fix(erlang): integer guard, cache dedup, dead macro cleanup

binding/erlang/README.md

@@ -158,11 +158,11 @@ System:
   CPU    : Apple M4
   Cores  : 10 cores / 10 threads
   Erlang : Erlang/OTP 29 [erts-17.0.2] [source] [64-bit] [smp:10:10] [ds:10:10:10] [async-threads:1] [jit] [dtrace]
-  Loaded : 487169 IPs in 1.574 s
+  Loaded : 487169 IPs in 1.335 s
 
 Benchmarks:
-  file      total=  9.460s  count= 487169  qps=    51500.40  avg= 0.019417 ms/op (19.417 us/op)
+  cold      total=  9.601s  count= 487169  qps=    50740.66  avg= 0.019708 ms/op (19.708 us/op)
-  cache     total=  0.150s  count= 487169  qps=  3245240.41  avg= 0.000308 ms/op ( 0.308 us/op)
+  warm      total=  0.160s  count= 487169  qps=  3053164.29  avg= 0.000328 ms/op ( 0.328 us/op)
 
 Done.
 ```
@@ -174,11 +174,11 @@ System:
   CPU    : Apple M4
   Cores  : 10 cores / 10 threads
   Erlang : Erlang/OTP 29 [erts-17.0.2] [source] [64-bit] [smp:10:10] [ds:10:10:10] [async-threads:1] [jit] [dtrace]
-  Loaded : 638953 IPs in 2.724 s
+  Loaded : 638953 IPs in 2.949 s
 
 Benchmarks:
-  file      total= 16.326s  count= 638953  qps=    39137.65  avg= 0.025551 ms/op (25.551 us/op)
+  cold      total= 20.504s  count= 638953  qps=    31162.52  avg= 0.032090 ms/op (32.090 us/op)
-  cache     total=  0.469s  count= 638953  qps=  1361972.38  avg= 0.000734 ms/op ( 0.734 us/op)
+  warm      total=  0.444s  count= 638953  qps=  1437781.56  avg= 0.000696 ms/op ( 0.696 us/op)
 
 Done.
 ```

binding/erlang/README_zh.md

@@ -135,11 +135,11 @@ System:
   CPU    : Apple M4
   Cores  : 10 cores / 10 threads
   Erlang : Erlang/OTP 29 [erts-17.0.2] [source] [64-bit] [smp:10:10] [ds:10:10:10] [async-threads:1] [jit] [dtrace]
-  Loaded : 487169 IPs in 1.574 s
+  Loaded : 487169 IPs in 1.335 s
 
 Benchmarks:
-  file      total=  9.460s  count= 487169  qps=    51500.40  avg= 0.019417 ms/op (19.417 us/op)
+  cold      total=  9.601s  count= 487169  qps=    50740.66  avg= 0.019708 ms/op (19.708 us/op)
-  cache     total=  0.150s  count= 487169  qps=  3245240.41  avg= 0.000308 ms/op ( 0.308 us/op)
+  warm      total=  0.160s  count= 487169  qps=  3053164.29  avg= 0.000328 ms/op ( 0.328 us/op)
 
 Done.
 ```
@@ -151,11 +151,11 @@ System:
   CPU    : Apple M4
   Cores  : 10 cores / 10 threads
   Erlang : Erlang/OTP 29 [erts-17.0.2] [source] [64-bit] [smp:10:10] [ds:10:10:10] [async-threads:1] [jit] [dtrace]
-  Loaded : 638953 IPs in 2.724 s
+  Loaded : 638953 IPs in 2.949 s
 
 Benchmarks:
-  file      total= 16.326s  count= 638953  qps=    39137.65  avg= 0.025551 ms/op (25.551 us/op)
+  cold      total= 20.504s  count= 638953  qps=    31162.52  avg= 0.032090 ms/op (32.090 us/op)
-  cache     total=  0.469s  count= 638953  qps=  1361972.38  avg= 0.000734 ms/op ( 0.734 us/op)
+  warm      total=  0.444s  count= 638953  qps=  1437781.56  avg= 0.000696 ms/op ( 0.696 us/op)
 
 Done.
 ```

binding/erlang/benchmarks/xdb-benchmark.sh

@@ -10,7 +10,7 @@ PROJECT_ROOT="$(cd ../.. && pwd)"
 if [ "$VERSION" = "ipv6" ]; then
     DB_FILE="${PROJECT_ROOT}/data/ip2region_v6.xdb"
     DATA_FILE="${PROJECT_ROOT}/data/ipv6_source.txt"
-    DB_CONFIG="[{ipv4, \"ip2region.xdb\"}, {ipv6, \"${DB_FILE}\"}]"
+    DB_CONFIG="[{ipv6, \"${DB_FILE}\"}]"
 elif [ "$VERSION" = "ipv4" ]; then
     DATA_FILE="${PROJECT_ROOT}/data/ipv4_source.txt"
     DB_CONFIG="[{ipv4, \"ip2region.xdb\"}]"

binding/erlang/include/ip2region.hrl

@@ -4,20 +4,11 @@
 -define(NONE, none).
 -define(APP_NAME, ip2region).
 
--define(XDB_VECTOR_INDEX, ets_xdb_vector_index).
--define(XDB_SEGMENT_INDEX, ets_xdb_segement_index).
--define(IP2REGION_CACHE, ets_ip2region_cache).
-
-
 -define(XDB_HEADER_SIZE, 256).
 -define(XDB_VECTOR_COLS, 256).
 -define(XDB_VECTOR_INDEX_SIZE, 8).
 -define(XDB_VECTOR_INDEX_COUNT, (16#10000)). %% 256*256
 
--define(XDB_SEGMENT_INDEX_SIZE, 14).
-
--define(IP2REGION_POOL, ip2region_pool).
-
 %% xdb header field offsets (bytes)
 -define(XDB_HEADER_VERSION_OFFSET, 0).
 -define(XDB_HEADER_INDEX_POLICY_OFFSET, 2).
@@ -26,7 +17,6 @@
 -define(XDB_HEADER_END_INDEX_PTR_OFFSET, 12).
 -define(XDB_HEADER_IP_VERSION_OFFSET, 16).
 -define(XDB_HEADER_RUNTIME_PTR_BYTES_OFFSET, 18).
--define(XDB_HEADER_INFO_LENGTH, 256).
 
 %% IP versions as stored in xdb header
 -define(IP_VERSION_4, 4).
@@ -45,12 +35,8 @@
 -define(IP2REGION_CACHE_V6, ets_ip2region_cache_v6).
 
 %% Per-version pool names. ?IP2REGION_POOL is kept as the legacy v4 pool name.
+-define(IP2REGION_POOL, ip2region_pool).
 -define(IP2REGION_POOL_V4, ip2region_pool_v4).
 -define(IP2REGION_POOL_V6, ip2region_pool_v6).
 
--ifndef(IF).
--define(IF(C, T, F), case (C) of true -> (T); false -> (F) end).
--define(IF(C, T), ?IF(C, T, skip)).
--endif.
-
 -endif.

binding/erlang/src/ip2region_sup.erl

@@ -41,10 +41,6 @@ init([]) ->
 %%
 create_table() ->
     Opts = [named_table, set, public, {read_concurrency, true}, {keypos, 1}],
-    %% Legacy tables (kept for backward compatibility)
-    ensure_table(?XDB_VECTOR_INDEX, Opts),
-    ensure_table(?XDB_SEGMENT_INDEX, Opts),
-    ensure_table(?IP2REGION_CACHE, Opts),
     %% Version-specific tables for dual-stack support
     ensure_table(?XDB_VECTOR_INDEX_V4, Opts),
     ensure_table(?XDB_VECTOR_INDEX_V6, Opts),
@@ -64,11 +60,14 @@ pool_child_specs() ->
     {ok, PoolArgsCfg} = application:get_env(poolargs),
     Versions = [Version || {Version, _File} <- DbConfig],
     UseLegacyName = (Versions == [ipv4]),
+    V4PoolName = case UseLegacyName of
+                     true -> ?IP2REGION_POOL;
+                     false -> ?IP2REGION_POOL_V4
+                 end,
+    ok = application:set_env(?APP_NAME, v4_pool_name, V4PoolName),
     lists:foldl(
-        fun({ipv4, File}, Acc) when UseLegacyName ->
+        fun({ipv4, File}, Acc) ->
-                [make_pool_spec(?IP2REGION_POOL, ipv4, File, PoolArgsCfg) | Acc];
+                [make_pool_spec(V4PoolName, ipv4, File, PoolArgsCfg) | Acc];
-           ({ipv4, File}, Acc) ->
-                [make_pool_spec(?IP2REGION_POOL_V4, ipv4, File, PoolArgsCfg) | Acc];
            ({ipv6, File}, Acc) ->
                 [make_pool_spec(?IP2REGION_POOL_V6, ipv6, File, PoolArgsCfg) | Acc];
            (_, Acc) ->

binding/erlang/src/ip2region_util.erl

@@ -13,7 +13,7 @@
 
 -spec ip_version(Ip :: tuple() | list() | binary() | integer()) ->
     ipv4 | ipv6 | {error, atom()}.
-ip_version(Ip) when is_integer(Ip) -> ipv4;
+ip_version(Ip) when is_integer(Ip), Ip >= 0, Ip =< 16#FFFFFFFF -> ipv4;
 ip_version({_, _, _, _}) -> ipv4;
 ip_version({_, _, _, _, _, _, _, _}) -> ipv6;
 ip_version(Ip) when is_binary(Ip) ->
@@ -29,7 +29,7 @@ ip_version(_) ->
 
 -spec ip_to_bytes(Ip :: tuple() | list() | binary() | integer()) ->
     {ok, ipv4 | ipv6, binary()} | {error, atom()}.
-ip_to_bytes(Ip) when is_integer(Ip) ->
+ip_to_bytes(Ip) when is_integer(Ip), Ip >= 0, Ip =< 16#FFFFFFFF ->
     {ok, ipv4, <<Ip:32>>};
 ip_to_bytes({A, B, C, D}) ->
     {ok, ipv4, <<A, B, C, D>>};
@@ -51,7 +51,7 @@ ip_to_bytes(_) ->
 
 -spec ipv4_to_n(Ip :: tuple() | list() | binary() | integer()) ->
     non_neg_integer() | {error, atom()}.
-ipv4_to_n(IntIp) when is_integer(IntIp) -> IntIp;
+ipv4_to_n(IntIp) when is_integer(IntIp), IntIp >= 0, IntIp =< 16#FFFFFFFF -> IntIp;
 ipv4_to_n({A, B, C, D}) ->
     <<N:32>> = <<A, B, C, D>>,
     N;

binding/erlang/src/ip2region_worker.erl

@@ -18,8 +18,7 @@
 
 -record(state, {
     xdb_fd,
-    version :: ipv4 | ipv6,
+    version :: ipv4 | ipv6
-    segment_index_size :: pos_integer()
 }).
 
 %%==========================================
@@ -63,14 +62,14 @@ init(Args) ->
     {ok, IoDevice} = file:open(XdbFileName, [read, binary]),
     {ok, HeaderBin} = file:read(IoDevice, ?XDB_HEADER_SIZE),
     {ok, Header} = ip2region_xdb:parse_header(HeaderBin),
-    Version = resolve_version(Header),
+    case resolve_version(Header) of
-    SegmentIndexSize = ip2region_xdb:segment_index_size(Version),
+        {ok, Version} ->
-    load_vector_index(IoDevice, Version),
+            load_vector_index(IoDevice, Version),
-    {ok, #state{
+            {ok, #state{xdb_fd = IoDevice, version = Version}};
-        xdb_fd = IoDevice,
+        {error, Reason} ->
-        version = Version,
+            file:close(IoDevice),
-        segment_index_size = SegmentIndexSize
+            {stop, Reason}
-    }}.
+    end.
 
