Protobuf: fix bugs that parsing complex syntax .proto files

Some .proto files contain complex syntax that does not be described in protobuf official site
(https://developers.google.com/protocol-buffers/docs/reference/proto3-spec).

1. Update 'epan/protobuf_lang_parser.lemon' to:
1) Support complex option names format (EBNF):
    optionName = ( ident | "(" fullIdent ")" ) { "." ( ident | "(" fullIdent ")" ) }
for example, "option (complex_opt2).(grault) = 654;".
2) Make enum body support 'reserved' section (EBNF):
    enumBody = "{" { reserved | option | enumField | emptyStatement } "}"
3) Allow the value of field or enumValue option to be "{ ... }" other than constant:
    enumValueOption = optionName  "=" ( constant | customOptionValue ) ";"
    fieldOption = optionName  "=" ( constant | customOptionValue ) ";"
4) Allow 'group' section missing 'label' (for example, in 'oneof' section).
5) Make 'oneof' section support 'option' and 'group' sections (BNF):
    oneof = "oneof" oneofName "{" { oneofField | option | group | emptyStatement } "}"
6) Ignore unused 'extend' section.
7) Fix the bug of one string being splitted into multi-lines.

2. Update 'epan/protobuf_lang_tree.c' to:
8) Fix the bug of parsing repeated option.

3. Update 'test/suite_dissection.py' to add test case for parsing complex syntax .proto files:
  test/protobuf_lang_files/complex_proto_files/unittest_custom_options.proto
  test/protobuf_lang_files/complex_proto_files/complex_syntax.proto
and dependency files:
  test/protobuf_lang_files/well_know_types/google/protobuf/any.proto
  test/protobuf_lang_files/well_know_types/google/protobuf/descriptor.proto

Refer to issue #17046

7 files changed, 865 insertions, 10 deletions

diff --git a/epan/protobuf_lang_parser.lemon b/epan/protobuf_lang_parser.lemon
index 36f9edb506..4aa1b9b85d 100644
--- a/epan/protobuf_lang_parser.lemon
+++ b/epan/protobuf_lang_parser.lemon
@@ -190,11 +190,17 @@ option ::= PT_OPTION optionName PT_ASSIGN constant PT_SEMICOLON.
 option ::= PT_OPTION optionName PT_ASSIGN customOptionValue PT_SEMICOLON.
 
 /* v2/v3: optionName = ( ident | "(" fullIdent ")" ) { "." ident } */
-optionName ::= exIdent.
-optionName(A) ::= PT_LPAREN exIdent(B) PT_RPAREN.
+/* Offical PBL bugfix: optionName = ( ident | "(" fullIdent ")" ) { "." ( ident | "(" fullIdent ")" ) } */
+extIdentInParentheses(A) ::= PT_LPAREN exIdent(B) PT_RPAREN.
     { A = B; A->v = pbl_store_string_token(state, g_strconcat("(", B->v, ")", NULL)); }
-optionName(A) ::= PT_LPAREN exIdent(B) PT_RPAREN exIdent(C).  /* Note that the exIdent contains "." */
-    { A = B; A->v = pbl_store_string_token(state, g_strconcat("(", B->v, ")", C->v, NULL)); }
+optionName ::= exIdent.
+optionName ::= extIdentInParentheses.
+optionName(A) ::= optionName(B) exIdent(C). // Note that the exIdent contains "."
+    { A = B; A->v = pbl_store_string_token(state, g_strconcat(B->v, C->v, NULL)); }
+optionName(A) ::= optionName(B) PT_DOT extIdentInParentheses(C).
+    { A = B; A->v = pbl_store_string_token(state, g_strconcat(B->v, ".", C->v, NULL)); }
+optionName(A) ::= optionName(B) extIdentInParentheses(C).
+    { A = B; A->v = pbl_store_string_token(state, g_strconcat(B->v, ".", C->v, NULL)); }
 
 /* Allow format which not defined in offical PBL specification like:
     option (google.api.http) = { post: "/v3alpha/kv/put" body: "*" };
@@ -203,9 +209,30 @@ optionName(A) ::= PT_LPAREN exIdent(B) PT_RPAREN exIdent(C).  /* Note that the e
 */
 customOptionValue ::= PT_LCURLY customOptionBody PT_RCURLY.
 
+/* The formal EBNF of customOptionBody seems to be */
+/*
+customOptionBody ::= .
+customOptionBody ::= customOptionBody optionField.
+customOptionBody ::= customOptionBody PT_COMMA optionField.
+customOptionBody ::= customOptionBody PT_SEMICOLON optionField.
+
+optionField ::= optionName PT_COLON constant.
+optionField ::= optionName PT_COLON customOptionValue.
+optionField ::= optionName customOptionValue.
+optionField ::= optionName PT_COLON array.
+
+array ::= PT_LBRACKET arrayBody PT_RBRACKET.
+arrayBodyConst ::= constant.
+arrayBodyConst ::= arrayBody PT_COMMA constant.
+arrayBodyCustom ::= customOptionValue.
+arrayBodyCustom ::= arrayBody PT_COMMA customOptionValue.
+arrayBody ::= arrayBodyConst.
+arrayBody ::= arrayBodyCustom.
+*/
+/* but for handling unexpected situations, we still use following EBNF  */
 customOptionBody ::= .
 customOptionBody ::= customOptionBody exIdent.
-customOptionBody ::= customOptionBody strLit.
+customOptionBody ::= customOptionBody PT_STRLIT.
 customOptionBody ::= customOptionBody symbolsWithoutCurly.
 customOptionBody ::= customOptionBody intLit.
 customOptionBody ::= customOptionBody customOptionValue.
@@ -275,7 +302,9 @@ enum(A) ::= PT_ENUM enumName(B) PT_LCURLY enumBody(C) PT_RCURLY.
     { A = C; pbl_set_node_name(A, B->ln, B->v); }
 
