WIP documents for memory interface and value domain #175
Conversation
|
Wow, so much feedback so quick! Thanks a lot. I hope I can go over it tonight. One thing I should have mentioned: the part in the glossary is hopefully uncontroversial, we should agree on it for merging. The "WIP" part is, just like the other WIP document, basically my personal proposal / opinion; I was hoping for a review to increase clarity but I do not think we should strive for consensus on the opinions expressed in those documents. |
RalfJung
force-pushed
the
memory-interface
branch
from
270dfe1 to
168249a
Compare
|
Okay, I think I responded to all comments and edited the documents accordingly. Thanks a ton for all your feedback! |
RalfJung
changed the title
| We define the `Byte` type as follows, where `Pointer` will later be instantiated with the `Memory::Pointer` associated type. | ||
|
|
||
| ```rust | ||
| enum Byte<Pointer> { |
There was a problem hiding this comment.
Maybe Byte is not a good name for this, as it can be viewed as re-defining the term. I am basically starting from "a byte is what gets stored at a memory address", but other people might think "a byte is 8 bits" or "a byte is an integer in 0..256", and then this type here makes no sense. It turns out that matching all these definitions at once is just not possible, though.
|
It's been two weeks, and this is just a WIP document. Any objections to merging? |
|
Let's do it! |
This also resolves #40. I concede to the majority opinion that the 4 terms in question are "value", "place", "value expression" and "place expression".
EDIT: Oh, I should also mention: just like the other WIP document, this is basically my personal proposal / opinion; I was hoping for a review to increase clarity but I do not think we should strive for consensus on the opinions expressed in those documents.