From b475ff40e5019a8af92b2d161c01f1bcab2b823d Mon Sep 17 00:00:00 2001
From: daimond113 <72147841+daimond113@users.noreply.github.com>
Date: Thu, 28 Nov 2024 18:18:40 +0100
Subject: [PATCH] feat: support specifying allowed external registries in
 config

---
 CHANGELOG.md                              |  1 +
 registry/src/endpoints/publish_version.rs | 24 ++------------
 src/source/pesde/mod.rs                   | 38 ++++++++++++++++++++++-
 src/util.rs                               | 11 ++++++-
 4 files changed, 51 insertions(+), 23 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 30645fb..23f08f2 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -12,6 +12,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Make `includes` use glob patterns by @daimond113
 - Use symlinks for workspace dependencies to not require reinstalling by @daimond113
 - Add `auth token` command to print the auth token for the index by @daimond113
+- Support specifying which external registries are allowed on registries by @daimond113
 
 ### Fixed
 - Install dependencies of packages in `x` command
diff --git a/registry/src/endpoints/publish_version.rs b/registry/src/endpoints/publish_version.rs
index 0b51c79..cae611b 100644
--- a/registry/src/endpoints/publish_version.rs
+++ b/registry/src/endpoints/publish_version.rs
@@ -66,25 +66,6 @@ struct DocEntryInfo {
     collapsed: bool,
 }
 
-fn compare_repo_urls(this: &gix::Url, external: &gix::Url) -> bool {
-    let this = this.to_bstring().to_string().to_lowercase();
-    let external = external.to_bstring().to_string().to_lowercase();
-
-    let this = if this.ends_with(".git") {
-        &this[..this.len() - 4]
-    } else {
-        &this
-    };
-
-    let external = if external.ends_with(".git") {
-        &external[..external.len() - 4]
-    } else {
-        &external
-    };
-
-    this == external
-}
-
 pub async fn publish_package(
     app_state: web::Data<AppState>,
     bytes: Bytes,
@@ -321,8 +302,9 @@ pub async fn publish_package(
                         .index
                         .as_deref()
                         .filter(|index| match gix::Url::try_from(*index) {
-                            Ok(_) if config.other_registries_allowed => true,
-                            Ok(url) => compare_repo_urls(source.repo_url(), &url),
+                            Ok(url) => config
+                                .other_registries_allowed
+                                .is_allowed(source.repo_url().clone(), url),
                             Err(_) => false,
                         })
                         .is_none()
diff --git a/src/source/pesde/mod.rs b/src/source/pesde/mod.rs
index e8c4667..919b302 100644
--- a/src/source/pesde/mod.rs
+++ b/src/source/pesde/mod.rs
@@ -265,6 +265,42 @@ fn default_archive_size() -> usize {
     4 * 1024 * 1024
 }
 
+/// The allowed registries for a package
+#[derive(Deserialize, Debug, Clone)]
+#[serde(untagged)]
+pub enum AllowedRegistries {
+    /// All registries are allowed
+    All(bool),
+    /// Only specific registries are allowed
+    #[serde(deserialize_with = "crate::util::deserialize_gix_url_hashset")]
+    Specific(HashSet<Url>),
+}
+
+impl Default for AllowedRegistries {
+    fn default() -> Self {
+        Self::All(false)
+    }
+}
+
+impl AllowedRegistries {
+    /// Whether the given URL is allowed
+    pub fn is_allowed(&self, mut this: Url, mut external: Url) -> bool {
+        // strip .git suffix to allow for more flexible matching
+        this.path = this.path.strip_suffix(b".git").unwrap_or(&this.path).into();
+        external.path = external
+            .path
+            .strip_suffix(b".git")
+            .unwrap_or(&external.path)
+            .into();
+
+        this == external
+            || (match self {
+                Self::All(all) => *all,
+                Self::Specific(urls) => urls.contains(&this) || urls.contains(&external),
+            })
+    }
+}
+
 /// The configuration for the pesde index
 #[derive(Deserialize, Debug, Clone)]
 #[serde(deny_unknown_fields)]
@@ -278,7 +314,7 @@ pub struct IndexConfig {
     pub git_allowed: bool,
     /// Whether other registries are allowed as a source for publishing packages
     #[serde(default)]
-    pub other_registries_allowed: bool,
+    pub other_registries_allowed: AllowedRegistries,
     /// Whether Wally is allowed as a source for publishing packages
     #[serde(default)]
     pub wally_allowed: bool,
diff --git a/src/util.rs b/src/util.rs
index b52fe1d..3baca73 100644
--- a/src/util.rs
+++ b/src/util.rs
@@ -2,7 +2,7 @@ use crate::AuthConfig;
 use gix::bstr::BStr;
 use serde::{Deserialize, Deserializer, Serializer};
 use sha2::{Digest, Sha256};
-use std::collections::BTreeMap;
+use std::collections::{BTreeMap, HashSet};
 
 pub fn authenticate_conn(
     conn: &mut gix::remote::Connection<
@@ -59,6 +59,15 @@ pub fn deserialize_gix_url_vec<'de, D: Deserializer<'de>>(
         .collect()
 }
 
+pub fn deserialize_gix_url_hashset<'de, D: Deserializer<'de>>(
+    deserializer: D,
+) -> Result<HashSet<gix::Url>, D::Error> {
+    HashSet::<String>::deserialize(deserializer)?
+        .into_iter()
+        .map(|v| gix::Url::from_bytes(BStr::new(&v)).map_err(serde::de::Error::custom))
+        .collect()
+}
+
 pub fn deserialize_git_like_url<'de, D: Deserializer<'de>>(
     deserializer: D,
 ) -> Result<gix::Url, D::Error> {