it provides for us, so it is a bit of a shame that right now it is purely internal and not many people outside of RemObjects get to see (or use) it – even if you do get to see and use the products it generates.

Who knows, maybe we will change that one day and make CI2 available for general use, be it as open source or as product. The main thing stopping that is that while the system is very flexible and configurable, it does still have a lot of “RemObjects-isms” – little bits and pieces that are made to work exactly how we need it, but aren’t precisely how everyone else would want it. And of course, it being used internally only, it can be pretty rough to set up and configure in places (for example, we add new products/repositories so seldomly that we have no tool for it, but simply add a new row to the database manually ;)).

But whether CI2 will ever see the light outside of our virtual halls or not, I hope you found this an interesting read, a good look behind the scenes of how we work here at RemObjects. And maybe it gave you a couple of ideas for your own build infrastructure.

At the very least, make sure to check out Train, our awesome build script engine. It’s open source and completely free – as is the .NET JavaScript engine that drives it.

Let me known what you think!


Profile photo of marc

by marc

Announcing Elements 8.1.85

July 16, 2015 in Elements

I’m happy to inform you that we have just released a new update to Elements, version 8.1.85.

As the version number implies, this release is a minor update over the previous Elements 8.1 from this Spring (the second one), but it is our most significant yet, with literally over 500 commits and fixes and improvements all across the tool chain.

We’ve been working on this release in parallel to the next big version (more on that below) and really focused on stability and bug fixes. We hope you will find 8.1.85 incredibly solid.

The release is available now, as trial and for registered users, and of course covers all three languages – Oxygene, C# and Swift – and includes Visual Studio, as well as significant improvements to Fire.

What’s Next?

We’re not slowing down, of course. The team has been hard at work on new features and improvements for the next major version, and at the end of this week, we’re planning on making the first beta of Elements 8.2 available. Like 8.1, Elements 8.2 will be a major new release, with significant new features and fast improvements. Already, we have over 1200 commits with new stuff over 8.1.85 that will become available in beta on Friday.

Changes in Elements 8.2 will include an update to Silver to support Swift 2.0, application templates for Windows 10 UAP, support for new Cocoa SDK features, such as ”kindof” and nullability info from Objective-C, language enhancements across the board, and much more.

Get It Now

As always, Elements 8.1.85 is a free update for all users with an active subscription (and of course Silver remains to be free for everybody). Having an active subscription also ensures that you will have access to the 8.2 Beta – so if your sub has lapsed, now is the perfect time to go and renew.

We hope you’ll enjoy the 8.1.85 release and that we’ll see you on the Elements 8.2 beta.

marc hoffman
Chief Architect

Profile photo of marc

by marc

WWDC 2015 and Elements

June 15, 2015 in "Silver", Fire, iOS, Mac, Nougat, watchOS, WWDC, Xcode

It’s been another exciting year at WWDC, even though i personally did not make it out to San Francisco this time. But staying home also has its upsides, as it means I had lots of time to dive into what’s actually been released.

As always there are many things new and coming in the Apple ecosystem that affect our Elements compiler, and with that Oxygene, C#, Silver and of course Fire. Let’s have a look.

New SDKs

Like every year, WWDC brings new editions of OS X and iOS, namely version 10.11 and 9.0. This time around, these new SDKs add not just new functionality (of which there is plentiful), but Apple also made some pretty drastic changes to the API headers, mostly in service of better inter-operation with Swift.

Because most of the headers use new Objective-C language features, such as annotations of nullability and (limited) generics, the new SDKs do not cleanly import with our current FXGen tool. After all, FXGen cannot know about new Objective-C syntaxes that just shipped.

But worry not. Myself and the team have been hard at work this week to update everything, so with the new Elements beta build we will post today (and the new official 8.1 update release coming next week), the new SDKs now import fine and are fully usable – including all the new frameworks and APIs. So get coding!

We’ll continue to work on the import so that for the next release (Elements 8.2, which goes into beta soon), we’ll be able to leverage a lot of the new information these SDKs headers now expose, which will allow us to represent Cocoa APIs even better in Silver, C# and Oxygene. For example, APIs will reflect the more accurate nullability information that’s now available, making them easier to deal with – especially from Silver.


Now, in addition to new iOS and OS X SDKs, Apple also shipped a third, brand new platform SDK: watchOS. On the SDK level, watchOS is very similar to the existing SDKs for Mac and iPhone – it’s just a bunch if frameworks.

In fact as part of the new work on the importer, the watchOS 2.0 SDK is already importing fine, and our compiler is already happily building against it. But of course fully integrating Watch support is more involved – there are many parts of the toolchain and the IDEs that we need to review and expand. Working on this will be a high priority over there next couple of months as watchOS gores thru the betas, and we plan on shipping watchOS support in Elements 8.2 in the Fall (and of course make it available incrementally in the betas, prior to that).

