Could vaxe be made to take it less exact with errors?
Opened this issue · 4 comments
I often have the situation that my lime project compiles from the console (or by :make) but vaxe autocompletion reports errors.
Also, sometimes I want autocompletion also my project does currently not compile.
For example, when I import:
import haxe.ui.toolkit.core.interfaces.IDisplayObject;
I always get:
Class not found : haxe.ui.toolkit.core.interfaces.IDisplayObject
also the project compiles (and I can use IDisplayOject successfully).
There's been some changes with how lime manages build attributes. It doesn't seem like they get exported properly in the lime-generated hxml. I have questions in to Josh about it. See also #50
Mmmh, when I call haxe autocomplete from the command line like this:
haxe --display Source/NuggetaHandler.hx@199 ./Export/linux64/cpp/haxe/release.hxml
It gives valid results.
But when doing autocomplete in vim at the same position, it fails. Can I somehow output what exact command is used for invoking haxe when autocompleting?
I'm trying to revisit some of these stalled bugs, are you still having problems? I know that Josh just updated openfl.
Hey,
vaxe works pretty good for me. Sometimes I have to delete the generated hxml file ... wonder if vaxe could be made to auto-re-generate it when libs have changed?!?
Anyhow, what still would be cool, imagine this example:
import StupidClass;
class Test {
public function test() {
this.
}
}
Stupid class does not exist and for that reason, the autocompletion at the "this." does not work. I often want to finish the code I am working on and correct the import later ...