Adds sanitize_html, a whitelist based HTML sanitizer. (#171)

Cargo.toml

@@ -64,6 +64,7 @@ pathfinding = { version = "4.9", optional = true }
 num-integer = { version = "0.1.46", optional = true }
 dmi = { version = "0.3.5", optional = true }
 tracy_full = { version = "1.7.1", optional = true }
+ammonia = { version = "4.0.0", optional = true }
 
 [features]
 default = [
@@ -77,6 +78,7 @@ default = [
     "log",
     "noise",
     "rustls_tls",
+    "sanitize",
     "sql",
     "time",
     "toml",
@@ -95,6 +97,7 @@ all = [
     "log",
     "noise",
     "rustls_tls",
+    "sanitize",
     "sql",
     "time",
     "toml",
@@ -117,6 +120,7 @@ git = ["gix", "chrono"]
 http = ["reqwest", "serde", "serde_json", "once_cell", "jobs"]
 json = ["serde", "serde_json"]
 log = ["chrono"]
+sanitize = ["ammonia", "serde_json"]
 sql = ["mysql", "serde", "serde_json", "once_cell", "dashmap", "jobs"]
 time = []
 toml = ["serde", "serde_json", "toml-dep"]

dmsrc/sanitize.dm

@@ -0,0 +1,9 @@
+/*
+ * Takes in a string and json_encode()"d lists to produce a sanitized string.
+ * This function operates on whitelists, there is currently no way to blacklist.
+ * Args:
+ * * text: the string to sanitize.
+ * * attribute_whitelist_json: a json_encode()'d list of HTML attributes to allow in the final string.
+ * * tag_whitelist_json: a json_encode()'d list of HTML tags to allow in the final string.
+ */
+#define rustg_sanitize_html(text, attribute_whitelist_json, tag_whitelist_json) RUSTG_CALL(RUST_G, "sanitize_html")(text, attribute_whitelist_json, tag_whitelist_json)

src/lib.rs

@@ -38,6 +38,8 @@ pub mod pathfinder;
 pub mod redis_pubsub;
 #[cfg(feature = "redis_reliablequeue")]
 pub mod redis_reliablequeue;
+#[cfg(feature = "sanitize")]
+pub mod sanitize;
 #[cfg(feature = "sql")]
 pub mod sql;
 #[cfg(feature = "time")]

src/sanitize.rs

@@ -0,0 +1,32 @@
+use crate::error::Result;
+use std::collections::HashSet;
+
+byond_fn!(fn sanitize_html(text, attribute_whitelist_json, tag_whitelist_json) {
+    match seriously_sanitize_html(text, attribute_whitelist_json, tag_whitelist_json) {
+        Ok(r) => return Some(r),
+        Err(e) => return Some(e.to_string())
+    }
+});
+
+fn seriously_sanitize_html(
+    text: &str,
+    attribute_whitelist_json: &str,
+    tag_whitelist_json: &str,
+) -> Result<String> {
+    let attribute_whitelist: HashSet<&str> = serde_json::from_str(attribute_whitelist_json)?;
+    let tag_whitelist: HashSet<&str> = serde_json::from_str(tag_whitelist_json)?;
+
+    let mut prune_url_schemes = ammonia::Builder::default().clone_url_schemes();
+    prune_url_schemes.insert("byond");
+
+    let sanitized = ammonia::Builder::empty()
+        .clean_content_tags(HashSet::from_iter(["script", "style"])) // Completely forbid script and style attributes.
+        .link_rel(Some("noopener")) // https://mathiasbynens.github.io/rel-noopener/
+        .url_schemes(prune_url_schemes)
+        .generic_attributes(attribute_whitelist)
+        .tags(tag_whitelist)
+        .clean(text)
+        .to_string();
+
+    Ok(sanitized)
+}