Fix langAltValue::read() parsing

+ Fix segmentation faults in langAlt parse
+ Fix mismatched quotation marks and incorrect values
+ Add Python testing
  + Some tests commented out as quotation marks are filtered, preventing them
    from running.
Closes #1481.

include/exiv2/error.hpp

@@ -249,6 +249,7 @@ namespace Exiv2 {
         kerInvalidXMP,
         kerTiffDirectoryTooLarge,
         kerInvalidTypeValue,
+        kerInvalidLangAltValue,
         kerInvalidMalloc,
         kerCorruptedMetadata,
         kerArithmeticOverflow,

src/error.cpp

@@ -158,6 +158,8 @@ namespace {
           N_("tiff directory length is too large") },
         { Exiv2::kerInvalidTypeValue,
           N_("invalid type in tiff structure") },
+        { Exiv2::kerInvalidLangAltValue,
+          N_("Invalid LangAlt value `%1'") }, // %1=value
         { Exiv2::kerInvalidMalloc,
           N_("invalid memory allocation request") },
         { Exiv2::kerCorruptedMetadata,

src/value.cpp

@@ -856,18 +856,39 @@ namespace Exiv2 {
     }
 
     int LangAltValue::read(const std::string& buf)
-    {
+    {        
         std::string b = buf;
         std::string lang = "x-default";
         if (buf.length() > 5 && buf.substr(0, 5) == "lang=") {
+            const char* ALPLHA = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
+            const char* ALPLHA_NUM = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789";
+            
             std::string::size_type pos = buf.find_first_of(' ');
             lang = buf.substr(5, pos-5);
             // Strip quotes (so you can also specify the language without quotes)
-            if (lang[0] == '"') lang = lang.substr(1);
-            if (lang[lang.length()-1] == '"') lang = lang.substr(0, lang.length()-1);
+            if (lang[0] == '"') {
+                lang = lang.substr(1);
+
+                if (lang == "" || lang.find('"') != lang.length()-1)
+                    throw Error(kerInvalidLangAltValue, buf); // ***
+            
+                lang = lang.substr(0, lang.length()-1);
+            }
+            
+            if (lang == "") throw Error(kerInvalidLangAltValue, buf); // ***
+
+            // Check language is in the correct format (see https://www.ietf.org/rfc/rfc3066.txt)
+            std::string::size_type charPos = lang.find_first_not_of(ALPLHA);
+            if (charPos != std::string::npos) {
+                if (lang[charPos] != '-' || lang.find_first_not_of(ALPLHA_NUM, charPos+1) != std::string::npos)
+                    throw Error(kerInvalidLangAltValue, buf);
+            }
+            
             b.clear();
             if (pos != std::string::npos) b = buf.substr(pos+1);
         }
+
+
         value_[lang] = b;
         return 0;
     }