 /* v2/v3: enumBody = "{" { option | enumField | emptyStatement } "}" */
+/* Offical PBL bugfix: enumBody = "{" { reserved | option | enumField | emptyStatement } "}" */
 enumBody(A) ::= .  { A = pbl_create_node(state->file, CUR_LINENO, PBL_ENUM, NAME_TO_BE_SET); }
+enumBody ::= enumBody reserved.
 enumBody ::= enumBody option.
 enumBody(A) ::= enumBody(B) enumField(C).  { A = B; pbl_add_child(A, C); }
 enumBody ::= enumBody emptyStatement.
@@ -296,7 +325,9 @@ enumValueOptions ::= enumValueOption.
 enumValueOptions ::= enumValueOptions PT_COMMA enumValueOption.
 
 /* v2/v3: enumValueOption = optionName "=" constant */
+/* Offical PBL bugfix: enumValueOption = optionName  "=" ( constant | customOptionValue ) ";" */
 enumValueOption ::= optionName PT_ASSIGN constant.
+enumValueOption ::= optionName PT_ASSIGN customOptionValue.
 
 /* v2: service = "service" serviceName "{" { option | rpc | stream | emptyStatement } "}" */
 /* v3: service = "service" serviceName "{" { option | rpc | emptyStatement } "}" */
@@ -373,21 +404,30 @@ fieldOptions(A) ::= fieldOptions(B) PT_COMMA fieldOption(C).
     { A = B; pbl_add_child(A, C); }
 
 /* v2/v3: fieldOption = optionName "=" constant */
+/* Offical PBL bugfix: fieldOption = optionName  "=" ( constant | customOptionValue ) ";" */
 fieldOption(A) ::= optionName(B) PT_ASSIGN constant(C).
     { A = pbl_create_option_node(state->file, B->ln, B->v, C); }
+fieldOption(A) ::= optionName(B) PT_ASSIGN customOptionValue.
+    { A = pbl_create_option_node(state->file, B->ln, B->v, pbl_store_string_token(state, g_strdup("{ ... }"))); }
 
 /* v2 only: group = label "group" groupName "=" fieldNumber messageBody */
+/* Offical PBL bugfix: there is no label if the 'group' is a member of oneof body */
+group(A) ::= PT_GROUP groupName(B) PT_ASSIGN fieldNumber PT_LCURLY messageBody(C) PT_RCURLY.
+    { A = C; pbl_set_node_name(A, B->ln, B->v); }
 group(A) ::= label PT_GROUP groupName(B) PT_ASSIGN fieldNumber PT_LCURLY messageBody(C) PT_RCURLY.
     { A = C; pbl_set_node_name(A, B->ln, B->v); }
 
 groupName ::= exIdent.
 
 /* v2/v3: oneof = "oneof" oneofName "{" { oneofField | emptyStatement } "}" */
+/* Offical PBL bugfix: oneof = "oneof" oneofName "{" { oneofField | option | group | emptyStatement } "}" */
 oneof(A) ::= PT_ONEOF oneofName(B) PT_LCURLY oneofBody(C) PT_RCURLY.
     { A = C; pbl_set_node_name(A, B->ln, B->v); }
 
 oneofBody(A) ::= .  { A = pbl_create_node(state->file, CUR_LINENO, PBL_ONEOF, NAME_TO_BE_SET); }
 oneofBody(A) ::= oneofBody(B) oneofField(C). { A = B; pbl_add_child(A, C); }
+oneofBody ::= oneofBody option.
+oneofBody ::= oneofBody group.
 oneofBody ::= oneofBody emptyStatement.
 
 /* v2/v3: oneofField = type fieldName "=" fieldNumber [ "[" fieldOptions "]" ] ";" */
@@ -436,11 +476,13 @@ Note that there is an error in BNF definition about reserved fieldName. It's str
 quoteFieldNames ::= strLit.
 quoteFieldNames ::= quoteFieldNames PT_COMMA strLit.
 
-/* v2 only: extend = "extend" messageType "{" {field | group | emptyStatement} "}" */
-extend(A) ::= PT_EXTEND(X) messageType(B) PT_LCURLY extendBody(C) PT_RCURLY.
-    { A = C; pbl_set_node_name(A, X->ln, pbl_store_string_token(state, g_strconcat(B, "Extend", NULL))); }
+/* v2/v3: extend = "extend" messageType "{" {field | group | emptyStatement} "}"
+Note that creating custom options uses extensions, which are permitted only for custom options in proto3.
+We don't use custom options while parsing packet, so we just ignore the 'extend'.
+*/
+extend(A) ::= PT_EXTEND messageType PT_LCURLY extendBody(B) PT_RCURLY.
+    { A = NULL; pbl_free_node(B); }
 
-/* v2 only */
 extendBody(A) ::= .  { A = pbl_create_node(state->file, CUR_LINENO, PBL_MESSAGE, NAME_TO_BE_SET); }
 extendBody(A) ::= extendBody(B) field(C). { A = B; pbl_add_child(A, C); }
 extendBody(A) ::= extendBody(B) group(C). { A = B; pbl_add_child(A, C); }
@@ -481,6 +523,8 @@ constant(A) ::= PT_MINUS exIdent(B). { A = pbl_store_string_token(state, g_strco
 exIdent ::= PT_IDENT.
 
 strLit(A) ::= PT_STRLIT(B).  { A = pbl_store_string_token(state, g_strndup(B->v + 1, strlen(B->v) - 2)); }
+/* support one string being splitted into multi-lines */
+strLit(A) ::= strLit(B) PT_STRLIT(C). { A = pbl_store_string_token(state, g_strconcat(B, g_strndup(C->v + 1, strlen(C->v) - 2), NULL)); }
 
 %code {
 
diff --git a/epan/protobuf_lang_tree.c b/epan/protobuf_lang_tree.c
index 25c9d2e1ac..b269ad3439 100644
--- a/epan/protobuf_lang_tree.c
+++ b/epan/protobuf_lang_tree.c
@@ -1009,7 +1009,17 @@ pbl_add_child(pbl_node_t* parent, pbl_node_t* child)
     }
 