Swift 2.0

The third and last big thing to mention is of course Swift. At WWDC, Apple introduced Swift 2.0, which brings many cool (and some awkward) improvements to the language. Like with version 1.0, there’s a lot tom like here, but also some things that have us scratching our heads in terms of syntax choices ;). But in general, we’re very happy with how Swift is evolving.

Of course we already started on bringing Silver, ur implementation of Swift for .NET, Java and Cocoa, up to speed with the latest language changes. Some parts were easy to do, while others require some more thought.

For example, Apple added ”Error Handling” to Swift, but it is very different from exception handling as it is needed on .NET and Java (and supported in Silver via a Language Extension). We have some cool ideas hewn to integrate the two in a way that’s convenient and intuitive, but we’re still fleshing out the details.

We’ll be working on adding Swift 2.0 support over the course of the next couple of months, as part of Elements 8.2 (and remember that Swift 2.0 itself is in beta right now as well, and bound to change more between now and the time Xcode 7 ships. In fact, in there WWDC sessions Apple was already talking about features not in the current beta release yet).

We’ll update a page on our documentation site with progress on Swift 2.0 support as things move forward.


Let us know what you think. And make sure to check out Elements, if you have not already!

Profile photo of marc

by marc

Announcing Data Abstract and RemObjects SDK 8.3

May 27, 2015 in "Silver", Data Abstract, Delphi, Visual Studio

We’re happy to announce that we just shipped Data Abstract and RemObjects SDK 8.3, a set of significant updates to our flagship emoting and multi-tier frameworks.


Probably the most exciting change for Delphi users in this release is the inclusion of support for the “NEXTGEN” Delphi compiler, and with that support for building FireMonkey applications for iOS and Android in Delphi. We know that many of you felt strongly about being able to use Delphi to build apps for these platforms, and with DA8.3 you now can.

This support of course comes in parallel with our existing support for building apps for these platforms with out native libraries for Cocoa and for Java, respectively. Our platform-native libraries will continue to be a major focus point, and our recommended solution for building great mobile apps.


Version 8.3 also adds official support for using RO/DA in combination with our new a (and free) Silver compiler that we shipped last month. You can now use Data Abstract and RemObjects SDK to build clients and servers using the Swift language. Servers can be created for .NET/Mono, and clients are of course supported for .NET/Mono, Cocoa and Java/Android.

This is a change that should be exciting booth existing RO/DA users who want to use Swift, as well as to existing Swift developers looking to write server code to accompany their client apps.


This release also adds official support for Visual Studio 2015 (Release Candidate, or later), and includes tentative support for Fire, our new native development environment for Mac.

You can use DA with all three Elements languages (Oxygene, C# and Swift) in Fire, and in Visual Studio it of course supports Visual C# and Visual Basic, as well.


Version 8.3 is available now, both as full versions for licensed users, and as free trial. As always, this is a free update for all users with an active subscription.


Profile photo of marc

by marc

A Small Update

May 17, 2015 in "Silver", Data Abstract, Elements, Fire


This just to let you know that a couple of days ago we shipped a small interim update to Elements 8.1, which was originally released at the end of last month. This update, versioned, focuses on fixing a few minor but annoying issues in Elements (such as an endless reboot loop when installing the Visual Studio Shell on some older Windows systems) and a couple of Silver bugs, but adds no major new features to the core product.

Fire has also seen a few bugfixes as well as some small enhancements, such as support for the Force Touch feature on new MacBooks to invoke Peek at Definition (just force-press on an identifier to see its definition – very handy). As always, check out the full change log for there full scoop.

In other news, we’re still preparing our 8.3 update for RemObjects SDK and Data Abstract. A release candidate is available as Gamma now, and it looks like those bits will ship virtually unchanged, this coming week. New features in 8.3 include support for the “nextgen” Delphi/Mobile compiler, as well as for Silver. Stay tuned for the RTM.


Profile photo of marc

by marc

Announcing Elements 8.1, with Swift and Fire

April 30, 2015 in .NET, "Silver", Cocoa, Elements, Fire, Java

Elements 8.1

We are absolutely thrilled to announce the immediate availability of Elements 8.1, the next major version of our Elements compiler with Oxygene and RemObjects C#.

Don’t let the .1 version number trick you, Elements 8.1 is a significant and major update that brings a wide range of features and improvements, building on our 8.0 release late last year. From support for Visual Studio 2015 to IDE-integrated Help, from new language features in Oxygene to full support for C# 6.0 syntax, from Android GUI designer integration to iOS Extension templates, there’s bound to be something new and exciting for everybody.

But of course the most significant new feature in Elements 8.1 is Silver, our implementation of Apple’s Swift programming language.

