Rust: fix reading lists from options.yml

Author: Paolo Tranquilli

rust/extractor/src/config.rs

@@ -1,6 +1,9 @@
+mod deserialize_vec;
+
 use anyhow::Context;
 use clap::Parser;
 use codeql_extractor::trap;
+use deserialize_vec::deserialize_newline_or_comma_separated;
 use figment::{
     providers::{Env, Format, Serialized, Yaml},
     value::Value,
@@ -14,7 +17,7 @@ use ra_ap_intern::Symbol;
 use ra_ap_paths::Utf8PathBuf;
 use ra_ap_project_model::{CargoConfig, CargoFeatures, CfgOverrides, RustLibSource};
 use rust_extractor_macros::extractor_cli_config;
-use serde::{Deserialize, Deserializer, Serialize};
+use serde::{Deserialize, Serialize};
 use std::fmt::Debug;
 use std::ops::Not;
 use std::path::PathBuf;
@@ -37,14 +40,6 @@ impl From<Compression> for trap::Compression {
     }
 }
 
-// required by the extractor_cli_config macro.
-fn deserialize_newline_or_comma_separated<'a, D: Deserializer<'a>, T: for<'b> From<&'b str>>(
-    deserializer: D,
-) -> Result<Vec<T>, D::Error> {
-    let value = String::deserialize(deserializer)?;
-    Ok(value.split(['\n', ',']).map(T::from).collect())
-}
-
 #[extractor_cli_config]
 pub struct Config {
     pub scratch_dir: PathBuf,

rust/extractor/src/config/deserialize_vec.rs

@@ -0,0 +1,49 @@
+use serde::de::Visitor;
+use serde::Deserializer;
+use std::fmt::Formatter;
+use std::marker::PhantomData;
+
+// phantom data ise required to allow parametrizing on `T` without actual `T` data
+struct VectorVisitor<T: From<String>>(PhantomData<T>);
+
+impl<T: From<String>> VectorVisitor<T> {
+    fn new() -> Self {
+        VectorVisitor(PhantomData)
+    }
+}
+
+impl<'de, T: From<String>> Visitor<'de> for VectorVisitor<T> {
+    type Value = Vec<T>;
+
+    fn expecting(&self, formatter: &mut Formatter) -> std::fmt::Result {
+        formatter.write_str("either a sequence, or a comma or newline separated string")
+    }
+
+    fn visit_str<E: serde::de::Error>(self, value: &str) -> Result<Vec<T>, E> {
+        Ok(value
+            .split(['\n', ','])
+            .map(|s| T::from(s.to_owned()))
+            .collect())
+    }
+
+    fn visit_seq<A>(self, mut seq: A) -> Result<Vec<T>, A::Error>
+    where
+        A: serde::de::SeqAccess<'de>,
+    {
+        let mut ret = Vec::new();
+        while let Some(el) = seq.next_element::<String>()? {
+            ret.push(T::from(el));
+        }
+        Ok(ret)
+    }
+}
+
+/// deserialize into a vector of `T` either of:
+/// * a sequence of elements serializable into `String`s, or
+/// * a single element serializable into `String`, then split on `,` and `\n`
+/// This is required to be in scope when the `extractor_cli_config` macro is used.
+pub(crate) fn deserialize_newline_or_comma_separated<'a, D: Deserializer<'a>, T: From<String>>(
+    deserializer: D,
+) -> Result<Vec<T>, D::Error> {
+    deserializer.deserialize_seq(VectorVisitor::new())
+}