如何消除OpenModelica外部函数数组输出的显式映射警告？

在OpenModelica项目中，使用单输出数组的外部函数时出现翻译警告，提示未使用显式映射，OpenModelica会将其作为输入处理但不符合Modelica规范。现有外部函数声明和对应C代码可正常编译返回值，但警告无法消除，需修改代码解决。

警告信息 #

[1] 18:05:18 Translation Warning
[MyModel: 3:1-11:25]: An external declaration with a single output without explicit mapping is defined as having the output as the lhs, but language C does not support this for array variables. OpenModelica will put the output as an input (as is done when there is more than 1 output), but this is not according to the Modelica Specification. Use an explicit mapping instead of the implicit one to suppress this warning.

原代码 #

Modelica函数声明 #

within MyPackage;

impure function MyExternalFunction "My External Function"
input Real input1 "Input 1";
input Real input2 "Input 2";
input Real input3 "Input 3";
output Real result[2] "result1, result2";
external "C"  annotation(
    IncludeDirectory = "modelica://MyPackage/",
    Include = "#include \"my_code.c\""
);
end MyExternalFunction;

对应C代码 #

void MyExternalFunction(double input1, double input2, double input3, double* result, size_t n) {
    result[0] = 1.2;  // result 1
    result[1] = 3.4;  // result 2
    n = 2;
}

修改方案 #

1. 为Modelica外部函数添加显式参数映射 #

核心是在external声明中明确指定参数的传递顺序和映射关系，替代默认的隐式映射。修改后的Modelica代码如下：

within MyPackage;

impure function MyExternalFunction "My External Function"
input Real input1 "Input 1";
input Real input2 "Input 2";
input Real input3 "Input 3";
output Real result[2] "result1, result2";
external "C" MyExternalFunction(input1, input2, input3, result, size(result)) annotation(
    IncludeDirectory = "modelica://MyPackage/",
    Include = "#include \"my_code.c\""
);
end MyExternalFunction;

这里通过MyExternalFunction(input1, input2, input3, result, size(result))明确告诉OpenModelica：将输入参数依次传递，把输出数组result和它的长度size(result)作为参数传入C函数，完全符合Modelica规范中显式映射的要求，从而消除警告。

2. 优化C代码（可选） #

原C代码中n = 2;属于无效操作，因为n是传值参数，修改后无法反馈到Modelica侧。建议调整C代码，基于传入的数组长度做安全赋值，避免越界：

void MyExternalFunction(double input1, double input2, double input3, double* result, size_t n) {
    // 确保数组长度足够时再赋值，避免访问越界
    if (n >= 1) {
        result[0] = 1.2;
    }
    if (n >= 2) {
        result[1] = 3.4;
    }
}

内容的提问来源于stack exchange，提问作者vindarmagnus