With Silver, Swift joins the ranks of Oxygene and C# as a third Elements language, and is supported across all three platforms, and with all the bells and whistles you have come to expect from Elements. What’s more, we have decided to make Silver completely free to use for everyone — it will be included for free in your active Elements subscription, and new users who are interested only in Swift can use Silver without requiring an Elements license at all.

The second major new thing shipping with Elements 8.1 — and very close to my heart personally, because it represents the last years of my work – is Fire. Fire is our new native development environment for the Mac, written and designed from the ground up for Elements around our ideas of what a modern and lightweight IDE should look like in 2015. Fire is designed to be fast and nimble, yet powerful.

Fire is not quite ready for the “1.0” moniker yet, so the current version is still considered a preview – but it’s a production stable preview that you should be able to use for your day-to-day work (I personally have been working exclusively in Fire since the beginning of 2014 – that’s 16 months now). Even though still in preview, Fire is available as free public trial download now, and also supports the free Silver compiler – in addition to, of course, Oxygene and C#.

Get Elements 8.1 now

Elements 8.1 is available for download now – both for Windows with Visual Studio, and for Mac inside Fire. You can grab your copy at


marc hoffman
Chief Architect,
RemObjects Software

Profile photo of marc

by marc

An Update on RemObjects Silver

March 31, 2015 in "Silver"


I wanted to take a brief moment to give you an update on the status of Silver, our Swift compiler for .NET, Java/Android and Cocoa.

First of all, thank you again for trying out Silver. Interest in the compiler has been tremendous, with over 1600 people using the beta now. We could not be more thrilled.


Things have really been coming together over the last weeks and the past few weekly beta drops – the beta we shipped last Friday (build .1727) is really solid.

We’re ready for the next step, so this coming week we are planning to promote the current codebase to our lockdown/gamma branches. That means we’ll start preparing to ship based on this codebase, allowing only the most critical and stability-focused fixes to go into the branch between now and when we officially ship “Elements 8.1” (which will include updates to Oxygene and C#, as well as the first official “RTM” release of Silver).

We expect the next build you see to be a Gamma build, and hope to have the “master” Elements 8.1 release with you and all Elements users before the end of April.

Of course, even as we lock down for 8.1, work is already ongoing for the next releases as well, and we’ll make beta versions of that available, too.

Help Us Make Silver a Reality

If you like Silver and what we are doing, and would like to support our efforts, I would like to once again ask you to consider helping us out with a small (or large, if you prefer) financial contribution via the links here, or by purchasing a full Elements 8 license (which will give you all the great features of Silver using the C# and Oxygene languages, as well) here.

Silver is and will be free to use at no charge by everybody – that includes the command line compiler and the full IDE experience in Visual Studio and Fire. Your contributions help us finance the continued development of Silver and all of Elements, and enable us to make it available for free to everybody who wants it.

Thanx Again!

So, thank you again for your interest in Silver, and for taking the time to read this. We really appreciate your thoughts and feedback on Silver (and all of Elements), so please do not hesitate to contact us and let us now, via email or on the Talk forums.

Like you, we are very much looking forward to and excited about the upcoming release.

marc hoffman
Chief Architect,
RemObjects Software

Profile photo of marc

by marc

Oxygene for Delphi Developers: New Features

January 28, 2015 in Elements, Oxygene

Oxygene has come so far from where Delphi left the Pascal language when it stopped innovating in the late ’90s that it’s hard to provide a simple and concise overview of what’s “new” in Oxygene for Delphi developers – there’s just so much.

This topic will try to provide brief sections on most of the major improvements and new language features that Oxygene provides, covering them with a brief introduction and then linking off to the main language documentation where you can find more details.

Pretty much all of these features, with the one exception of Generics, will be new to you, whether you are coming from Delphi 7 or a more recent Delphi version such as XE7, because the Delphi language really hasn’t changed much over the past 15 years.


While pretty much all code in Oxygene lives inside types, this first section looks at new kinds of types that Oxygene introduces (such as tuples and sequences), and fundamentally new things you can do with types (such as nullability). Let’s get started.


Sequences are a special type that exist on a similar level to Arrays, and can also be thought of as a collection of elements.

Different from arrays, sequences do not imply a specific form of data storage, but can represent any collection of elements that is accessible in a specific order. This could be an array (and as a matter of fact, all arrays can be treated as a sequence) or a different data store.

One major advantage of working with data via sequences is that your code can start to work on the first items of a sequence before the entire sequence has been generated. Your code might even stop working the sequence at some point, and the remainder of it never will be generated. This allows for some pretty powerful use. For example, you can query a large database table, and have rows fetched as you need them. You could even have an infinite sequence that generates all the digits of Pi, and choose to process only the first 10,000.

Sequences are defined with the sequence of keyword combo, parallel to arrays:

var lCustomers: sequence of Customer;

Sequences can be looped over with the regular for each loops that recent versions of Delphi have also introduced, and they also work great with LINQ and Oxygene’s from expression syntax for LINQ.

Aside from many APIs in the frameworks that already expose sequences, the Iterators and For Loop Expressions features discussed below help you define your own sequences in Oxygene.


Tuples are the second new kind of “container” type introduced by Oxygene. You can think of a tuples as a group of two or more strongly typed variables that can be used and passed around in combination – for example passed into or returned from a method.

Each member of a tuple can have a different, well-defined type (for example, you can have a tuple of a number and a string), but different than a record, members of a tuple have no individual names. Commonly, tuples are used in a more light-weight fashion, in places where declaring an explicit record type somewhere else would seem overkill.

Types are defined with the tuple of keyword combo:

var lError: tuple of (Integer, String) := (404, "Page not found")

You can access individual members of a tuple using their numeric index, such as lError.0. You can also assign tuples back into individual variables by using a tuple literal on the left side of an assignment:

(lCode, lMessage) := WebRequest.GetError();

Future Types

A Future Type is variant of a type that promises to have a value at a later time, but that value might not be calculated or obtained yet, and may be derived either asynchronously in the background, or the first time the future’s value is accessed.

Any ordinary type known in Oxygene can be used as future by prefixing its type name with the future keyword:

var lCount: future Integer := lSomeSequence.Count; // the count of a sequence might be costly to determine

An ordinary future as in the example above will be evaluated the first time the value is accessed. All future access to the variable will use that same value. In essence, the future enables lCount to be referred to multiple times, but ensures it won’t actually be calculated until (and unless) it is actually accessed. Within any subsequent code, lCount can be used just as if it were an ordinary Integer, so it could for example be used in arithmetic expressions such as lCount/4+lCount.

Futures really shine when used in combination with async expressions, as covered below. A future initialized with an asynchronous expression will start calculating its value in the background automatically, so it might already be available when first accessed. As such, futures really help writing parallelized code that can take advantage of multi-core CPUs, with very little work.

Anonymous Classes

Anonymous Classes provide a syntax to quickly define and instantiate a simple class – commonly containing only data, but no code – from inside the code that will use it. Anonymous classes are not often used on their own, but they really shine when used in combination with Sequences and LINQ, as they enable you to filter down or combine data from sequences into new objects on the fly.

For example, as you are processing a lot of Customers and their Orders, you might want to generate a new list that contains each Customer and their total order volume, and then loop over that list. Anonymous classes make that easy without having to clumsily define a class for this. In particular, the select clause of LINQ from expressions will commonly define new anonymous classes.

Anonymous classes are defined using the new class keyword combo:

var lCustomerData := new class(CustomerID: lCustomerID, OrderVolume: lOrders.Sum);

Anonymous Interfaces

Anonymous Interfaces are very similar to anonymous classes, and are used to define a new class inline that satisfies (i.e. implements) a given interface. This is commonly used on the Java and Android platform, where rather than Delphi- or .NET-style Events, controls usually are assigned a delegate object that implements a given interface in order to receive callbacks when events happen – such as to react to the click of a button.

Anonymous interfaces allow you define such a class inline and implement one or more handler methods without having to implement the interface on the containing class (which can be awkward if you need to provide different handlers to different controls – for example two different click events on two different buttons).

You can think of anonymous interfaces as an extension or a more sophisticated version of Anonymous Methods. In fact, an anonymous method is considered the same as an anonymous interface with just one method.

Anonymous classes are defined using the new interface keyword combo:

fButton.delegate := new interface(OnClick := method begin
// handle the click here

Partial Types

The Partial Types feature allows the definition of a Class or Record to be split across two or more source files. This is mainly used for three different types of scenarios:

  • Complex or very large classes can be split up to keep the individual code files more manageable.
  • Classes that are shared across platforms (for example via Shared Projects can have one part that’s shared, and another that provides platform-specific logic, without needing excessive $IFDEFing.
  • Some UI frameworks, such as WinForms and WPF will use one part for user code, while a second part is maintained by the visual designer or build tool chain.


In Oxygene, like Delphi, simple value types that are stored on the stack will always have a value (a default of 0, if not otherwise initialized), while reference types (mostly Class) that are stored on the heap will be nil unless initialized.

Oxygene, however, provides a way to override this. A variable, field or parameter of value type can be marked as nullable type to indicate that it will default to (and can be assigned) nil. Similarly, a variable of reference type can be marked as not nullable, causing the compiler to enforce it to always be assigned a valid value and never be nil.

Most interestingly, and unique to Oxygene and the other Elements languages, nullable value types can be used in code, including arithmetic expressions, just as their regular counterparts. The Nullability will filter through, so that any expression using a nullable type will in turn also be nullable – and in true tertiary boolean logic, an actual nil value in an arithmetic expression will turn the whole expression nil.

var x := nullable Int; // nil
var y := 5;
var z := 10*x+y; //z will be nullable, and nil

You can read more about nullability here.

