| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is the one major refinement on the C programming model I'm planning
to introduce in mu. Instead of Rust's menagerie of pointer types and
static checking, I want to introduce just one new type, and use it to
perform ref-counting at runtime.
So far all we're doing is updating new's interface. The actual
ref-counting implementation is next.
One implication: I might sometimes need duplicate implementations for a
recipe with allocated vs vanilla addresses of the same type. So far it
seems I can get away with just always passing in allocated addresses;
the situations when you want to pass an unallocated address to a recipe
should be few and far between.
|
| |
|
|
|
|
|
|
|
|
| |
I wasn't seeing errors because I wasn't using /contained-in in products
yet. But it seems to work fine even after.
One reason this isn't an invasive change is that it's opt-in. Most of
the time it isn't triggered. You have to add the /contained-in check to
trigger a check. But that's just like regular const checks in other
languages: if you don't specify immutability you get no checks.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This isn't complete yet:
a) If you copy a variable the copy is not checked for mutations.
b) /contained-in might be a hack. It's needed because sometimes I want
to pass in two pointers to a data structure, say for deleting something
from a list. Both are conceptually the same structure, so it's
unnecessary for both to be products as well. There's also technical
reasons you *can't* return both, because if you pass in the same
variable to both ingredients (say you want to remove the first element
of a list), the products now must write to the same variable as well
(thanks to our earlier /same-as-ingredient constraint), and what value
gets written last is not something we want to be thinking about.
c) Even if we stick with /contained-in, it's ambiguous. I'm using it
sometimes to say "a belongs to b", sometimes to say "a _will_ belong to
b after the recipe returns. Perhaps that distinction doesn't matter.
We'll see.
d) Should we be encouraged to say ingredients are contained in products?
At the moment 'push' works only because of mu's incomplete analysis.
Once we fix a) above, it's unclear what the right header should be.
e) edit/ isn't close to working yet.
(No more commit numbers since I'm now starting a branch, and can't rely
on a stable ordering as I rebase. For the same reason I'm not including
the changes to .traces/, to minimize merge conflicts.)
|
| |
|
|
| |
Great that it just worked after the previous commit.
|
| |
|
|
|
|
| |
Fix the drawback in the previous commit: if an ingredient is just a
literal 0 we'll skip its type-checking and hope to map type ingredients
elsewhere.
|
| |
|
|
|
|
|
|
|
| |
I was failing to specialize calls containing literals. And then I had to
deal with whether literals should map to numbers or characters. (Answer:
both.)
One of the issues that still remains: shape-shifting recipes can't be
called with literals for addresses, even if it's 0.
|
| |
|
|
|
|
| |
Lessons with Caleb uncovered a problem with type ingredients: I can call
shape-shifting recipes like 'push' from the commandline but not inside
the edit/ or sandbox/ apps.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The rule is, an address ingredient is only modifiable if:
a) it's also a product
b) it's /contained-in some other ingredient+product
Only if an ingredient is a modifiable can you:
a) call get-address or index-address on it (the only way to write to it)
b) call other recipes that also return it in a product
I still don't check copies of the address. That's next.
Core mu passes this check, but none of the example apps do. edit/ and
sandbox/ are known to fail.
|
| |
|