[perf] detect_text_format() should check for valid utf-8 once

Also, turn of ASAN for distributed slices.

Author: tstack

src/base/distributed_slice.hh

@@ -36,7 +36,8 @@
 
 #if defined(__APPLE__)
 #    define DIST_SLICE(id) \
-        __attribute__((used, \
+        __attribute__((no_sanitize_address, \
+                       used, \
                        section("__DATA,__ds_" DS_STRINGIZE( \
                            id) ",regular,no_dead_strip")))
 #    define DIST_SLICE_BEGIN(id) \
@@ -45,7 +46,8 @@
         __asm("section$end$__DATA$__ds_" DS_STRINGIZE(id))
 #else
 #    define DIST_SLICE(id) \
-        __attribute__((used, section("ds_" DS_STRINGIZE(id))))
+        __attribute__(( \
+            no_sanitize_address, used, section("ds_" DS_STRINGIZE(id))))
 #    define DIST_SLICE_BEGIN(id) __asm("__start_ds_" DS_STRINGIZE(id))
 #    define DIST_SLICE_END(id)   __asm("__stop_ds_" DS_STRINGIZE(id))
 #endif
@@ -91,15 +93,9 @@ struct dist_slice_container {
             return this->sia_array[index];
         }
 
-        iterator begin()
-        {
-            return this->sia_array.begin();
-        }
+        iterator begin() { return this->sia_array.begin(); }
 
-        iterator end()
-        {
-            return this->sia_array.end();
-        }
+        iterator end() { return this->sia_array.end(); }
 
         void clear()
         {

src/base/intern_string.hh

@@ -620,14 +620,15 @@ struct string_fragment {
             if (this->qs_in_escape) {
                 this->qs_in_escape = false;
                 return true;
-            } else if (ch == '\\') {
+            }
+            if (ch == '\\') {
                 this->qs_in_escape = true;
                 return true;
-            } else if (ch == '"') {
+            }
+            if (ch == '"') {
                 return false;
-            } else {
-                return true;
             }
+            return true;
         }
     };
 
@@ -715,7 +716,7 @@ struct string_fragment {
             static_cast<std::string_view::size_type>(this->length())};
     }
 
-    enum class case_style {
+    enum class case_style : uint8_t {
         lower,
         upper,
         camel,
@@ -732,8 +733,8 @@ struct string_fragment {
     }
 
     const char* sf_string;
-    int sf_begin;
-    int sf_end;
+    int32_t sf_begin;
+    int32_t sf_end;
 };
 
 inline bool

src/bookmarks.hh

@@ -184,13 +184,18 @@ public:
     static std::vector<string_fragment> get_type_names();
 
     template<typename T, std::size_t N>
-    explicit bookmark_type_t(const T (&name)[N])
+    constexpr explicit bookmark_type_t(const T (&name)[N])
         : bt_name(string_fragment::from_const(name))
     {
     }
 
+    ~bookmark_type_t() = default;
+
     bookmark_type_t(const bookmark_type_t&) = delete;
-    bookmark_type_t(const bookmark_type_t&&) = delete;
+    bookmark_type_t(bookmark_type_t&&) = delete;
+
+    bookmark_type_t& operator=(const bookmark_type_t&) = delete;
+    bookmark_type_t& operator=(bookmark_type_t&&) = delete;
 
     const string_fragment& get_name() const { return this->bt_name; }
 

src/lnav.ico

@@ -34,6 +34,7 @@
 #include "text_format.hh"
 
 #include "base/from_trait.hh"
+#include "base/is_utf8.hh"
 #include "base/itertools.enumerate.hh"
 #include "base/itertools.hh"
 #include "base/lnav_log.hh"
@@ -183,6 +184,11 @@ detect_text_format(string_fragment sf,
 )",
         PCRE2_MULTILINE | PCRE2_CASELESS | PCRE2_EXTENDED);
 
+    auto utf_res = is_utf8(sf);
+    if (!utf_res.is_valid()) {
+        return text_format_t::TF_UNKNOWN;
+    }
+
     if (path) {
         while (FILTER_EXTS.count(path->extension()) > 0) {
             path = path->stem();
@@ -256,47 +262,47 @@ detect_text_format(string_fragment sf,
         }
     }
 
-    if (DIFF_MATCHERS.find_in(sf).ignore_error()) {
+    if (DIFF_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_DIFF;
     }
 
-    if (SH_MATCHERS.find_in(sf).ignore_error()) {
+    if (SH_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_SHELL_SCRIPT;
     }
 
-    if (MAN_MATCHERS.find_in(sf).ignore_error()) {
+    if (MAN_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_MAN;
     }
 
-    if (PYTHON_MATCHERS.find_in(sf).ignore_error()) {
+    if (PYTHON_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_PYTHON;
     }
 
-    if (RUST_MATCHERS.find_in(sf).ignore_error()) {
+    if (RUST_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_RUST;
     }
 
-    if (JAVA_MATCHERS.find_in(sf).ignore_error()) {
+    if (JAVA_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_JAVA;
     }
 
-    if (C_LIKE_MATCHERS.find_in(sf).ignore_error()) {
+    if (C_LIKE_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_C_LIKE;
     }
 
-    if (LNAV_MATCHERS.find_in(sf).ignore_error()) {
+    if (LNAV_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_LNAV_SCRIPT;
     }
 
-    if (SQL_MATCHERS.find_in(sf).ignore_error()) {
+    if (SQL_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_SQL;
     }
 
-    if (XML_MATCHERS.find_in(sf).ignore_error()) {
+    if (XML_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_XML;
     }
 
-    if (INI_MATCHERS.find_in(sf).ignore_error()) {
+    if (INI_MATCHERS.find_in(sf, PCRE2_NO_UTF_CHECK).ignore_error()) {
         return text_format_t::TF_INI;
     }