Mapped Types

Mapped type are a unique feature of the Elements compiler. They let you create compatibility wrappers for types without ending up with classes that contain the real type. The wrappers will be eliminated by the compiler and rewritten to use the type the mapping maps to.

When working with Oxygene, you will most commonly use mapped types (for example as provided by the Sugar cross-platform library). Using mapped types is seamless, and they behave just like regular non-mapped types.

You will not often need to implement mapped types yourself, but for when you do, Oxygene – like RemObjects C# and Swift – provides a syntax for implementing mapped types when needed, with the mapped keyword.

Please refer to the Mapped Types topic in the Language Concepts section for more details.

Type Members

That covers actual types, and as you see, Oxygene has quite a lot to offer. Next, let’s have a look at what you can do within those types (and in particular, Class or Record. It’s also worth mentioning that in Oxygene, Records are elevated to be pretty much as powerful as classes: In addition to fields, they can contain properties and methods, just like their siblings.

Pretty much the only difference between the two kinds of types is that classes are heap based – they get created in memory as needed, and variables refer to their memory location. More than one variable can point to the same class instance, and you can pass class instances all around your program. Records are stack based and value types. Two variables of record type will always point to unique copies of the record, and passing a record as parameter or assigning it to a second field or variable will create a copy of its data.

Advanced Fields

Fields in classes and records work and behave pretty much as you know them from Delphi. The only new feature for fields is that they can be marked with the readonly directive, which indicates that they can only be written from the Constructor or via an initializer, but are then immutable.

Fields can also be initialized in line, and when they are, their type can be omitted if it can be inferred from the initial value.

fCount := 5; readonly; // fCount will be an Integer

Advanced Properties

Just like fields, Properties in principle work as in Delphi, but as mentioned above are also supported in Records, not just Classes.

That said, Oxygene vastly expands the syntax for declaring properties, making them a lot more convenient to define and work with. All of these features are covered in detail in the Properties section.

  • Like fields, properties can me marked readonly.
  • Like fields, properties can be initialized inline.
  • Properties can be declared without read and write clause, and will automatically be backed by an implicitly created field.
  • Properties themselves can be marked virtual and be overridden, which is cleaner than relying on virtual getters/setters as Delphi does.
  • Properties can be defined in Interfaces.
  • Properties can define different visibility for the getter and setter, for example letting you declare a property that is public readable but only private or protected writable, which can be very powerful.
  • Properties can be marked as locked to synchronize their access to be thread-safe.
  • Properties can be marked to generate Notifications when they change, via the notify directive.
  • Properties can be marked as lazy and have their initialization deferred until they are first accessed.
  • Properties can use more complex expressions than just a field or method name for their read and write statement.

Advanced Methods

Methods also work just as in Delphi, and are supported in Records as well, not just Classes. As mentioned in the Minor Differences topic, Oxygene introduces a new method keyword that we recommend to use for methods, instead of the old procedure and function keywords. It emphasizes the Object-Oriented nature of Oxygene, and deemphasizes the largely irrelevant difference of whether a method returns a value or not. But procedure and function still work as well, in Delphi Language Compatibility Mode.

But once again, Oxygene expands the syntax for declaring methods, all of which is covered in detail in the Methods section.

  • Like properties, methods can be marked as locked to synchronize their access to be thread-safe.
  • Methods can me marked as async to indicate that they will automatically execute in the background. async methods with a return value will return a Future.
  • Methods can me marked as empty if they are placeholders that perform no function. This saves creating an empty method body.
  • Methods can be marked as inline, and their logic will then be embedded into the calling code for performance optimization.

In Oxygene, methods can use a new “multi-part method name” syntax that embraces Cocoa naming conventions (but is available on all platforms, and for all three languages) and makes for more readable and expressive method calls. You can read more in the respective section in the Methods topic.

Methods can also define pre- and post-conditions to validate their arguments and their results, which is covered further down on this page and under Class Contracts.


Iterators are a special kind of method that makes it easy to implement dynamically generated Sequences. Marked with the iterator directive, an iterator method can write regular linear code that can use the yield statement to add values to the sequence. yield works similar to exit in that it returns a value, except that the execution flow of the iterator method keeps going, because the returned value is just one of many that make up the final sequence.

Multi-Cast Events and Blocks

Oxygene introduces a new kind of member for classes and records: Events. While in Delphi events are essentially properties of a special type, and thus get no special syntax, events in Oxygene are fundamentally different and separate from regular properties, and are defined with the event keyword.

Events are multi-cast, meaning that more than one handler can be assigned to an event using the += operator that Oxygene introduces exclusively for events. When the event is triggered, all assigned handlers will be called.

