Previous post by @yegor256 explained declarative WHILE nature. Let’s examine more complex example when while objects are nested.

In the recently released version 0.28.14 we’ve changed the iterating algorithm of the bool.while object. Until now, by our mistake, it was imperative. Now, it’s declarative. The difference is in the result of its dataization. The previous imperative version was “returning” a data object. The new declarative one returns the latest body of the loop (without dataization!). The difference is huge (thanks to it, many of our tests broke).

In the recently released version 0.28.14 we introduced a new language feature, which is called “object adoption” — because it allows changing of an object’s parent. Every object in EO has a parent object, which is set to it when it’s born (either formed or copied). Until the recent release it was not possible to change the parent object using EO language. However, it was possible to do this from inside an atom (through Java). Now, there is no special API inside Java, but the feature is available through EO.

There are “atoms” in EO language, which are objects implemented by the runtime platform, not by a composition of other EO objects. Most notable examples of atoms are int.plus, float.times, and bool.while. Here is a quick intruction to creating your own atoms.

XMIR is a dialect of XML, which we use to represent a parsed EO program. It is a pretty simple format, which has a few important tricks, which I share below in this blog post. You may also want to check our schema: XMIR.xsd.

As described in this blog post place step of EO build process copies compiled files from dependencies’ JARs into target/classes folder to have them in classpath during compilation later. You can check PlaceMojo class for exact logic of this step.
All copied files are registered within “placed catalog,” which is normally stored as target/eo/placed.csv file. Let’s consider structure of the catalog in more details.

As in all other programming languages, an object while is used to iterate over a set of statements while a condition is true. In EO this object not only does this, but is also dataized. And the result of such dataization was the number of iterations performed. Now this behavior has changed: the result of such dataization can now be any object (depending on the condition). It can be divided into two parts: when the condition is TRUE and when the condition is FALSE.

EO parser is written in ANTLR4, Java, and XSL. It is packaged as a multi-module Maven project, in objectionary/eo GitHub repository. In order to compile an EO program to Java you may either use our Maven plugin or our eoc command line toolkit. Moreover, you can also use our parser programmatically in order to generate XMIR from EO. Here is how you do it in your Java/Kotlin/Clojure/Groovy/etc. project.