     node = (pbl_node_t*) g_hash_table_lookup(parent->children_by_name, child->name);
-    if (node && child->file && parent->file
+    if (node && node->nodetype == PBL_OPTION && child->nodetype == PBL_OPTION
+        && ((pbl_option_descriptor_t*)node)->value && ((pbl_option_descriptor_t*)child)->value) {
+        /* repeated option can be set many times like:
+             string fieldWithComplexOption5 = 5 [(rules).repeated_int = 1, (rules).repeated_int = 2];
+           we just merge the old value and new value in format /old_value "," new_value/.
+        */
+        gchar* oval = ((pbl_option_descriptor_t*)node)->value;
+        gchar* nval = ((pbl_option_descriptor_t*)child)->value;
+        ((pbl_option_descriptor_t*)child)->value = g_strconcat(oval, ",", nval, NULL);
+        g_free(nval);
+    } else if (node && child->file && parent->file
         && child->file->pool && child->file->pool->error_cb) {
         child->file->pool->error_cb(
             "Protobuf: Warning: \"%s\" of [%s:%d] is already defined in file [%s:%d].\n",
diff --git a/test/protobuf_lang_files/complex_proto_files/complex_syntax.proto b/test/protobuf_lang_files/complex_proto_files/complex_syntax.proto
new file mode 100644
index 0000000000..359ead4aa6
--- /dev/null
+++ b/test/protobuf_lang_files/complex_proto_files/complex_syntax.proto
@@ -0,0 +1,82 @@
+// Test more complex syntax of *.proto files.
+syntax = "proto3";
+package wireshark.protobuf.test.complex.syntax;
+import "google/protobuf/descriptor.proto";
+
+// equal to "testing.multiline.strings"
+option java_package = "testing."
+                      'multiline.'
+"strings";
+
+// user defined options for messages
+extend google.protobuf.MessageOptions {
+  bool disabled = 1071;
+  bool ignored = 1072;
+  TestMultiLinesOption mlinemsg = 1073;
+}
+
+// user defined options for oneof types
+extend google.protobuf.OneofOptions {
+  bool required = 1071;
+}
+
+// user defined options for fields
+extend google.protobuf.FieldOptions {
+  FieldRules rules = 1071;
+}
+// test extend google.protobuf.FieldOptions twice
+extend google.protobuf.FieldOptions {
+  string multilines = 1072;
+}
+
+message FieldRules {
+  oneof type {
+    BoolRules     bool     = 13;
+    StringRules   string   = 14;
+  }
+  repeated uint32 repeated_uint = 15;
+}
+
+message StringRules {
+  uint64 min_bytes = 2;
+  BoolRules morebool = 3;
+  string astr = 4;
+}
+
+message BoolRules {
+  bool const = 1;
+  repeated bool repeated_bool = 2;
+  repeated int32 repeated_int = 3;
+}
+
+message TestMultiLinesOption {
+  string mlines = 1;
+}
+
+message ComplexDefinedMessage {
+  option (mlinemsg).mlines = "first line"
+  "second line";
+
+  // test complex field options
+  string fieldWithComplexOption1 = 1 [(wireshark.protobuf.test.complex.syntax.rules).string = {min_bytes: 1}];
+  string fieldWithComplexOption2 = 2 [(rules).string = {min_bytes: 2 astr: "abc" }];
+  string fieldWithComplexOption3 = 3 [(rules).string.morebool = {const: true, repeated_bool: [false, true], repeated_int: [1, 2]}];
+  string fieldWithComplexOption4 = 4 [(rules).string = {min_bytes: 1; morebool { const: true }}];
+  string fieldWithComplexOption5 = 5 [(rules).repeated_uint = 1, (rules).repeated_uint = 2];
+
+  // test oneof custom option
+  oneof oneofWithOption {
+    option (wireshark.protobuf.test.complex.syntax.required) = true;
+    int32 field1 = 11;
+    string field2 = 12;
+  }
+
+  // test multilines strings
+  uint32 fieldWithMultilineStringOption = 20 [(wireshark.protobuf.test.complex.syntax.multilines) = "first line"
+  'Second line' ];
+}
+
+// add this message for testing whether this file was successfully parsed
+message TestFileParsed {
+  optional int32 last_field_for_wireshark_test = 1;
+}
diff --git a/test/protobuf_lang_files/complex_proto_files/unittest_custom_options.proto b/test/protobuf_lang_files/complex_proto_files/unittest_custom_options.proto
new file mode 100644
index 0000000000..d04c9b7289
--- /dev/null
+++ b/test/protobuf_lang_files/complex_proto_files/unittest_custom_options.proto
@@ -0,0 +1,395 @@
+// This file is from https://github.com/protocolbuffers/protobuf/blob/3.14.x/src/google/protobuf/unittest_custom_options.proto
+// To reduce the file size, some comments have been removed.
+// Message 'TestFileParsed' is added at the end of file for testing whether this file was successfully parsed.
+
+syntax = "proto2";
+
+option cc_generic_services = true;
+option java_generic_services = true;
+option py_generic_services = true;
+
+option (file_opt1) = 9876543210;
+
+import "google/protobuf/any.proto";
+import "google/protobuf/descriptor.proto";
+
+package protobuf_unittest;
+
+// Some simple test custom options of various types.
+extend google.protobuf.FileOptions {
+  optional uint64 file_opt1 = 7736974;
+}
+
+extend google.protobuf.MessageOptions {
+  optional int32 message_opt1 = 7739036;
+}
+
+extend google.protobuf.FieldOptions {
+  optional fixed64 field_opt1 = 7740936;
+  optional int32 field_opt2 = 7753913 [default = 42];
+}
+
+extend google.protobuf.OneofOptions {
+  optional int32 oneof_opt1 = 7740111;
+}
+
+extend google.protobuf.EnumOptions {
+  optional sfixed32 enum_opt1 = 7753576;
+}
+
+extend google.protobuf.EnumValueOptions {
+  optional int32 enum_value_opt1 = 1560678;
+}
+
+extend google.protobuf.ServiceOptions {
+  optional sint64 service_opt1 = 7887650;
+}
+
+enum MethodOpt1 {
+  METHODOPT1_VAL1 = 1;
+  METHODOPT1_VAL2 = 2;
+}
+
+extend google.protobuf.MethodOptions {
+  optional MethodOpt1 method_opt1 = 7890860;
+}
+
+message TestMessageWithCustomOptions {
+  option message_set_wire_format = false;
+  option (message_opt1) = -56;
+
+  optional string field1 = 1 [ctype = CORD, (field_opt1) = 8765432109];
+
+  oneof AnOneof {
+    option (oneof_opt1) = -99;
+
+    int32 oneof_field = 2;
+  }
+
+  enum AnEnum {
+    option (enum_opt1) = -789;
+
+    ANENUM_VAL1 = 1;
+    ANENUM_VAL2 = 2 [(enum_value_opt1) = 123];
+  }
+}
+
+message CustomOptionFooRequest {}