Multi-cast Events are almost exclusively used on the .NET platform, since the Cocoa, Java and Android platforms have different default mechanisms to deal with this concept – such as the Anonymous Interfaces discussed earlier on Java, or more traditional delegate classes on Cocoa. But the Event syntax and infrastructure is nonetheless available on all three platforms, should you wish to use it.

Custom Operators

Finally, Oxygene allows you to define Custom Operators for your classes and records, allowing them to participate naturally in arithmetic expressions. For example, you can define the + operator for a record representing a Complex number or a Matrix, allowing code that consumes the new record (or class) to seamlessly add two values together.

You can read more in the Custom Operators section.


We’ve now covered both types and their members, so next, let’s take a look at what Oxygene lets you do inside those members, most particularly Methods-like members, in terms of the kinds of Statements you can write.

Inline vars and Type Inference

Most prominently, Oxygene does away with the need for an explicit var section at the top of each method where all the method’s local variables need to be declared.

Instead, Oxygene lets you declare variables throughout the flow of your method where they are needed, with the new var statement. This makes code easier to understand, as variables can be declared closer to where they are used, and even inside nested scopes, such as if blocks or Loops.

More importantly, the var statement supports type inference, so you can, for example, assign the result of a method call to a new local variable without restating (or even knowing) the exact type. Variables defined with inferred type will of course still be strongly typed.

Type Inference is of course especially important when working with Anonymous Classes discussed above, since these classes don’t even have a known type name that could be explicitly stated. Type inference is the only way to declare a variable holding such a type (or a Sequences of such types).

Infinite Loops

Mostly a curiosity but handy at times, Oxygene introduces a new loop type that runs indefinitely, with the loop keyword. A loop loop, also called an infinite loop, has no pre-determined exit condition and keeps running until it is broken out of with break or exit.

While not used often, loop does make for cleaner code and lets you avoid awkward and unintuitive while true or repeat until false loops.

Improved For Loops

for loops have also bee greatly expanded in Oxygene.

For one, Oxygene adds a new for each/in variation in addition to the trusted for/to loop. for each loops run over all members of a collection, Array or Sequences, without your code having to maintain an indexer manually. (More recent versions of Delphi have adopted this loop style as well, so you might already be familiar with it.)

for each loops also have two advanced syntaxes.

  • Via the index keyword, you can introduce a second loop variable that keeps track of the count of loop iterations, without you having to increment the variable yourself. Essentially, index gives you the best of both for each and for/to loops, in one.
  • Via the matching keyword, you can limit the loop to only execute for those members of a collection that are of a specific sub-type.

Currently on .NET only, both loop types can also be made to run multiple loop iterations in parallel on different threads, via the parallel keyword.

Exception handling

Exception Handling has been expanded over Delphi’s in two ways:

  • A single try block can be followed by both a finally block and one or more except blocks. There no longer is any need to nest two try blocks just to leverage both types of handler.
  • except blocks can be expanded using where clauses to further filter which exception a given block will catch, using criteria other than just the mere exception type.

Advanced Case Statements

Oxygene expands the case statement to be more flexible. case statements can work on strings (smartly using a hash table in the background for efficiently finding the proper case to execute). This not only saves code over needing to write multiple if/else if/else if statements, but is also faster.

The case statement can also execute different branches based on the type of its parameter, via the new case/type of syntax.

Refer to the case Statements topic for more details.


Similarly to the locked directive on Methods and Properties already mentioned above, the locking statement can protect a section of code against parallel execution on multiple threads, making it very easy to write code that is ready for parallelization. Via its parameter, the locking statement gives you flexibility for how granularly to synchronize execution – for example per instance, or globally.


While Oxygene uses GC or ARC on all platforms and you do not usually need to worry about memory and object lifetime management, sometimes your code will interact with external resources (such as file or network handles) that do need to be released in a clean and timely fashion.

The using statement allows you to write a block of code that will run and make use of a specific object and automatically dispose of the object at the end. Essentially, using is a convenient way to encode a try/finally block that makes sure your object and external (“unmanaged”) resources get cleaned up.


With statements out of the way, let’s look at some of the improved Expression types Oxygene provides.

Colon Operator

Small but immensely powerful, the Colon (:) Operator is a team favorite of all the features in Oxygene.

Delphi and Oxygene normally use the Dot (.) operator to access members such as Properties or Methods of a class reference. This is something so natural and so frequently done, we mostly don’t even think about this as a special expression.

When trying to access a member of a class reference that happens to be nil, an exception is raised. In Delphi, that is the dreaded Access Violation straight from the CPU, in Oxygene it’s a Null Reference Exception, often chummily called “NRE”.

NREs are great when they happen on truly broken code, as they report the failure in a clean and obvious manner. But oftentimes it would be nice to be able to write code that doesn’t need to care if an object is nil or not. That’s where the Colon (:) Operator comes in.

