I was testing an import as written at [1], which is not a documented form[2], and discovered that such imports introduce names into the global namespace:
```
void main() {
//import io = std.stdio; // Error.
import io = std.stdio : writeln; // Executes.
writeln("test");
}
```
It appears that `import io = std.stdio : writeln;` is being interpreted as `import io = std.stdio : writeln=writeln;`. If this is the case, should an implicit `name=name` be allowed?
```
void main() {
import io = std.stdio : writeln; // This builds and runs.
writeln("test"); // Executes.
//write("test"); // Error.
io.writeln("test"); // Executes.
io.write("test"); // Executes.
}
```
[1]: http://forum.dlang.org/post/[email protected]
[2]: Sections 4.6 - 4.8: https://dlang.org/spec/module.html#renamed_imports
Comment #1 by dlang — 2017-08-16T11:35:19Z
"Ambiguous" may be the wrong term; the compiler knows exactly what it's doing. It may be ambiguous to the programmer reading and writing it. In that same forum thread, we read that one person (two including myself) expected this to work differently[1] (e.g., did not expect symbols to be introduced into the global namespace).
[1]: http://forum.dlang.org/post/[email protected]
Comment #2 by robert.schadek — 2024-12-13T18:54:04Z