+
+message CustomOptionFooResponse {}
+
+message CustomOptionFooClientMessage {}
+
+message CustomOptionFooServerMessage {}
+
+service TestServiceWithCustomOptions {
+  option (service_opt1) = -9876543210;
+
+  rpc Foo(CustomOptionFooRequest) returns (CustomOptionFooResponse) {
+    option (method_opt1) = METHODOPT1_VAL2;
+  }
+}
+
+message DummyMessageContainingEnum {
+  enum TestEnumType {
+    TEST_OPTION_ENUM_TYPE1 = 22;
+    TEST_OPTION_ENUM_TYPE2 = -23;
+  }
+}
+
+message DummyMessageInvalidAsOptionType {}
+
+extend google.protobuf.MessageOptions {
+  optional bool bool_opt = 7706090;
+  optional int32 int32_opt = 7705709;
+  optional int64 int64_opt = 7705542;
+  optional uint32 uint32_opt = 7704880;
+  optional uint64 uint64_opt = 7702367;
+  optional sint32 sint32_opt = 7701568;
+  optional sint64 sint64_opt = 7700863;
+  optional fixed32 fixed32_opt = 7700307;
+  optional fixed64 fixed64_opt = 7700194;
+  optional sfixed32 sfixed32_opt = 7698645;
+  optional sfixed64 sfixed64_opt = 7685475;
+  optional float float_opt = 7675390;
+  optional double double_opt = 7673293;
+  optional string string_opt = 7673285;
+  optional bytes bytes_opt = 7673238;
+  optional DummyMessageContainingEnum.TestEnumType enum_opt = 7673233;
+  optional DummyMessageInvalidAsOptionType message_type_opt = 7665967;
+}
+
+message CustomOptionMinIntegerValues {
+  option (bool_opt) = false;
+  option (int32_opt) = -0x80000000;
+  option (int64_opt) = -0x8000000000000000;
+  option (uint32_opt) = 0;
+  option (uint64_opt) = 0;
+  option (sint32_opt) = -0x80000000;
+  option (sint64_opt) = -0x8000000000000000;
+  option (fixed32_opt) = 0;
+  option (fixed64_opt) = 0;
+  option (sfixed32_opt) = -0x80000000;
+  option (sfixed64_opt) = -0x8000000000000000;
+}
+
+message CustomOptionMaxIntegerValues {
+  option (bool_opt) = true;
+  option (int32_opt) = 0x7FFFFFFF;
+  option (int64_opt) = 0x7FFFFFFFFFFFFFFF;
+  option (uint32_opt) = 0xFFFFFFFF;
+  option (uint64_opt) = 0xFFFFFFFFFFFFFFFF;
+  option (sint32_opt) = 0x7FFFFFFF;
+  option (sint64_opt) = 0x7FFFFFFFFFFFFFFF;
+  option (fixed32_opt) = 0xFFFFFFFF;
+  option (fixed64_opt) = 0xFFFFFFFFFFFFFFFF;
+  option (sfixed32_opt) = 0x7FFFFFFF;
+  option (sfixed64_opt) = 0x7FFFFFFFFFFFFFFF;
+}
+
+message CustomOptionOtherValues {
+  option (int32_opt) = -100;  // To test sign-extension.
+  option (float_opt) = 12.3456789;
+  option (double_opt) = 1.234567890123456789;
+  option (string_opt) = "Hello, \"World\"";
+  option (bytes_opt) = "Hello\0World";
+  option (enum_opt) = TEST_OPTION_ENUM_TYPE2;
+}
+
+message SettingRealsFromPositiveInts {
+  option (float_opt) = 12;
+  option (double_opt) = 154;
+}
+
+message SettingRealsFromNegativeInts {
+  option (float_opt) = -12;
+  option (double_opt) = -154;
+}
+
+message ComplexOptionType1 {
+  optional int32 foo = 1;
+  optional int32 foo2 = 2;
+  optional int32 foo3 = 3;
+  repeated int32 foo4 = 4;
+
+  extensions 100 to max;
+}
+
+message ComplexOptionType2 {
+  optional ComplexOptionType1 bar = 1;
+  optional int32 baz = 2;
+
+  message ComplexOptionType4 {
+    optional int32 waldo = 1;
+
+    extend google.protobuf.MessageOptions {
+      optional ComplexOptionType4 complex_opt4 = 7633546;
+    }
+  }
+
+  optional ComplexOptionType4 fred = 3;
+  repeated ComplexOptionType4 barney = 4;
+
+  extensions 100 to max;
+}
+
+message ComplexOptionType3 {
+  optional int32 qux = 1;
+
+  optional group ComplexOptionType5 = 2 {
+    optional int32 plugh = 3;
+  }
+}
+
+extend ComplexOptionType1 {
+  optional int32 quux = 7663707;
+  optional ComplexOptionType3 corge = 7663442;
+}
+
+extend ComplexOptionType2 {
+  optional int32 grault = 7650927;
+  optional ComplexOptionType1 garply = 7649992;
+}
+
+extend google.protobuf.MessageOptions {
+  optional protobuf_unittest.ComplexOptionType1 complex_opt1 = 7646756;
+  optional ComplexOptionType2 complex_opt2 = 7636949;
+  optional ComplexOptionType3 complex_opt3 = 7636463;
+  optional group ComplexOpt6 = 7595468 {
+    optional int32 xyzzy = 7593951;
+  }
+}
+
+message VariousComplexOptions {
+  option (.protobuf_unittest.complex_opt1).foo = 42;
+  option (protobuf_unittest.complex_opt1).(.protobuf_unittest.quux) = 324;
+  option (.protobuf_unittest.complex_opt1).(protobuf_unittest.corge).qux = 876;
+  option (protobuf_unittest.complex_opt1).foo4 = 99;
+  option (protobuf_unittest.complex_opt1).foo4 = 88;
+  option (complex_opt2).baz = 987;
+  option (complex_opt2).(grault) = 654;
+  option (complex_opt2).bar.foo = 743;
+  option (complex_opt2).bar.(quux) = 1999;
+  option (complex_opt2).bar.(protobuf_unittest.corge).qux = 2008;
+  option (complex_opt2).(garply).foo = 741;
+  option (complex_opt2).(garply).(.protobuf_unittest.quux) = 1998;
+  option (complex_opt2).(protobuf_unittest.garply).(corge).qux = 2121;
+  option (ComplexOptionType2.ComplexOptionType4.complex_opt4).waldo = 1971;
+  option (complex_opt2).fred.waldo = 321;
+  option (complex_opt2).barney = {
+    waldo: 101
+  };
+  option (complex_opt2).barney = {
+    waldo: 212
+  };
+  option (protobuf_unittest.complex_opt3).qux = 9;
+  option (complex_opt3).complexoptiontype5.plugh = 22;
+  option (complexopt6).xyzzy = 24;
+}
+
+message AggregateMessageSet {
+  option message_set_wire_format = true;
+
+  extensions 4 to max;
+}
+
+message AggregateMessageSetElement {
+  extend AggregateMessageSet {
+    optional AggregateMessageSetElement message_set_extension = 15447542;
+  }
+  optional string s = 1;
+}
+
+message Aggregate {
+  optional int32 i = 1;
+  optional string s = 2;
+
+  optional Aggregate sub = 3;
+
+  optional google.protobuf.FileOptions file = 4;
+  extend google.protobuf.FileOptions {
+    optional Aggregate nested = 15476903;
+  }
+
+  optional AggregateMessageSet mset = 5;
+
+  optional google.protobuf.Any any = 6;
+}
+
+extend google.protobuf.FileOptions {
+  optional Aggregate fileopt = 15478479;
+}
+extend google.protobuf.MessageOptions {
+  optional Aggregate msgopt = 15480088;
+}
+extend google.protobuf.FieldOptions {
+  optional Aggregate fieldopt = 15481374;
+}
+extend google.protobuf.EnumOptions {
+  optional Aggregate enumopt = 15483218;
+}
+extend google.protobuf.EnumValueOptions {
+  optional Aggregate enumvalopt = 15486921;
+}
+extend google.protobuf.ServiceOptions {
+  optional Aggregate serviceopt = 15497145;
+}
+extend google.protobuf.MethodOptions {
+  optional Aggregate methodopt = 15512713;
+}
+
+option (fileopt) = {
+  s: 'FileAnnotation'
+  i: 100
+
+  sub { s: 'NestedFileAnnotation' }
+
+  file {
+    [protobuf_unittest.fileopt] { s: 'FileExtensionAnnotation' }
+  }
+
+  mset {
+    [protobuf_unittest.AggregateMessageSetElement.message_set_extension] {
+      s: 'EmbeddedMessageSetElement'
+    }
+  }
+
+  any {
+    [type.googleapis.com/protobuf_unittest.AggregateMessageSetElement] {
+      s: 'EmbeddedMessageSetElement'
+    }
+  }
+};
+
+message AggregateMessage {
+  option (msgopt) = {
+    i: 101
+    s: 'MessageAnnotation'
+  };
+
+  optional int32 fieldname = 1 [(fieldopt) = { s: 'FieldAnnotation' }];
+}
+
+service AggregateService {
+  option (serviceopt) = {
+    s: 'ServiceAnnotation'
+  };
+
+  rpc Method(AggregateMessage) returns (AggregateMessage) {
+    option (methodopt) = {
+      s: 'MethodAnnotation'
+    };
+  }
+}
+
+enum AggregateEnum {
+  option (enumopt) = {
+    s: 'EnumAnnotation'
+  };
+
+  VALUE = 1 [(enumvalopt) = { s: 'EnumValueAnnotation' }];
+}
+
+message NestedOptionType {
+  message NestedMessage {
+    option (message_opt1) = 1001;
+
+    optional int32 nested_field = 1 [(field_opt1) = 1002];
+  }
+  enum NestedEnum {
+    option (enum_opt1) = 1003;
+
+    NESTED_ENUM_VALUE = 1 [(enum_value_opt1) = 1004];
+  }
+  extend google.protobuf.FileOptions {
+    optional int32 nested_extension = 7912573 [(field_opt2) = 1005];
+  }
+}
+
+message OldOptionType {
+  enum TestEnum { OLD_VALUE = 0; }
+  required TestEnum value = 1;
+}
+
+message NewOptionType {
+  enum TestEnum {
+    OLD_VALUE = 0;
+    NEW_VALUE = 1;
+  }
+  required TestEnum value = 1;
+}
+
+extend google.protobuf.MessageOptions {
+  optional OldOptionType required_enum_opt = 106161807;
+}
+
+message TestMessageWithRequiredEnumOption {
+  option (required_enum_opt) = {
+    value: OLD_VALUE
+  };
+}
+
+// add this message for testing whether this file was successfully parsed
+message TestFileParsed {
+  optional int32 last_field_for_wireshark_test = 1;
+}
diff --git a/test/protobuf_lang_files/well_know_types/google/protobuf/any.proto b/test/protobuf_lang_files/well_know_types/google/protobuf/any.proto
new file mode 100644
index 0000000000..8ff870dd82
--- /dev/null
+++ b/test/protobuf_lang_files/well_know_types/google/protobuf/any.proto
@@ -0,0 +1,22 @@
+// This file is from https://github.com/protocolbuffers/protobuf/blob/3.14.x/src/google/protobuf/any.proto
+// To reduce the file size, some comments have been removed.
+
+syntax = "proto3";
+
+package google.protobuf;
+
+option csharp_namespace = "Google.Protobuf.WellKnownTypes";
+option go_package = "google.golang.org/protobuf/types/known/anypb";
+option java_package = "com.google.protobuf";
+option java_outer_classname = "AnyProto";
+option java_multiple_files = true;
+option objc_class_prefix = "GPB";
+
+message Any {
+  // A URL/resource name that uniquely identifies the type of the serialized
+  // protocol buffer message.
+  string type_url = 1;
+
+  // Must be a valid serialized protocol buffer of the above specified type.
+  bytes value = 2;
+}
diff --git a/test/protobuf_lang_files/well_know_types/google/protobuf/descriptor.proto b/test/protobuf_lang_files/well_know_types/google/protobuf/descriptor.proto
new file mode 100644
index 0000000000..ece32e1571
--- /dev/null
+++ b/test/protobuf_lang_files/well_know_types/google/protobuf/descriptor.proto
@@ -0,0 +1,282 @@
+// This file is from https://github.com/protocolbuffers/protobuf/blob/3.14.x/src/google/protobuf/descriptor.proto
+// To reduce the file size, some comments have been removed.
+
+syntax = "proto2";
+
+package google.protobuf;
+
+option go_package = "google.golang.org/protobuf/types/descriptorpb";
+option java_package = "com.google.protobuf";
+option java_outer_classname = "DescriptorProtos";
+option csharp_namespace = "Google.Protobuf.Reflection";
+option objc_class_prefix = "GPB";
+option cc_enable_arenas = true;
+
+option optimize_for = SPEED;
+
+message FileDescriptorSet {
+  repeated FileDescriptorProto file = 1;
+}
+
+message FileDescriptorProto {
+  optional string name = 1;
+  optional string package = 2;
+  repeated string dependency = 3;
+  repeated int32 public_dependency = 10;
+  repeated int32 weak_dependency = 11;
+  repeated DescriptorProto message_type = 4;
+  repeated EnumDescriptorProto enum_type = 5;
+  repeated ServiceDescriptorProto service = 6;
+  repeated FieldDescriptorProto extension = 7;
+  optional FileOptions options = 8;
+  optional SourceCodeInfo source_code_info = 9;
+  optional string syntax = 12;
+}
+
+message DescriptorProto {
+  optional string name = 1;
+  repeated FieldDescriptorProto field = 2;
+  repeated FieldDescriptorProto extension = 6;
+  repeated DescriptorProto nested_type = 3;
+  repeated EnumDescriptorProto enum_type = 4;
+
+  message ExtensionRange {
+    optional int32 start = 1;
+    optional int32 end = 2;
+    optional ExtensionRangeOptions options = 3;
+  }
+  repeated ExtensionRange extension_range = 5;
+  repeated OneofDescriptorProto oneof_decl = 8;
+  optional MessageOptions options = 7;
+
+  message ReservedRange {
+    optional int32 start = 1;
+    optional int32 end = 2;
+  }
+  repeated ReservedRange reserved_range = 9;
+  repeated string reserved_name = 10;
+}
+
+message ExtensionRangeOptions {
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message FieldDescriptorProto {
+  enum Type {
+    TYPE_DOUBLE = 1;
+    TYPE_FLOAT = 2;
+    TYPE_INT64 = 3;
+    TYPE_UINT64 = 4;
+    TYPE_INT32 = 5;
+    TYPE_FIXED64 = 6;
+    TYPE_FIXED32 = 7;
+    TYPE_BOOL = 8;