 handle_call(Request, From, State) ->
     try
@@ -137,18 +136,18 @@ do_info(Info, State) ->
 
 resolve_version(Header) ->
     case ip2region_xdb:header_version(Header) of
-        2 -> ipv4;
+        2 -> {ok, ipv4};
         3 ->
             case ip2region_xdb:header_ip_version(Header) of
-                ?IP_VERSION_4 -> ipv4;
+                ?IP_VERSION_4 -> {ok, ipv4};
-                ?IP_VERSION_6 -> ipv6;
+                ?IP_VERSION_6 -> {ok, ipv6};
-                _ -> ipv4
+                V -> {error, {invalid_xdb_ip_version, V}}
             end;
-        _ -> ipv4
+        V -> {error, {invalid_xdb_version, V}}
     end.
 
 load_vector_index(IoDevice, Version) ->
-    Table = vector_index_table(Version),
+    Table = ip2region_xdb:vector_index_table(Version),
     case ets:info(Table, size) of
         undefined ->
             Opts = [named_table, set, public, {read_concurrency, true}, {keypos, 1}],
@@ -162,7 +161,7 @@ load_vector_index(IoDevice, Version) ->
 
 load_vector_index_data(IoDevice, Table) ->
     {ok, VectorIndexBin} =
-        file:read(IoDevice, ?XDB_VECTOR_INDEX_COUNT * 8),
+        file:read(IoDevice, ?XDB_VECTOR_INDEX_COUNT * ?XDB_VECTOR_INDEX_SIZE),
     load_vector_index_aux(VectorIndexBin, 0, Table).
 
 load_vector_index_aux(<<>>, _Index, _Table) -> ok;
@@ -170,24 +169,15 @@ load_vector_index_aux(<<SPtr:32/little, EPtr:32/little, VectorIndexBin/binary>>,
     ets:insert(Table, {Index, SPtr, EPtr}),
     load_vector_index_aux(VectorIndexBin, Index + 1, Table).
 
-search_ip(IoDevice, IpInt, State) when is_integer(IpInt) ->
+search_ip(IoDevice, Ip, #state{version = Version}) ->
-    search_ip(IoDevice, <<IpInt:32>>, State);
+    SegSize = ip2region_xdb:segment_index_size(Version),
-search_ip(IoDevice, Ip, #state{version = Version, segment_index_size = SegSize}) ->
+    VectorTable = ip2region_xdb:vector_index_table(Version),
-    CacheTable = cache_table(Version),
+    SegmentTable = ip2region_xdb:segment_index_table(Version),
-    VectorTable = vector_index_table(Version),
+    <<A:8, B:8, _/binary>> = Ip,
-    SegmentTable = segment_index_table(Version),
+    VectorIdx = A * ?XDB_VECTOR_COLS + B,
-    case ets:lookup(CacheTable, Ip) of
+    [{_, SPtr, EPtr}] = ets:lookup(VectorTable, VectorIdx),
-        [{_, RegionInfo}] ->
+    search_ip(IoDevice, Ip, SPtr, EPtr, 0,
-            RegionInfo;
+              (EPtr - SPtr) div SegSize, SegSize, Version, SegmentTable).
-        _ ->
-            <<A:8, B:8, _/binary>> = Ip,
-            VectorIdx = A * ?XDB_VECTOR_COLS + B,
-            [{_, SPtr, EPtr}] = ets:lookup(VectorTable, VectorIdx),
-            RegionInfo = search_ip(IoDevice, Ip, SPtr, EPtr, 0,
-                                   (EPtr - SPtr) div SegSize, SegSize, Version, SegmentTable),
-            ets:insert_new(CacheTable, {Ip, RegionInfo}),
-            RegionInfo
-    end.
 
 search_ip(IoDevice, Ip, SPtr, EPtr, Low, High, SegSize, Version, SegmentTable) when Low =< High ->
     Middle = (Low + High) bsr 1,
@@ -240,12 +230,3 @@ decode_segment_index(Bin, ?XDB_SEGMENT_INDEX_SIZE_V4) ->
 decode_segment_index(Bin, ?XDB_SEGMENT_INDEX_SIZE_V6) ->
     <<SIp:16/binary, EIp:16/binary, DataLen:16/little, DataPtr:32/little>> = Bin,
     {SIp, EIp, DataLen, DataPtr}.
-
-vector_index_table(ipv4) -> ?XDB_VECTOR_INDEX_V4;
-vector_index_table(ipv6) -> ?XDB_VECTOR_INDEX_V6.
-
-segment_index_table(ipv4) -> ?XDB_SEGMENT_INDEX_V4;
-segment_index_table(ipv6) -> ?XDB_SEGMENT_INDEX_V6.
-
-cache_table(ipv4) -> ?IP2REGION_CACHE_V4;
-cache_table(ipv6) -> ?IP2REGION_CACHE_V6.

binding/erlang/src/ip2region_xdb.erl

@@ -14,6 +14,7 @@
          header_start_index_ptr/1, header_end_index_ptr/1,
          header_ip_version/1, header_runtime_ptr_bytes/1]).
 -export([segment_index_size/1]).
+-export([vector_index_table/1, segment_index_table/1, cache_table/1]).
 
 -include("ip2region.hrl").
 
@@ -55,3 +56,12 @@ header_runtime_ptr_bytes(#xdb_header{runtime_ptr_bytes = B}) -> B.
 -spec segment_index_size(ipv4 | ipv6) -> pos_integer().
 segment_index_size(ipv4) -> ?XDB_SEGMENT_INDEX_SIZE_V4;
 segment_index_size(ipv6) -> ?XDB_SEGMENT_INDEX_SIZE_V6.
+
+vector_index_table(ipv4) -> ?XDB_VECTOR_INDEX_V4;
+vector_index_table(ipv6) -> ?XDB_VECTOR_INDEX_V6.
+
+segment_index_table(ipv4) -> ?XDB_SEGMENT_INDEX_V4;
+segment_index_table(ipv6) -> ?XDB_SEGMENT_INDEX_V6.
+
+cache_table(ipv4) -> ?IP2REGION_CACHE_V4;
+cache_table(ipv6) -> ?IP2REGION_CACHE_V6.

binding/erlang/src/xdb.erl

@@ -13,7 +13,7 @@
 -export([search/1]).
 
 -spec search(Ip :: tuple() | list() | binary() | integer()) ->
-    Result :: list() | {error, Reason::atom()}.
+    Result :: list() | {error, term()}.
 search(Ip) ->
     case ip2region_util:ip_to_bytes(Ip) of
         {ok, ipv4, IpBin} ->
@@ -25,23 +25,22 @@ search(Ip) ->
     end.
 
 do_search(PoolName, Version, IpBin) ->
-    CacheTable = cache_table(Version),
+    CacheTable = ip2region_xdb:cache_table(Version),
     case ets:lookup(CacheTable, IpBin) of
         [{_, Region}] -> Region;
         _ ->
             Worker = poolboy:checkout(PoolName, true, infinity),
             try
-                ip2region_worker:search(Worker, IpBin)
+                Region = ip2region_worker:search(Worker, IpBin),
+                ets:insert(CacheTable, {IpBin, Region}),
+                Region
             after
                 poolboy:checkin(PoolName, Worker)
             end
     end.
 
 v4_pool() ->
-    case whereis(?IP2REGION_POOL) of
+    case application:get_env(?APP_NAME, v4_pool_name) of
-        Pid when is_pid(Pid) -> ?IP2REGION_POOL;
+        {ok, PoolName} -> PoolName;
-        undefined -> ?IP2REGION_POOL_V4
+        undefined -> ?IP2REGION_POOL
     end.
-
-cache_table(ipv4) -> ?IP2REGION_CACHE_V4;
-cache_table(ipv6) -> ?IP2REGION_CACHE_V6.

binding/erlang/src/xdb_benchmark.erl

@@ -46,7 +46,7 @@ cpu_info() ->
         {unix, linux} ->
             Model = linux_cpu_field("model name"),
             Clock = format_clock_mhz(linux_cpu_field("cpu MHz")),
-            Cores = trim(os:cmd("grep -c '^processor' /proc/cpuinfo 2>/dev/null")),
+            Cores = string:trim(os:cmd("grep -c '^processor' /proc/cpuinfo 2>/dev/null")),
             Threads = Cores,
             {Model, Clock, Cores, Threads};
         _ ->
@@ -54,28 +54,28 @@ cpu_info() ->
     end.
 
