Provide more debug info

Author: atuchin-m

Cargo.toml

@@ -94,7 +94,7 @@ harness = false
 default = ["embedded-domain-resolver", "full-regex-handling", "single-thread"]
 full-regex-handling = []
 single-thread = [] # disables `Send` and `Sync` on `Engine`.
-regex-debug-info = []
+debug-info = []
 css-validation = ["cssparser", "selectors"]
 content-blocking = []
 embedded-domain-resolver = ["addr"] # Requires setting an external domain resolver if disabled.

src/blocker.rs

@@ -487,13 +487,13 @@ impl Blocker {
         regex_manager.set_discard_policy(new_discard_policy);
     }
 
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub fn discard_regex(&self, regex_id: u64) {
         let mut regex_manager = self.borrow_regex_manager();
         regex_manager.discard_regex(regex_id);
     }
 
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub fn get_regex_debug_info(&self) -> crate::regex_manager::RegexDebugInfo {
         let regex_manager = self.borrow_regex_manager();
         regex_manager.get_debug_info()

src/engine.rs

@@ -58,6 +58,14 @@ pub struct Engine {
     filter_data_context: FilterDataContextRef,
 }
 
+#[cfg(feature = "debug-info")]
+pub struct EngineDebugInfo {
+    pub regex_debug_info: crate::regex_manager::RegexDebugInfo,
+
+    pub flatbuffer_size: usize,
+    pub resources_total_length: usize,
+}
+
 impl Default for Engine {
     fn default() -> Self {
         Self::from_filter_set(FilterSet::new(false), false)
@@ -241,14 +249,18 @@ impl Engine {
         self.blocker.set_regex_discard_policy(new_discard_policy);
     }
 
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub fn discard_regex(&mut self, regex_id: u64) {
         self.blocker.discard_regex(regex_id);
     }
 
-    #[cfg(feature = "regex-debug-info")]
-    pub fn get_regex_debug_info(&self) -> crate::regex_manager::RegexDebugInfo {
-        self.blocker.get_regex_debug_info()
+    #[cfg(feature = "debug-info")]
+    pub fn get_debug_info(&self) -> EngineDebugInfo {
+        EngineDebugInfo {
+            regex_debug_info: self.blocker.get_regex_debug_info(),
+            flatbuffer_size: self.filter_data_context.memory.data().len(),
+            resources_total_length: self.resources.resources_total_length(),
+        }
     }
 
     /// Serializes the `Engine` into a binary format so that it can be quickly reloaded later.

src/regex_manager.rs

@@ -39,7 +39,7 @@ const DEFAULT_DISCARD_UNUSED_TIME: Duration = Duration::from_secs(180);
 
 /// Reports [`RegexManager`] metrics that may be useful for creating an optimized
 /// [`RegexManagerDiscardPolicy`].
-#[cfg(feature = "regex-debug-info")]
+#[cfg(feature = "debug-info")]
 pub struct RegexDebugInfo {
     /// Information about each regex contained in the [`RegexManager`].
     pub regex_data: Vec<RegexDebugEntry>,
@@ -48,7 +48,7 @@ pub struct RegexDebugInfo {
 }
 
 /// Describes metrics about a single regex from the [`RegexManager`].
-#[cfg(feature = "regex-debug-info")]
+#[cfg(feature = "debug-info")]
 pub struct RegexDebugEntry {
     /// Id for this particular regex, which is constant and unique for its lifetime.
     ///
@@ -312,7 +312,7 @@ impl RegexManager {
     }
 
     /// Discard one regex, identified by its id from a [`RegexDebugEntry`].
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub fn discard_regex(&mut self, regex_id: u64) {
         self.map
             .iter_mut()
@@ -322,7 +322,7 @@ impl RegexManager {
             });
     }
 
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub(crate) fn get_debug_regex_data(&self) -> Vec<RegexDebugEntry> {
         use itertools::Itertools;
         self.map
@@ -336,13 +336,13 @@ impl RegexManager {
             .collect_vec()
     }
 
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub(crate) fn get_compiled_regex_count(&self) -> usize {
         self.compiled_regex_count
     }
 
     /// Collect metrics that may be useful for creating an optimized [`RegexManagerDiscardPolicy`].
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     pub fn get_debug_info(&self) -> RegexDebugInfo {
         RegexDebugInfo {
             regex_data: self.get_debug_regex_data(),

tests/unit/engine.rs

@@ -218,6 +218,14 @@ mod tests {
         let mut engine = Engine::from_rules_parametrised(rules, Default::default(), false, true);
         let data = engine.serialize().to_vec();
 
+        #[cfg(feature = "debug-info")]
+        {
+            let debug_info = engine.get_debug_info();
+            let expected_size = 8_527_344_f32;
+            assert!(debug_info.flatbuffer_size >= (expected_size * 0.99) as usize);
+            assert!(debug_info.flatbuffer_size <= (expected_size * 1.01) as usize);
+        }
+
         let expected_hash: u64 = if cfg!(feature = "css-validation") {
             2942520321544562177
         } else {

tests/unit/filters/network_matchers.rs

@@ -678,7 +678,7 @@ mod match_tests {
 
     #[test]
     #[ignore] // Not going to handle lookaround regexes
-    #[cfg(feature = "regex-debug-info")]
+    #[cfg(feature = "debug-info")]
     fn check_lookaround_regex_handled() {
         {
             let filter = r#"/^https?:\/\/([0-9a-z\-]+\.)?(9anime|animeland|animenova|animeplus|animetoon|animewow|gamestorrent|goodanime|gogoanime|igg-games|kimcartoon|memecenter|readcomiconline|toonget|toonova|watchcartoononline)\.[a-z]{2,4}\/(?!([Ee]xternal|[Ii]mages|[Ss]cripts|[Uu]ploads|ac|ajax|assets|combined|content|cov|cover|(img\/bg)|(img\/icon)|inc|jwplayer|player|playlist-cat-rss|static|thumbs|wp-content|wp-includes)\/)(.*)/$image,other,script,~third-party,xmlhttprequest,domain=~animeland.hu"#;

tests/unit/regex_manager.rs

@@ -1,4 +1,4 @@
-#[cfg(all(test, feature = "regex-debug-info"))]
+#[cfg(all(test, feature = "debug-info"))]
 mod tests {
     use super::super::*;