chore: add deserialization tests with corrupted inputs

[nsarlin-zama]

closes: please link all relevant issues

PR content/description

This PR adds a test that runs known corrupted ct lists to check they are correctly deserialized (meaning errors are caught but does not trigger a panic).

The lfs part (pull, github cache, ...) has been copied from the backward data tests.

---

This change is 

[IceTDrinker]

Looks reasonable, I let the others take a look

thanks !

@IceTDrinker reviewed 21 files and all commit messages, and made 3 comments.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on nsarlin-zama, soonum, and SouchonTheo).

---

tests/corrupted_inputs_deserialization.rs line 117 at r1 (raw file):

}

fn handle_ct_list(input: &[u8], conformance_params: &CompactCiphertextListConformanceParams) {

could be worth printing to see up to where a list progresses given most of them likely won't make it passed deserialization

same for the proven stuff below

---

tests/corrupted_inputs_deserialization.rs line 161 at r1 (raw file):

#[test]
#[cfg(feature = "integer")]

I think we can get rid of that feature, to group once #3380 is merged to get rid of ZK and shortint features too I think

[nsarlin-zama]

• I added the strings feature to tfhe because one of the samples has a string
• also added a black_box to explicitly states that the ops should not be optimized away

@nsarlin-zama made 3 comments.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on IceTDrinker, soonum, and SouchonTheo).

---

tests/corrupted_inputs_deserialization.rs line 117 at r1 (raw file):

Previously, IceTDrinker wrote…

could be worth printing to see up to where a list progresses given most of them likely won't make it passed deserialization

same for the proven stuff below

yes good idea

---

tests/corrupted_inputs_deserialization.rs line 161 at r1 (raw file):

Previously, IceTDrinker wrote…

I think we can get rid of that feature, to group once #3380 is merged to get rid of ZK and shortint features too I think

I removed it here so that it's less work to do it later

[IceTDrinker]

Looks good to me, should we have someone else review the changes ?

Otherwise I can approve

@IceTDrinker partially reviewed 5 files and all commit messages, made 1 comment, and resolved 2 discussions.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on IceTDrinker, mayeul-zama, soonum, SouchonTheo, and tmontaigu).

[tmontaigu]

@tmontaigu partially reviewed 21 files and all commit messages, and made 3 comments.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on mayeul-zama, nsarlin-zama, soonum, and SouchonTheo).

---

tests/corrupted_inputs_deserialization.rs line 60 at r4 (raw file):

fn process_inputs(dir: &Path, mut handler: impl FnMut(&[u8])) -> u64 {
    let mut total_tests = 0;
    let entries: Vec<_> = std::fs::read_dir(dir)

minor, but I dont think we actually need to collect ?

---

tests/corrupted_inputs_deserialization.rs line 71 at r4 (raw file):

            total_tests += 1;
            let mut input = Vec::new();
            File::open(&path)

std::fs:read exists as a convenient wrapper of this
https://doc.rust-lang.org/std/fs/fn.read.html

---

tests/corrupted_inputs_deserialization.rs line 82 at r4 (raw file):

}

macro_rules! test_type {

Minor since this is a test, but it can be done without macro

fn test_integer<FheType>(exp: &CompactCiphertextListExpander, i: usize)
where
    FheType: HlExpandable
        + Tagged
        + Clone
        + Add<FheType, Output = FheType>
        + Sub<FheType, Output = FheType>
        + Mul<FheType, Output = FheType>,
{
    let ct = match exp.get::<FheType>(i) {
        Ok(Some(ct)) => ct,
        Ok(None) => {
            println!("No ct found at idx {}\n", i);
            return;
        }
        Err(e) => {
            println!("Error caught while trying to get ct at idx {}:\n{e}\n", i);
            return;
        }
    };
    let res = ct.clone() + ct.clone();
    let res = res * ct.clone();
    let _ = std::hint::black_box(res - ct);
}

[nsarlin-zama]

@nsarlin-zama made 3 comments and resolved 1 discussion.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on mayeul-zama, soonum, SouchonTheo, and tmontaigu).

---

tests/corrupted_inputs_deserialization.rs line 60 at r4 (raw file):

Previously, tmontaigu (tmontaigu) wrote…

minor, but I dont think we actually need to collect ?

done

---

tests/corrupted_inputs_deserialization.rs line 71 at r4 (raw file):

Previously, tmontaigu (tmontaigu) wrote…

std::fs:read exists as a convenient wrapper of this
https://doc.rust-lang.org/std/fs/fn.read.html

done

---

tests/corrupted_inputs_deserialization.rs line 82 at r4 (raw file):

Previously, tmontaigu (tmontaigu) wrote…

Minor since this is a test, but it can be done without macro

fn test_integer<FheType>(exp: &CompactCiphertextListExpander, i: usize)
where
    FheType: HlExpandable
        + Tagged
        + Clone
        + Add<FheType, Output = FheType>
        + Sub<FheType, Output = FheType>
        + Mul<FheType, Output = FheType>,
{
    let ct = match exp.get::<FheType>(i) {
        Ok(Some(ct)) => ct,
        Ok(None) => {
            println!("No ct found at idx {}\n", i);
            return;
        }
        Err(e) => {
            println!("Error caught while trying to get ct at idx {}:\n{e}\n", i);
            return;
        }
    };
    let res = ct.clone() + ct.clone();
    let res = res * ct.clone();
    let _ = std::hint::black_box(res - ct);
}

I'm always happy when I can avoid a macro :)
done

[tmontaigu]

@tmontaigu reviewed 1 file and all commit messages, and resolved 2 discussions.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on mayeul-zama, soonum, and SouchonTheo).

[IceTDrinker]

@IceTDrinker reviewed 5 files and all commit messages, and made 1 comment.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on mayeul-zama, soonum, and SouchonTheo).