make source name validation consistent with subgraph name validation (#…

…6219)

Cargo.lock

@@ -193,8 +193,10 @@ dependencies = [
  "multimap 0.10.0",
  "nom",
  "nom_locate",
+ "once_cell",
  "petgraph",
  "pretty_assertions",
+ "regex",
  "ron",
  "rstest",
  "serde",

apollo-federation/Cargo.toml

@@ -29,7 +29,9 @@ lazy_static = "1.4.0"
 line-col = "0.2.1"
 multimap = "0.10.0"
 nom = "7.1.3"
+once_cell = "1.19.0"
 petgraph = { version = "0.6.4", features = ["serde-1"] }
+regex = "1.10.5"
 serde.workspace = true
 serde_json.workspace = true
 serde_json_bytes.workspace = true

apollo-federation/src/sources/connect/validation/snapshots/validation_tests@invalid_chars_in_source_name.graphql.snap

@@ -6,9 +6,30 @@ input_file: apollo-federation/src/sources/connect/validation/test_data/invalid_c
 [
     Message {
         code: InvalidSourceName,
-        message: "There are invalid characters in `@source(name: \"u$ers\")`. Only alphanumeric and underscores are allowed.",
+        message: "`@source(name: \"u$ers\")` is invalid; all source names must follow pattern '^[a-zA-Z][a-zA-Z0-9_-]{0,63}$",
         locations: [
             6:17..6:24,
         ],
     },
+    Message {
+        code: InvalidSourceName,
+        message: "`@source(name: \"1\")` is invalid; all source names must follow pattern '^[a-zA-Z][a-zA-Z0-9_-]{0,63}$",
+        locations: [
+            7:17..7:20,
+        ],
+    },
+    Message {
+        code: InvalidSourceName,
+        message: "`@source(name: \"no.dots\")` is invalid; all source names must follow pattern '^[a-zA-Z][a-zA-Z0-9_-]{0,63}$",
+        locations: [
+            8:17..8:26,
+        ],
+    },
+    Message {
+        code: InvalidSourceName,
+        message: "`@source(name: \"areallylongnamethatisoversixtythreecharacterstakesalongwhiletotypebutthisshoulddoit\")` is invalid; all source names must follow pattern '^[a-zA-Z][a-zA-Z0-9_-]{0,63}$",
+        locations: [
+            10:11..10:96,
+        ],
+    },
 ]

apollo-federation/src/sources/connect/validation/snapshots/validation_tests@multiple_errors.graphql.snap

@@ -13,7 +13,7 @@ input_file: apollo-federation/src/sources/connect/validation/test_data/multiple_
     },
     Message {
         code: InvalidSourceName,
-        message: "There are invalid characters in `@source(name: \"u$ers\")`. Only alphanumeric and underscores are allowed.",
+        message: "`@source(name: \"u$ers\")` is invalid; all source names must follow pattern '^[a-zA-Z][a-zA-Z0-9_-]{0,63}$",
         locations: [
             6:17..6:24,
         ],

apollo-federation/src/sources/connect/validation/source_name.rs

@@ -17,6 +17,19 @@ use super::Message;
 use crate::sources::connect::spec::schema::SOURCE_NAME_ARGUMENT_NAME;
 use crate::sources::connect::validation::graphql::SchemaInfo;
 
+// Adding a module to allow changing clippy lints for the regex
+#[allow(clippy::expect_used)]
+mod patterns {
+    use once_cell::sync::Lazy;
+    use regex::Regex;
+
+    /// This is the same regular expression used for subgraph names
+    pub(super) static SOURCE_NAME_REGEX: Lazy<Regex> = Lazy::new(|| {
+        Regex::new(r"^[a-zA-Z][a-zA-Z0-9_-]{0,63}$")
+            .expect("this regex to check source names is valid")
+    });
+}
+
 pub(super) fn validate_source_name_arg(
     field_name: &Name,
     object_name: &Name,
@@ -112,7 +125,7 @@ impl SourceName {
                 directive_name,
                 value: arg.value.clone(),
             }
-        } else if str_value.chars().all(|c| c.is_alphanumeric() || c == '_') {
+        } else if patterns::SOURCE_NAME_REGEX.is_match(str_value) {
             Self::Valid {
                 value: arg.value.clone(),
                 directive_name,
@@ -132,7 +145,8 @@ impl SourceName {
                 value,
                 directive_name,
             } => Err(Message {
-                message: format!("There are invalid characters in {coordinate}. Only alphanumeric and underscores are allowed.", coordinate = source_name_value_coordinate(&directive_name, &value)),
+                // This message is the same as Studio when trying to publish a subgraph with an invalid name
+                message: format!("{coordinate} is invalid; all source names must follow pattern '^[a-zA-Z][a-zA-Z0-9_-]{{0,63}}$", coordinate = source_name_value_coordinate(&directive_name, &value)),
                 code: Code::InvalidSourceName,
                 locations: value.line_column_range(sources).into_iter().collect(),
             }),

apollo-federation/src/sources/connect/validation/test_data/invalid_chars_in_source_name.graphql

@@ -4,8 +4,24 @@ extend schema
     import: ["@connect", "@source"]
   )
   @source(name: "u$ers", http: { baseURL: "http://127.0.0.1" })
+  @source(name: "1", http: { baseURL: "http://127.0.0.1" })
+  @source(name: "no.dots", http: { baseURL: "http://127.0.0.1" })
+  @source(
+    name: "areallylongnamethatisoversixtythreecharacterstakesalongwhiletotypebutthisshoulddoit"
+    http: { baseURL: "http://127.0.0.1" }
+  )
 
 type Query {
-  resources: [String!]!
+  resources1: [String!]!
     @connect(source: "u$ers", http: { GET: "/resources" }, selection: "$")
+  resources2: [String!]!
+    @connect(source: "1", http: { GET: "/resources" }, selection: "$")
+  resources3: [String!]!
+    @connect(source: "no.dots", http: { GET: "/resources" }, selection: "$")
+  resources4: [String!]!
+    @connect(
+      source: "areallylongnamethatisoversixtythreecharacterstakesalongwhiletotypebutthisshoulddoit"
+      http: { GET: "/resources" }
+      selection: "$"
+    )
 }