尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證正在載入完整 README。 以下為已儲存的摘錄。
Extend your Java code with Liquid Types and catch bugs earlier!
LiquidJava is an additional type checker for Java, based on liquid types and typestates, which provides stronger safety guarantees to Java programs at compile-time. With this extension, you can use LiquidJava directly in VS Code, with real-time diagnostic reporting, syntax highlighting for refinements, and an interactive webview for displaying details about diagnostics and state machine diagrams.檢視完整 README ↗LiquidJava is an additional type checker for Java, based on liquid types and typestates, which provides stronger safety guarantees to Java programs at compile-time. With this extension, you can use LiquidJava directly in VS Code, with real-time diagnostic reporting, syntax highlighting for refinements, and an interactive webview for displaying details about diagnostics and state machine diagrams.
@Refinement("a > 0")
int a = 3; // okay
a = -8; // type error!
To try out the extension, install it from the VS Code Marketplace or the Open VSX Marketplace. Additionally, you'll need the Language Support for Java(TM) by Red Hat VS Code extension, and add the liquidjava-api dependency to your Java project:
<dependency>
<groupId>io.github.liquid-java</groupId>
<artifactId>liquidjava-api</artifactId>
<version>0.0.7</version>
</dependency>
repositories {
mavenCentral()
}
dependencies {
implementation 'io.github.liquid-java:liquidjava-api:0.0.7'
}
A repository with LiquidJava examples is available at liquidjava-examples. You can try them out without setting up your local environment using GitHub Codespaces.
Liquid types extend a language with logical predicates over the basic types. They allow developers to restrict the values that a variable, parameter or return value can have. These kinds of constraints help to catch more bugs before the program is executed. For example, they allow us to prevent bugs like array index out-of-bounds or division by zero at compile-time.
To refine a variable, field, parameter or return value, use the @Refinement annotation with a predicate as an argument. The predicate must be a boolean expression that uses the name of the variable being refined (or _) to refer to its value. You can also provide a custom message to be included in the error message when the refinement is violated. Some examples include:
@Refinement("x > 0") // x must be greater than 0
int x;
@Refinement("0 <= _ && _ <= 100") // y must be between 0 and 100
int y;
@Refinement(value="z % 2 == 0 ? z >= 0 : z < 0", msg="z must be positive if even, negative if odd")
int z;
@Refinement("_ >= 0")
int absDiv(int a, @Refinement(value="b != 0", msg="cannot divide by zero") int b) {
int res = a / b;
return res >= 0 ? res : -res;
}
To simplify the usage of refinements, you can create predicate aliases using the @RefinementAlias annotation, and apply them inside other refinements:
@RefinementAlias("Percentage(int v) { 0 <= v && v <= 100 }")
public class MyClass {
// x must be between 0 and 100
@Refinement("Percentage(x)")
int x = 25;
}
Beyond basic refinements, LiquidJava also supports object state modeling via typestates, which allows developers to specify when a method can or cannot be called based on the state of the object. You can also provide a custom error message for when the method precondition is violated. For example:
@StateSet({"open", "closed"})
public class MyFile {
@StateRefinement(to="open(this)")
public MyFile() {}
@StateRefinement(from="open(this)", msg="file must be open to read")
public void read() {}
@StateRefinement(from="open(this)", to="closed(this)", msg="file must be open to close")
public void close() {}
}
MyFile f = new MyFile();
f.read();
f.close();
f.read(); // type error: file must be open to read
Finally, LiquidJava also provides ghost variables, which are used to track additional information about the program state when typestates aren't enough, with the @Ghost annotation. Additionally, you can also refine external libraries using the @ExternalRefinementsFor annotation. Here is an example of the java.util.Stack class refined with LiquidJava, using a size ghost variable to track the number of elements in the stack:
@ExternalRefinementsFor("java.util.Stack")
@Ghost("int size")
public interface StackRefinements<E> {
public void Stack();
@StateRefinement(to="size(this) == size(old(this)) + 1") // increments size by 1
public boolean push(E elem);
@StateRefinement(from="size(this) > 0", to="size(this) == size(old(this)) - 1", msg="cannot pop from an empty stack") // decrements size by 1
public E pop();
@StateRefinement(from="size(this) > 0", msg="cannot peek from an empty stack")
public E peek();
}
Stack<String> s = new Stack<>();
s.push("hello");
s.pop();
s.pop(); // type error: cannot pop from an empty stack
You can find more examples of how to use LiquidJava on the LiquidJava Website. To learn how to use LiquidJava, you can also follow the LiquidJava tutorial.
For more information, check the following repositories: