fix rewrites of ordered_hash_map so the key is also rewritten

[TomerStarkware]

No description provided.

[reviewable-StarkWare]

This change is 

[orizi]

Reviewed all commit messages.
Reviewable status: 0 of 3 files reviewed, 2 unresolved discussions (waiting on @TomerStarkware)

---

crates/cairo-lang-semantic/src/expr/inference.rs line 387 at r1 (raw file):

                .type_assignment
                .iter()
                .map(|(k, v)| (*k, inference_id_replacer.rewrite(*v).no_err()))

this sort of rewrite sounds like the same thing.

Code quote:

            type_assignment: self
                .type_assignment
                .iter()
                .map(|(k, v)| (*k, inference_id_replacer.rewrite(*v).no_err()))

---

crates/cairo-lang-semantic/src/substitution.rs line 213 at r1 (raw file):

        *value = temp_value
            .into_iter()
            .map(|(mut k, mut v)| {

This makes map rewrite much heavier - as it actually has to create a new map.

this wasn't this way on purpose, we need to either makes sure this doesn't hurt performance (note the explicit calling for this sort of rewrites in other areas, like the inference rewrite), or just call this heavy version when required.

Code quote:

        let mut result = RewriteResult::NoChange;
        let temp_value = std::mem::take(value);
        *value = temp_value
            .into_iter()
            .map(|(mut k, mut v)| {

[TomerStarkware]

Reviewable status: 0 of 3 files reviewed, 2 unresolved discussions (waiting on @orizi)

---

crates/cairo-lang-semantic/src/substitution.rs line 213 at r1 (raw file):

Previously, orizi wrote…

This makes map rewrite much heavier - as it actually has to create a new map.

this wasn't this way on purpose, we need to either makes sure this doesn't hurt performance (note the explicit calling for this sort of rewrites in other areas, like the inference rewrite), or just call this heavy version when required.

Now I am only cloning the keys when there is no rewrite,
also we only use this method with copiable keys so we can change T to be copiable

---

crates/cairo-lang-semantic/src/expr/inference.rs line 387 at r1 (raw file):

Previously, orizi wrote…

this sort of rewrite sounds like the same thing.

here the key is not rewritable

[orizi]

Reviewed 2 of 3 files at r1, all commit messages.
Reviewable status: 2 of 3 files reviewed, 1 unresolved discussion (waiting on @TomerStarkware)

---

crates/cairo-lang-semantic/src/substitution.rs line 213 at r1 (raw file):

Previously, TomerStarkware wrote…

Now I am only cloning the keys when there is no rewrite,
also we only use this method with copiable keys so we can change T to be copiable

do note that you should know this from the type of the keys (if they have 'identity-rewrite') - but maybe that's enough.

[orizi]

Reviewed 1 of 1 files at r2.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @TomerStarkware)