+    TYPE_STRING = 9;
+    TYPE_GROUP = 10;
+    TYPE_MESSAGE = 11;
+    TYPE_BYTES = 12;
+    TYPE_UINT32 = 13;
+    TYPE_ENUM = 14;
+    TYPE_SFIXED32 = 15;
+    TYPE_SFIXED64 = 16;
+    TYPE_SINT32 = 17;
+    TYPE_SINT64 = 18;
+  }
+
+  enum Label {
+    LABEL_OPTIONAL = 1;
+    LABEL_REQUIRED = 2;
+    LABEL_REPEATED = 3;
+  }
+
+  optional string name = 1;
+  optional int32 number = 3;
+  optional Label label = 4;
+  optional Type type = 5;
+  optional string type_name = 6;
+  optional string extendee = 2;
+  optional string default_value = 7;
+  optional int32 oneof_index = 9;
+  optional string json_name = 10;
+  optional FieldOptions options = 8;
+  optional bool proto3_optional = 17;
+}
+
+message OneofDescriptorProto {
+  optional string name = 1;
+  optional OneofOptions options = 2;
+}
+
+message EnumDescriptorProto {
+  optional string name = 1;
+  repeated EnumValueDescriptorProto value = 2;
+  optional EnumOptions options = 3;
+
+  message EnumReservedRange {
+    optional int32 start = 1;
+    optional int32 end = 2;
+  }
+
+  repeated EnumReservedRange reserved_range = 4;
+  repeated string reserved_name = 5;
+}
+
+message EnumValueDescriptorProto {
+  optional string name = 1;
+  optional int32 number = 2;
+  optional EnumValueOptions options = 3;
+}
+
+message ServiceDescriptorProto {
+  optional string name = 1;
+  repeated MethodDescriptorProto method = 2;
+  optional ServiceOptions options = 3;
+}
+
+message MethodDescriptorProto {
+  optional string name = 1;
+  optional string input_type = 2;
+  optional string output_type = 3;
+  optional MethodOptions options = 4;
+  optional bool client_streaming = 5 [default = false];
+  optional bool server_streaming = 6 [default = false];
+}
+
+message FileOptions {
+  optional string java_package = 1;
+  optional string java_outer_classname = 8;
+  optional bool java_multiple_files = 10 [default = false];
+  optional bool java_generate_equals_and_hash = 20 [deprecated=true];
+  optional bool java_string_check_utf8 = 27 [default = false];
+  enum OptimizeMode {
+    SPEED = 1;
+    CODE_SIZE = 2;
+    LITE_RUNTIME = 3;
+  }
+  optional OptimizeMode optimize_for = 9 [default = SPEED];
+  optional string go_package = 11;
+  optional bool cc_generic_services = 16 [default = false];
+  optional bool java_generic_services = 17 [default = false];
+  optional bool py_generic_services = 18 [default = false];
+  optional bool php_generic_services = 42 [default = false];
+  optional bool deprecated = 23 [default = false];
+  optional bool cc_enable_arenas = 31 [default = true];
+  optional string objc_class_prefix = 36;
+  optional string csharp_namespace = 37;
+  optional string swift_prefix = 39;
+  optional string php_class_prefix = 40;
+  optional string php_namespace = 41;
+  optional string php_metadata_namespace = 44;
+  optional string ruby_package = 45;
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+  reserved 38;
+}
+
+message MessageOptions {
+  optional bool message_set_wire_format = 1 [default = false];
+  optional bool no_standard_descriptor_accessor = 2 [default = false];
+  optional bool deprecated = 3 [default = false];
+  optional bool map_entry = 7;
+  reserved 8;
+  reserved 9;
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message FieldOptions {
+  optional CType ctype = 1 [default = STRING];
+  enum CType {
+    STRING = 0;
+    CORD = 1;
+    STRING_PIECE = 2;
+  }
+  optional bool packed = 2;
+  optional JSType jstype = 6 [default = JS_NORMAL];
+  enum JSType {
+    JS_NORMAL = 0;
+    JS_STRING = 1;
+    JS_NUMBER = 2;
+  }
+  optional bool lazy = 5 [default = false];
+  optional bool deprecated = 3 [default = false];
+  optional bool weak = 10 [default = false];
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+
+  reserved 4;
+}
+
+message OneofOptions {
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message EnumOptions {
+  optional bool allow_alias = 2;
+  optional bool deprecated = 3 [default = false];
+  reserved 5;
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message EnumValueOptions {
+  optional bool deprecated = 1 [default = false];
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message ServiceOptions {
+  optional bool deprecated = 33 [default = false];
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message MethodOptions {
+  optional bool deprecated = 33 [default = false];
+  enum IdempotencyLevel {
+    IDEMPOTENCY_UNKNOWN = 0;
+    NO_SIDE_EFFECTS = 1;
+    IDEMPOTENT = 2;
+  }
+  optional IdempotencyLevel idempotency_level = 34
+      [default = IDEMPOTENCY_UNKNOWN];
+  repeated UninterpretedOption uninterpreted_option = 999;
+  extensions 1000 to max;
+}
+
+message UninterpretedOption {
+  message NamePart {
+    required string name_part = 1;
+    required bool is_extension = 2;
+  }
+  repeated NamePart name = 2;
+  optional string identifier_value = 3;
+  optional uint64 positive_int_value = 4;
+  optional int64 negative_int_value = 5;
+  optional double double_value = 6;
+  optional bytes string_value = 7;
+  optional string aggregate_value = 8;
+}
+
+message SourceCodeInfo {
+  repeated Location location = 1;
+  message Location {
+    repeated int32 path = 1 [packed = true];
+    repeated int32 span = 2 [packed = true];
+    optional string leading_comments = 3;
+    optional string trailing_comments = 4;
+    repeated string leading_detached_comments = 6;
+  }
+}
+
+message GeneratedCodeInfo {
+  repeated Annotation annotation = 1;
+  message Annotation {
+    repeated int32 path = 1 [packed = true];
+    optional string source_file = 2;
+    optional int32 begin = 3;
+    optional int32 end = 4;
+  }
+}
diff --git a/test/suite_dissection.py b/test/suite_dissection.py
index c509a77c04..29fe232387 100644
--- a/test/suite_dissection.py
+++ b/test/suite_dissection.py
@@ -225,6 +225,26 @@ class case_dissect_protobuf(subprocesstest.SubprocessTestCase):
             ))
         self.assertTrue(self.grepOutput('tutorial.AddressBook'))
 
+    def test_protobuf_complex_syntax(self, cmd_tshark, features, dirs, capture_file):
+        '''Test Protobuf parsing complex syntax .proto files'''
+        well_know_types_dir = os.path.join(dirs.protobuf_lang_files_dir, 'well_know_types').replace('\\', '/')
+        complex_proto_files_dir = os.path.join(dirs.protobuf_lang_files_dir, 'complex_proto_files').replace('\\', '/')
+        self.assertRun((cmd_tshark,
+                '-r', capture_file('protobuf_udp_addressbook_with_image_ts.pcapng'),
+                '-o', 'uat:protobuf_search_paths: "{}","{}"'.format(well_know_types_dir, 'FALSE'),
+                '-o', 'uat:protobuf_search_paths: "{}","{}"'.format(complex_proto_files_dir, 'TRUE'),
+                '-o', 'protobuf.preload_protos: TRUE',
+                '-o', 'protobuf.pbf_as_hf: TRUE',
+                '-Y', 'pbf.wireshark.protobuf.test.complex.syntax.TestFileParsed.last_field_for_wireshark_test'
+                      ' && pbf.protobuf_unittest.TestFileParsed.last_field_for_wireshark_test',
+            ))
+        # the output must be empty and not contain something like:
+        #   tshark: "pbf.xxx.TestFileParsed.last_field_for_wireshark_test" is neither a field nor a protocol name.
+        # or
+        #   tshark: Protobuf: Error(s)
+        self.assertFalse(self.grepOutput('.last_field_for_wireshark_test'))
+        self.assertFalse(self.grepOutput('Protobuf: Error'))
+
 @fixtures.mark_usefixtures('test_env')
 @fixtures.uses_fixtures
 class case_dissect_tcp(subprocesstest.SubprocessTestCase):