Markus Triska
Markus Triska
Today I found an interesting article that may help with this issue: https://tunglevo.com/note/crafting-interpreters-with-rust-on-garbage-collection/
> The mark phase of the GC algorithm seems to require a bit vector eventually. @UWN's message indicates that this bit vector can be constructed **on marking**, as a result...
> The marking algorithm works by scanning the heap one cell at a time in either direction. What do you mean with "in either direction"? Strings are only stored in...
All I can recommend is to try to **keep everything as simple and robust as possible**. A single mark phase starting from root nodes should be enough to register the locations...
We want to support **string differences** efficiently. Consider for example the following *partial* string (i.e., we don't yet know its length, since the tail is a variable): `[a,b,c|Ls]`. It must...
I this this is addressed by 5418a40c53f738b0569aff72febeae59be786b8b?
Yes, it seems so!
Hm, I overlooked `fn new_char_inline`, which I think *did* create such atoms, but now no longer does. The intention of this issue was to remove an unnecessary conditional, but now...
Thank you a lot! Specifically, in the following example: ?- X+\(X=a). X = a. Would we say that `(X=a)` is a lambda expression with **no parameters**? It appears so! (I...
This change improves SICStus compatibility? Then it seems good to implement it. Thank you a lot! Please also see the other points raised in discussions because they may be more...