Kay Hayen

Well, you can always use `--onefile-temp-spec` and not use the temp folder, but e.g. a user cache folder, and if you want things to delete, just use a volatile value...

I think whatever `mount` does, we can do too of course. Querying file system and flags is surely simple API. Having a fallback to something is useful. I just hate...

If we made it such, that the spec was a list of things, that are tried, that would allow to specifiy such ugly things as `/run` and have it work...

I don't see myself doing this anytime soon, unless a commercial sponsor level subscriber asks for it.

This is the opposite of the other changes suggested apparently, without this, stdlib will only be included in as so far as it is technically required to do so for...

Not sure if I got what you mean, but I think I will make sure I implement the consideration of these follow options, which may already work, not sure now.

Seeing my latest comment in your PR. I believe it's all the same actually. The decision function just needs to take into account the suggested options, and be used in...

So, this is not working still, also tried --nofollow-import-to ``` python bin/nuitka --report=out.xml --noinclude-custom-mode=poplib:nofollow tests/onefile/HelloWorldTest.py ``` We need to make that scan respect this kind of option already. Not sure...

So, anti-bloat somehow won't work on stdlib, which is fine, but `--nofollow-import-to=poplib` can be made to work by moving the decision to after the no case branch indeed. This needs...

Well, I think IPython ought to work, it's just causing too big compiles. What might easily happen is that we break IPython when we do changes and do not guard...