If you use : instead of . to call a member, Oxygene will automatically check whether the object you are trying to call into is valid or not. If the object is valid, the call will proceed as normal, same as with .. But if the object is nil, then rather than raising an NRE, Oxygene will just skip the call altogether and return nil as the result.

Consider this example:

var lKnownSiblings := lSomeObject.Parent:GetChildren();

This code will call the GetChildren method of the object from the Parent property of lSomeObject. But what if Parent is not assigned (for example because data is incomplete, or because lSomeObject is the root of the hierarchy)? Because the code uses :, the call to GetChildren will simply be omitted, and lKnownSiblings will be set to nil.

The Colon (:) Operator allows you to write code that’s simpler (often avoiding many nested if assigned(...) checks) and less error prone.

Double Boolean Comparisons

Double Boolean Comparisons allow you to compare three values in one step with a ternary operator – for example to check if a given value falls between two boundaries.

if 5 <= Count <= 10 then writeLn('between five and ten');

Lambda Expressions

[Lambda Expressions] provide a convenient shortcut syntax for writing short Anonymous Methods without the overhead of a full method/begin/end declaration. Lambda expressions are commonly used for single-statement methods, where they consist of an (optional) parameter list, the special -> operator, and the single statement. For example:

var lFives := lMyCollection.Where(x -> x.Value = 5); // filter list to items with value 5

Lambda expressions can be used anywhere anonymous methods can be used – for example as event handler assignments or as Block parameters to methods. One very common scenario, as shown in the example, is to use them with the LINQ query operators.

If Expressions

if expressions take the regular if statement and allow it to be used for expressions. As such, the if expression evaluates a condition, and then returns one value or the other.

var lLabel := if lList.Count = 1 then 'Item' else 'Items';

Case Expressions

Similar to if expressions, case allow the regular case statement syntax to be used as an expression to return a conditional value:

var lCountString := case lList.Count of
0: 'none';
1: 'one';
2: 'two';
else 'more than i can count';

For Loop Expressions

You are probably seeing a pattern here. For Loop Expressions are the expression version of the regular for loop statement. Since a for loop, by its nature, can run for many iterations, the result of a for loop expression is a Sequences of values:

var lSomeEvenNumbers := for i := 1 to 100 yield i*2;

Similar to Iterators, for loop expressions use the yield keyword to add a value to the generated sequence. Also like iterators, the value of a for loop expression merely represents the functional logic for generating the sequence. The loop code does not actually run until the sequence is enumerated.

Async Expressions

async expressions allow a statement or an expression to be executed and evaluated asynchronously on a background thread, returning a Future Type.

Calling an async expression will return immediately, and execution will begin in the background (immediately, or once a free thread is available based on system resources).

Await Expressions (.NET)

Available on .NET only, the await expression construct can be used to “unwrap” otherwise asynchronous code so that future results can be dealt with in a linear fashion. Under the hood, await will break the method into different parts, scheduling them to be executed asynchronously once the awaited actions have been completed.

Please refer to the await Expressions topic for more details.

From (LINQ) Expressions

A huge topic on their own, from expressions provide a rather sophisticated sub-language that allows you to use an SQL-like syntax to work with Sequences of objects in a strongly-typed fashion. They form the basis of LINQ support.

var lFilteredData := from c in lCustomers
where c.Name.StartsWith('O') // filter by name
order by c.DateOfBirth // order by date
select c.Name, c.Address; // and return only two fields
// via a new anonymous class

Note: Although LINQ technology originated on .NET, Oxygene makes it and from expressions available on all platforms.

“is not” / “not in”

Oxygene expands the standard is type check operator and the in operator that checks for membership in a set to allow for more natural use with the not keyword.

if not (x is Button) then … // traditional Delphi
if not (5 in MySet) then … // traditional Delphi

if x is not Button then … // Oxygene
if 5 not in MySet then … // Oxygene

Class Contracts

Last but not least, Oxygene introduces a major language feature called Class Contracts that allows you to write self-testing code in a “Design-by-Contract” fashion.

Class Contracts consist of two syntax features:

  • Inside method implementations, you can add code to check for pre-conditions and post-conditions using the require and ensure keywords, as shown in the “Method Implementation Syntax” section of the Methods topic.

  • On a class (or record) level, you can define Invariants that are used to define a fixed state the type must fulfill at any given time. This makes it easy to detect bugs where any method or property setter leaves the type in an inconsistent state.

You can read more about these features in the Class Contracts topic.


This article originally appeared on our new Elements docs site at Check out our whole section for newcomers from Delphi at

Profile photo of marc

by marc

Staring into the Fire: Code Completion

September 8, 2014 in Fire

This post is part of a series in which I want to take a closer look at some of the features and functionality in Fire, our new IDE for the Mac. The topic for today is Code Completion.

Code completion, also called IntelliSense in Visual Studio parlance, is a crucial part of writing code these days. Instead of reading docs or knowing types and their members by heart, we rely on Code Completion to let us know and discover what APIs are there for us to use. And gone are the days of trying to save keystrokes by making names easy to type — because CC can help us type complex names, easily.

Of course Fire comes with state-of-the-art Code Completion in its code editor, driven by the same back-end engine that also provides CC data for our languages in Visual Studio. But while the content is the same, Fire takes the CC user interface to the next level, as we have really put a lot of thought into how we can improve the general experience.

In Fire, Code Completion will automatically activate as you type, wherever sensible. So you don’t need to think about it much, and it will automatically be there to help you, should you need it. Of course you can also manually invoke CC at the current cursor location, if so desired. You can either press Escape (which is the Mac default for this), or — if you’re still used to Visual Studio, and maybe switching a lot between the two IDEs — you can use Cmd-Space (after changing the global Spotlight shortcut, which you probably already did to use that combo on your VM, anyways).

CC is fast, but sometimes even the tiniest of interruptions can be annoying, so CC in Fire is completely asynchronous. Even while CC gathers info to show you, you can keep typing, or cursor elsewhere. If you type more of the current identifier, CC will adjust as you type to narrow your search. If you type or move away from the current identifier, CC will close (or never show, if you’re a really fast typer ;).

The CC “Soft Selection”

As Code Completion gets activated, you will notice that Fire will add a soft highlight to the current “token” that it considers relevant. This is a helpful indicator as to what part of your code will be replaced should you accept an item from the CC list.

If you use Visual Studio, you might have (consciously or subconsciously) noticed that sometimes when you select an item from CC, VS will replace the full token you are on, and other times, it inserts the new text in front of the existing text. If you’re like me, on more than one occasion you’ll have been annoyed that it did the opposite of what you wanted or expected. Fire’s visual indicator helps with that. For example, if you start typing at the very front of a token, chances are that you are adding a new token in front and don’t want to replace what’s already there, so CC will not absorb the text behind the cursor:

But if you invoke CC in the middle of an identifier, chances are you want to replace the whole identifier, so Fire’s UI will reflect that:

Not only does the “soft selection” show you what to expect, but Fire is also smart about picking the right “mode” for CC, depending on the context it finds itself in.

Partial CC

Fire’s CC also has a nifty way that helps you type complex names. Take the following example of very common Cocoa code:

dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_NORMAL, …)

That is a huge pain to type, even with CC, because each the three identifiers starts with “dispatch_”, and there’s a huge list of those. So you end up either typing most of the names manually or doing a lot of scrolling through the CC list. Not anymore!

In Fire, when you press Tab or underscore (_), CC will automatically advance to the next underscore in the current item (or to the next item with an underscore. So for example, if you just type dis_g_g_q(, CC will smartly navigate you through the list so that you end up with dispatch_get_global_queue(|):

Similarly, Tab also jumps ahead between parts of PascalCased names. So it’s just NSFil<Tab>Ma<Tab> and you have NSFileManager.

The Devil’s in the Details

There are other small details that streamline the CC experience. When CC gets invoked and there’s no current token to soft-select yet (the most common case when you’re starting a new identifier), Fire will show a small “sliver” of a soft selection to the right of the cursor to show you CC is coming:

If you accept CC for a method by typing (, Fire will automatically insert the pair of parenthesis and put the cursor between them — if you accept CC with a different key, for example with space or enter, it will not.

Of course CC also handles multi-part method names in Oxygene, C# and Silver (even if they look like named parameters in the latter) — inserting the full method signature for you, and giving you “jump points” to let you tab from one parameter to the next:

If you ignore CC and “type through”, CC will smartly do the right thing when you come to the end of the identifier. For example, it will adjust the case of what you typed to the proper item in CC. But not if the version you typed would also be valid, as in the case below (which of course only happens in case sensitive languages, i.e. C# and Silver):

When you start typing in the middle of an existing identifier, CC will not drop down automatically. Chances are you’re trying to fix a typo or otherwise know what you’re doing, and our use testing showed that the CC dropdown just gets in the way. (Of course you can still invoke CC manually via Escape for that location — it will then cover the entire identifier.)

Just like these, there are lots of little tweaks and and “special cases” — some even language-specific — that fine-tune CC to behave as you’d want it to and not get in the way. Most of them being cases that you won’t actually notice — except by lack of being annoyed ;).


This post gave you a quick look at what Code Completion is like in Fire. Code Completion is implemented as part of Fire’s Code Editor, which we have written from scratch to give you the best possible code editing experience. I’ll talk more about the Code Editor in general in a future post.

Stay tuned, check out Fire, and make sure to let us know what you think!