Make up your mind -- do you want to try it "without the goal phrase", or do you want to try it "the evolutionary way"? You can't have it both ways.
Evolution requires a nondiscrete selective force (along with replication and variation). If you don't understand this basic thing about evolution, it's no wonder you've got all sorts of incorrect presumptions about what it is or is not capable of doing.
But hey, that's only been known since 1859, perhaps you haven't kept up.
For evolution to occur, there has to be some method by which variations are "graded" in some fashion by whether they are more or less "successful", and this has to affect their reproductive rates. In nature, that takes place by the sheer fact of life itself -- those individuals which are better suited to survive long enough to reproduce (and reproduce successfully, and/or more often) are the ones who are going to pass on more of their genes than the ones who do so less successfully.
If you're going to use your "hello world" as an *evolutionary* example, instead of an irrelevant exercise in random walks, you're going to have to include some sort of method by which those programs that get *closer* to a "hello world" output are more likely to replicate than those which are *farther* from that output.
Evolution is an exercise in hill-climbing, not wandering aimlessly around on a flat surface hoping to bump into a lone invisible flagpole.
You can amaze yourself with false premises all day.
Indeed, so why don't you cut it out? Your "examples" are false premises in that they specifically rule out actually applying evolution. And then you declare "victory" when your non-evolutionary scenarios are self-evidently not going to evolve successfully.
If you want to discuss evolution, put for an example which meets the requirements for evolution to occur. Stop throwing out "false premises all day".
The only question that remains is, are you doing this on purpose or through error?