tests/bugfixes/github/test_issue_1481.py

@@ -0,0 +1,160 @@
+# -*- coding: utf-8 -*-
+
+import system_tests
+
+@system_tests.CopyFiles("$data_path/exiv2-empty.jpg")
+class CheckXmpLangAltValues(metaclass=system_tests.CaseMeta):
+
+    url = "https://github.com/Exiv2/exiv2/issues/1481"
+
+# Python unittest is filtering out empty pairs of "" and flags up an error if 
+# a mismatched number of quotes is used inside the commands[] (see
+# https://docs.python.org/3/library/shlex.html#parsing-rules).
+#
+# This means that some of the tests in the github issue cannot be run.
+
+    langAltValue = [
+        # 1. No language value
+        """lang= test1-1""",
+        """lang=\" test1-2""",
+
+        # 2. Empty language value
+        """lang=\"\" test2""",
+ 
+        # 3. Mismatched and/or incorrect positioning of quotation marks
+        """lang=\"\"test3-1""",
+        """lang=\"test3-2""",
+        """lang=\"en-UK test3-3""",
+        """lang=en-US\" test3-4""",
+        """lang=test3-5\"""",
+        """lang=test3-6\"\"""",
+        
+        # 4. Invalid characters in language part
+        """lang=en-UK- test4-1""",
+        """lang=en=UK test4-2""",
+    ]
+
+    filename = system_tests.path("$data_path/exiv2-empty_copy.jpg")
+
+    commands = [
+                # 1. No language value
+                """$exiv2 -M"set Xmp.dc.title """ + langAltValue[0]  + """" $filename""",
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[1]  + """" $filename""",
+                """$exiv2 -px                                               $filename""",
+        
+                # 2. Empty language value
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[2]  + """" $filename""",
+#               """$exiv2 -px                                               $filename""",
+        
+                # 3. Mismatched and/or incorrect positioning of quotation marks
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[3]  + """" $filename""",
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[4]  + """" $filename""",
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[5]  + """" $filename""",
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[6]  + """" $filename""",
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[7]  + """" $filename""",
+#               """$exiv2 -M"set Xmp.dc.title """ + langAltValue[8]  + """" $filename""",
+#               """$exiv2 -px                                               $filename""",
+                
+                # 4. Invalid characters in language part
+                """$exiv2 -M"set Xmp.dc.title """ + langAltValue[9]  + """" $filename""",
+                """$exiv2 -M"set Xmp.dc.title """ + langAltValue[10] + """" $filename""",
+                """$exiv2 -px                                               $filename"""
+    ]
+
+    stdout = [
+              # 1. No language value
+              "",
+#             "",
+              "",
+              
+              # 2. Empty language value
+#             "",
+#             "",
+              
+              # 3. Mismatched and/or incorrect positioning of quotation marks
+#             "",
+#             "",
+#             "",
+#             "",
+#             "",
+#             "",
+#             "",
+              
+              # 4. Invalid characters in language part
+              "",
+              "",
+              ""
+    ]
+
+    stderr = [
+              # 1. No language value
+              """$exiv2_modify_exception_message $filename:
+$kerInvalidLangAltValue `""" + langAltValue[0]  + """'
+""",
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[1]  + """'
+# """,
+              "",
+                      
+              # 2. Empty language value
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[2]  + """'
+# """,
+#             "",
+        
+              # 3. Mismatched and/or incorrect positioning of quotation marks
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[3]  + """'
+# """,
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[4]  + """'
+# """,
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[5]  + """'
+# """,
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[6]  + """'
+# """,
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[7]  + """'
+# """,
+#             """$exiv2_modify_exception_message $filename:
+# $kerInvalidLangAltValue `""" + langAltValue[8]  + """'
+# """,
+#             "",
+              
+              # 4. Invalid characters in language part
+              """$exiv2_modify_exception_message $filename:
+$kerInvalidLangAltValue `""" + langAltValue[9]  + """'
+""",
+              """$exiv2_modify_exception_message $filename:
+$kerInvalidLangAltValue `""" + langAltValue[10]  + """'
+""",
+              ""
+    ]
+        
+    retval = [
+              # 1. No language value
+              1,
+#             1,
+              0,
+        
+              # 2. Empty language value
+#             1,
+#             0,
+              
+              # 3. Mismatched and/or incorrect positioning of quotation marks
+#             1,
+#             1,
+#             1,
+#             1,
+#             1,
+#             1,
+#             0,
+              
+              # 4. Invalid characters in language part
+              1,
+              1,
+              0
+    ]
+

tests/suite.conf

@@ -42,8 +42,10 @@ kerInvalidTypeValue: invalid type in tiff structure
 kerNotAJpeg : This does not look like a JPEG image
 kerNoImageInInputData: Input data does not contain a valid image
 kerFileContainsUnknownImageType: The file contains data of an unknown image type
+kerInvalidLangAltValue: Invalid LangAlt value
 addition_overflow_message: Overflow in addition
 exiv2_exception_message: Exiv2 exception in print action for file
+exiv2_modify_exception_message: Exiv2 exception in modify action for file
 exiv2_overflow_exception_message: std::overflow_error exception in print action for file
 exception_in_extract: Exiv2 exception in extract action for file
 uncaught_exception: Uncaught exception: