chore(capi+go): don't enable rules profiling by default.

Also improves the profiling-related documentation in the Go library.

capi/Cargo.toml

@@ -12,7 +12,7 @@ homepage.workspace = true
 
 [features]
 # The `capi` feature is required by `cargo-c`.
-default = ["capi", "rules-profiling"]
+default = ["capi"]
 capi = []
 
 # When enabled, the serialization of compiled rules include native code for
@@ -29,7 +29,7 @@ native-code-serialization = ["yara-x/native-code-serialization"]
 
 # Enables rules profiling.
 #
-# This feature is enabled by default.
+# This feature is disabled by default.
 rules-profiling = ["yara-x/rules-profiling"]
 
 

capi/include/yara_x.h

@@ -80,6 +80,9 @@ typedef enum YRX_RESULT {
   SERIALIZATION_ERROR,
   // An error returned when a rule doesn't have any metadata.
   NO_METADATA,
+  // An error returned in cases where some API is not supported because the
+  // library was not built with the required features.
+  NOT_SUPPORTED,
 } YRX_RESULT;
 
 // A compiler that takes YARA source code and produces compiled rules.
@@ -708,7 +711,7 @@ enum YRX_RESULT yrx_scanner_set_global_float(struct YRX_SCANNER *scanner,
 // Iterates over the top N most expensive rules, calling the callback for
 // each rule.
 //
-// Requires the `rules-profiling` feature.
+// Requires the `rules-profiling` feature, otherwise the
 //
 // See [`YRX_MOST_EXPENSIVE_RULES_CALLBACK`] for more details.
 enum YRX_RESULT yrx_scanner_iter_most_expensive_rules(struct YRX_SCANNER *scanner,

capi/src/lib.rs

@@ -152,6 +152,9 @@ pub enum YRX_RESULT {
     SERIALIZATION_ERROR,
     /// An error returned when a rule doesn't have any metadata.
     NO_METADATA,
+    /// An error returned in cases where some API is not supported because the
+    /// library was not built with the required features.
+    NOT_SUPPORTED,
 }
 
 /// Returns the error message for the most recent function in this API

capi/src/scanner.rs

@@ -61,11 +61,11 @@ pub unsafe extern "C" fn yrx_scanner_set_timeout(
     scanner: *mut YRX_SCANNER,
     timeout: u64,
 ) -> YRX_RESULT {
-    if scanner.is_null() {
-        return YRX_RESULT::INVALID_ARGUMENT;
-    }
+    let scanner = match scanner.as_mut() {
+        Some(s) => s,
+        None => return YRX_RESULT::INVALID_ARGUMENT,
+    };
 
-    let scanner = scanner.as_mut().unwrap();
     scanner.inner.set_timeout(Duration::from_secs(timeout));
 
     YRX_RESULT::SUCCESS
@@ -84,16 +84,16 @@ pub unsafe extern "C" fn yrx_scanner_scan(
 ) -> YRX_RESULT {
     _yrx_set_last_error::<ScanError>(None);
 
-    if scanner.is_null() {
-        return YRX_RESULT::INVALID_ARGUMENT;
-    }
+    let scanner = match scanner.as_mut() {
+        Some(s) => s,
+        None => return YRX_RESULT::INVALID_ARGUMENT,
+    };
 
     let data = match slice_from_ptr_and_len(data, len) {
         Some(data) => data,
         None => return YRX_RESULT::INVALID_ARGUMENT,
     };
 
-    let scanner = scanner.as_mut().unwrap();
     let scan_results = scanner.inner.scan(data);
 
     if let Err(err) = scan_results {
@@ -178,9 +178,10 @@ pub unsafe extern "C" fn yrx_scanner_set_module_output(
     data: *const u8,
     len: usize,
 ) -> YRX_RESULT {
-    if scanner.is_null() {
-        return YRX_RESULT::INVALID_ARGUMENT;
-    }
+    let scanner = match scanner.as_mut() {
+        Some(s) => s,
+        None => return YRX_RESULT::INVALID_ARGUMENT,
+    };
 
     let module_name = match CStr::from_ptr(name).to_str() {
         Ok(name) => name,
@@ -195,8 +196,6 @@ pub unsafe extern "C" fn yrx_scanner_set_module_output(
         None => return YRX_RESULT::INVALID_ARGUMENT,
     };
 
-    let scanner = scanner.as_mut().unwrap();
-
     match scanner.inner.set_module_output_raw(module_name, data) {
         Ok(_) => {
             _yrx_set_last_error::<ScanError>(None);
@@ -216,9 +215,10 @@ unsafe extern "C" fn yrx_scanner_set_global<
     ident: *const c_char,
     value: T,
 ) -> YRX_RESULT {
-    if scanner.is_null() {
-        return YRX_RESULT::INVALID_ARGUMENT;
-    }
+    let scanner = match scanner.as_mut() {
+        Some(s) => s,
+        None => return YRX_RESULT::INVALID_ARGUMENT,
+    };
 
     let ident = match CStr::from_ptr(ident).to_str() {
         Ok(ident) => ident,
@@ -228,8 +228,6 @@ unsafe extern "C" fn yrx_scanner_set_global<
         }
     };
 