 sysctl(Key) ->
-    trim(os:cmd("sysctl -n " ++ Key ++ " 2>/dev/null")).
+    string:trim(os:cmd("sysctl -n " ++ Key ++ " 2>/dev/null")).
 
 linux_cpu_field(Key) ->
     Cmd = "grep -m1 '^" ++ Key ++ "' /proc/cpuinfo 2>/dev/null | cut -d: -f2- | sed 's/^ *//'",
-    trim(os:cmd(Cmd)).
+    string:trim(os:cmd(Cmd)).
 
 first_non_empty(["" | Rest]) -> first_non_empty(Rest);
 first_non_empty([Val | _]) -> Val;
 first_non_empty([]) -> "".
 
 format_clock(HzStr) ->
-    case string:to_integer(trim(HzStr)) of
+    case string:to_integer(string:trim(HzStr)) of
         {ok, Hz, _} when Hz > 1000000000 ->
             lists:flatten(io_lib:format("~.2f GHz", [Hz / 1000000000]));
         {ok, Hz, _} when Hz > 1000000 ->
-            lists:flatten(io_lib:format("~.2f GHz", [Hz / 1000000000]));
+            lists:flatten(io_lib:format("~.2f MHz", [Hz / 1000000]));
         _ ->
             ""
     end.
 
 format_clock_mhz(MhzStr) ->
-    case string:to_float(trim(MhzStr)) of
+    case string:to_float(string:trim(MhzStr)) of
         {ok, Mhz, _} ->
             lists:flatten(io_lib:format("~.3f GHz", [Mhz / 1000]));
         _ ->
@@ -108,8 +108,8 @@ load_test_data(Fd, IpList) ->
 run(IpList) ->
     garbage_collect(),
     io:format("~nBenchmarks:~n", []),
-    run_test("file", IpList),
+    run_test("cold", IpList),
-    run_test("cache", IpList),
+    run_test("warm", IpList),
     io:format("~nDone.~n", []).
 
 run_test(Label, IpList) ->
@@ -128,5 +128,3 @@ run_test_aux([]) -> ok;
 run_test_aux([Ip | Tail]) ->
     xdb:search(Ip),
     run_test_aux(Tail).
-
-trim(Str) -> string:trim(Str).

binding/erlang/test/ip2region_sup_test.erl

@@ -6,8 +6,11 @@ pools_started_test() ->
     application:stop(ip2region),
     application:unload(ip2region),
     ok = application:load(ip2region),
-    {ok, Cwd} = file:get_cwd(),
+    TestDir = filename:dirname(?FILE),
-    V6File = filename:join([Cwd, "..", "..", "data", "ip2region_v6.xdb"]),
+    ErlangDir = filename:dirname(TestDir),
+    BindingDir = filename:dirname(ErlangDir),
+    RepoRoot = filename:dirname(BindingDir),
+    V6File = filename:join([RepoRoot, "data", "ip2region_v6.xdb"]),
     ok = application:set_env(ip2region, db, [
         {ipv4, "ip2region.xdb"},
         {ipv6, V6File}

binding/erlang/test/ip2region_util_test.erl

@@ -30,7 +30,11 @@ invalid_inputs_test_() ->
         ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes("xxx.0.8.0")),
         ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes("::ggg")),
         ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes({1,2,3})),
-        ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes(atom))
+        ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes(atom)),
+        ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes(-1)),
+        ?_assertEqual({error, bad_ip_format}, ip2region_util:ip_to_bytes(16#100000000)),
+        ?_assertEqual({error, bad_ip_format}, ip2region_util:ipv4_to_n(-1)),
+        ?_assertEqual({error, bad_ip_format}, ip2region_util:ipv4_to_n(16#100000000))
     ].
 
 legacy_ipv4_to_n_test_() ->