Replace MRB_NO_DIRECT_THREADING with MRB_USE_VM_SWITCH_DISPATCH

[dearblue]

The correct technique now is "Token Threading".
However, the new name was chosen because it was felt that if a different technique was added in the future, there would be no need to replace it.

[matz]

I am not sure how proposed wording is better. Could you explain?

[dearblue]

🐈‍⬛ My only source of information is https://ja.wikipedia.org/wiki/スレッデッドコード, which may be suspect.

• The current implementation is based on "Token threading".
• "Direct threading" is not "Token threading".
• "Token threading" is not "Direct threading".
• "Threaded code" includes "Token threading", "Direct threading", "Huffman threading" and etc..

[dearblue]

I fixed the conflicts and rebased.

[matz]

I understand the wording (and I am sure it's valid, since Prof. Ertl mentioned). But since we see “switch threading” in the reference, so if used precisely, the term threaded code itself is meaningless. Let me think about it for a while.

[dearblue]

Oh, "Switch threading".
Hmmm, I'll leave this PR open for now, but switch to draft.

[dearblue]

How about MRB_USE_SWITCH_THREADING?
Or is MRB_USE_VM_SWITCH_DISPATCH more understandable?

[matz]

I prefer the term "switch dispatch" to "switch threading", for not using direct goto.

[dearblue]

Thanks for your comment.
Replaced with MRB_USE_VM_SWITCH_DISPATCH.

And DIRECT_THREADED has been merged into MRB_USE_VM_SWITCH_DISPATCH.
So, please note that #ifdef and #ifndef have been replaced by the reverse.