-    let scanner = scanner.as_mut().unwrap();
-
     match scanner.inner.set_global(ident, value) {
         Ok(_) => {
             _yrx_set_last_error::<ScanError>(None);
@@ -327,35 +325,40 @@ pub type YRX_MOST_EXPENSIVE_RULES_CALLBACK = extern "C" fn(
 /// Iterates over the top N most expensive rules, calling the callback for
 /// each rule.
 ///
-/// Requires the `rules-profiling` feature.
+/// Requires the `rules-profiling` feature, otherwise the
 ///
 /// See [`YRX_MOST_EXPENSIVE_RULES_CALLBACK`] for more details.
-#[cfg(feature = "rules-profiling")]
 #[no_mangle]
+#[allow(unused_variables)]
 pub unsafe extern "C" fn yrx_scanner_iter_most_expensive_rules(
     scanner: *mut YRX_SCANNER,
     n: usize,
     callback: YRX_MOST_EXPENSIVE_RULES_CALLBACK,
     user_data: *mut c_void,
 ) -> YRX_RESULT {
-    if scanner.is_null() {
-        return YRX_RESULT::INVALID_ARGUMENT;
-    }
-
-    let scanner = scanner.as_ref().unwrap();
+    #[cfg(not(feature = "rules-profiling"))]
+    return YRX_RESULT::NOT_SUPPORTED;
+
+    #[cfg(feature = "rules-profiling")]
+    {
+        let scanner = match scanner.as_ref() {
+            Some(s) => s,
+            None => return YRX_RESULT::INVALID_ARGUMENT,
+        };
 
-    for profiling_info in scanner.inner.most_expensive_rules(n) {
-        let namespace = CString::new(profiling_info.namespace).unwrap();
-        let rule = CString::new(profiling_info.rule).unwrap();
+        for profiling_info in scanner.inner.most_expensive_rules(n) {
+            let namespace = CString::new(profiling_info.namespace).unwrap();
+            let rule = CString::new(profiling_info.rule).unwrap();
+
+            callback(
+                namespace.as_ptr(),
+                rule.as_ptr(),
+                profiling_info.pattern_matching_time.as_secs_f64(),
+                profiling_info.condition_exec_time.as_secs_f64(),
+                user_data,
+            );
+        }
 
-        callback(
-            namespace.as_ptr(),
-            rule.as_ptr(),
-            profiling_info.pattern_matching_time.as_secs_f64(),
-            profiling_info.condition_exec_time.as_secs_f64(),
-            user_data,
-        );
+        YRX_RESULT::SUCCESS
     }
-
-    YRX_RESULT::SUCCESS
 }

go/scanner.go

@@ -247,6 +247,13 @@ func (s *Scanner) Scan(buf []byte) (*ScanResults, error) {
 	return scanResults, err
 }
 
+// ProfilingInfo contains profiling information about a YARA rule.
+//
+// For each rule it contains: the rule's namespace, the rule's name,
+// the time spent in matching patterns declared by the rule, and the time
+// spent evaluating the rule's condition.
+//
+// See [Scanner.MostExpensiveRules].
 type ProfilingInfo struct {
 	Namespace           string
 	Rule                string
@@ -276,16 +283,28 @@ func mostExpensiveRulesCallback(
 	})
 }
 
+// MostExpensiveRules returns information about the slowest rules and how much
+// time they spent matching patterns and executing their conditions.
+//
+// In order to use this function the YARA-X C library must be built with
+// support for rules profiling, which is done by enabling the `rules-profiling`
+// feature. Otherwise, calling this function will cause a panic.
 func (s *Scanner) MostExpensiveRules(n int) []ProfilingInfo {
 	profilingInfo := make([]ProfilingInfo, 0)
 	mostExpensiveRules := cgo.NewHandle(&profilingInfo)
 	defer mostExpensiveRules.Delete()
 
-	if C._yrx_scanner_iter_most_expensive_rules(
+	result := C._yrx_scanner_iter_most_expensive_rules(
 		s.cScanner,
 		C.size_t(n),
 		C.YRX_MOST_EXPENSIVE_RULES_CALLBACK(C.mostExpensiveRulesCallback),
-		C.uintptr_t(mostExpensiveRules)) != C.SUCCESS {
+		C.uintptr_t(mostExpensiveRules))
+
+	if result == C.NOT_SUPPORTED {
+		panic("MostExpensiveRules requires that the YARA-X C library is built with the `rules-profiling` feature")
+	}
+
+	if result != C.SUCCESS {
 		panic("yrx_scanner_iter_most_expensive_rules failed")
 	}
 

go/scanner_test.go

@@ -91,18 +91,6 @@ func TestScannerTimeout(t *testing.T) {
 	assert.ErrorIs(t, err, ErrTimeout)
 }
 
-func TestScannerMostExpensiveRules(t *testing.T) {
-	r, _ := Compile("rule t { strings: $a = /a(.*)*a/ condition: $a }")
-	s := NewScanner(r)
-	_, err := s.Scan(bytes.Repeat([]byte("a"), 5000))
-	assert.NoError(t, err)
-	profilingInfo := s.MostExpensiveRules(1)
-	assert.Equal(t, "t", profilingInfo[0].Rule)
-	assert.Equal(t, "default", profilingInfo[0].Namespace)
-	assert.Greater(t, profilingInfo[0].PatternMatchingTime, time.Duration(0))
-	assert.Greater(t, profilingInfo[0].ConditionExecTime, time.Duration(0))
-}
-
 func TestScannerMetadata(t *testing.T) {
 	r, _ := Compile(`rule t {
 			meta: