diff --git a/scripts/abi_check.py b/scripts/abi_check.py
index 502c7ae..e19f2c0 100755
--- a/scripts/abi_check.py
+++ b/scripts/abi_check.py
@@ -59,9 +59,7 @@
 
     @staticmethod
     def check_repo_path():
-        current_dir = os.path.realpath('.')
-        root_dir = os.path.dirname(os.path.dirname(os.path.realpath(__file__)))
-        if current_dir != root_dir:
+        if not all(os.path.isdir(d) for d in ["include", "library", "tests"]):
             raise Exception("Must be run from Mbed TLS root")
 
     def _setup_logger(self):
@@ -108,6 +106,12 @@
             stderr=subprocess.STDOUT
         )
         self.log.debug(worktree_output.decode("utf-8"))
+        version.commit = subprocess.check_output(
+            [self.git_command, "rev-parse", "HEAD"],
+            cwd=git_worktree_path,
+            stderr=subprocess.STDOUT
+        ).decode("ascii").rstrip()
+        self.log.debug("Commit is {}".format(version.commit))
         return git_worktree_path
 
     def _update_git_submodules(self, git_worktree_path, version):
@@ -163,6 +167,13 @@
                     os.path.join(root, file)
                 )
 
+    @staticmethod
+    def _pretty_revision(version):
+        if version.revision == version.commit:
+            return version.revision
+        else:
+            return "{} ({})".format(version.revision, version.commit)
+
     def _get_abi_dumps_from_shared_libraries(self, version):
         """Generate the ABI dumps for the specified git revision.
         The shared libraries must have been built and the module paths
@@ -177,7 +188,7 @@
                 "abi-dumper",
                 module_path,
                 "-o", output_path,
-                "-lver", version.revision
+                "-lver", self._pretty_revision(version),
             ]
             abi_dump_output = subprocess.check_output(
                 abi_dump_command,
@@ -222,70 +233,84 @@
                 if not problems.getchildren():
                     report.remove(problems)
 
+    def _abi_compliance_command(self, mbed_module, output_path):
+        """Build the command to run to analyze the library mbed_module.
+        The report will be placed in output_path."""
+        abi_compliance_command = [
+            "abi-compliance-checker",
+            "-l", mbed_module,
+            "-old", self.old_version.abi_dumps[mbed_module],
+            "-new", self.new_version.abi_dumps[mbed_module],
+            "-strict",
+            "-report-path", output_path,
+        ]
+        if self.skip_file:
+            abi_compliance_command += ["-skip-symbols", self.skip_file,
+                                       "-skip-types", self.skip_file]
+        if self.brief:
+            abi_compliance_command += ["-report-format", "xml",
+                                       "-stdout"]
+        return abi_compliance_command
+
+    def _is_library_compatible(self, mbed_module, compatibility_report):
+        """Test if the library mbed_module has remained compatible.
+        Append a message regarding compatibility to compatibility_report."""
+        output_path = os.path.join(
+            self.report_dir, "{}-{}-{}.html".format(
+                mbed_module, self.old_version.revision,
+                self.new_version.revision
+            )
+        )
+        try:
+            subprocess.check_output(
+                self._abi_compliance_command(mbed_module, output_path),
+                stderr=subprocess.STDOUT
+            )
+        except subprocess.CalledProcessError as err:
+            if err.returncode != 1:
+                raise err
+            if self.brief:
+                self.log.info(
+                    "Compatibility issues found for {}".format(mbed_module)
+                )
+                report_root = ET.fromstring(err.output.decode("utf-8"))
+                self._remove_extra_detail_from_report(report_root)
+                self.log.info(ET.tostring(report_root).decode("utf-8"))
+            else:
+                self.can_remove_report_dir = False
+                compatibility_report.append(
+                    "Compatibility issues found for {}, "
+                    "for details see {}".format(mbed_module, output_path)
+                )
+            return False
+        compatibility_report.append(
+            "No compatibility issues for {}".format(mbed_module)
+        )
+        if not (self.keep_all_reports or self.brief):
+            os.remove(output_path)
+        return True
+
     def get_abi_compatibility_report(self):
         """Generate a report of the differences between the reference ABI
         and the new ABI. ABI dumps from self.old_version and self.new_version
         must be available."""
-        compatibility_report = ""
+        compatibility_report = ["Checking evolution from {} to {}".format(
+            self._pretty_revision(self.old_version),
+            self._pretty_revision(self.new_version)
+        )]
         compliance_return_code = 0
         shared_modules = list(set(self.old_version.modules.keys()) &
                               set(self.new_version.modules.keys()))
         for mbed_module in shared_modules:
-            output_path = os.path.join(
-                self.report_dir, "{}-{}-{}.html".format(
-                    mbed_module, self.old_version.revision,
-                    self.new_version.revision
-                )
-            )
-            abi_compliance_command = [
-                "abi-compliance-checker",
-                "-l", mbed_module,
-                "-old", self.old_version.abi_dumps[mbed_module],
-                "-new", self.new_version.abi_dumps[mbed_module],
-                "-strict",
-                "-report-path", output_path,
-            ]
-            if self.skip_file:
-                abi_compliance_command += ["-skip-symbols", self.skip_file,
-                                           "-skip-types", self.skip_file]
-            if self.brief:
-                abi_compliance_command += ["-report-format", "xml",
-                                           "-stdout"]
-            try:
-                subprocess.check_output(
-                    abi_compliance_command,
-                    stderr=subprocess.STDOUT
-                )
-            except subprocess.CalledProcessError as err:
-                if err.returncode == 1:
-                    compliance_return_code = 1
-                    if self.brief:
-                        self.log.info(
-                            "Compatibility issues found for {}".format(mbed_module)
-                        )
-                        report_root = ET.fromstring(err.output.decode("utf-8"))
-                        self._remove_extra_detail_from_report(report_root)
-                        self.log.info(ET.tostring(report_root).decode("utf-8"))
-                    else:
-                        self.can_remove_report_dir = False
-                        compatibility_report += (
-                            "Compatibility issues found for {}, "
-                            "for details see {}\n".format(mbed_module, output_path)
-                        )
-                else:
-                    raise err
-            else:
-                compatibility_report += (
-                    "No compatibility issues for {}\n".format(mbed_module)
-                )
-                if not (self.keep_all_reports or self.brief):
-                    os.remove(output_path)
+            if not self._is_library_compatible(mbed_module,
+                                               compatibility_report):
+                compliance_return_code = 1
         for version in [self.old_version, self.new_version]:
             for mbed_module, mbed_module_dump in version.abi_dumps.items():
                 os.remove(mbed_module_dump)
         if self.can_remove_report_dir:
             os.rmdir(self.report_dir)
-        self.log.info(compatibility_report)
+        self.log.info("\n".join(compatibility_report))
         return compliance_return_code
 
     def check_for_abi_changes(self):
@@ -357,7 +382,9 @@
         )
         parser.add_argument(
             "-s", "--skip-file", type=str,
-            help="path to file containing symbols and types to skip"
+            help=("path to file containing symbols and types to skip "
+                  "(typically \"-s identifiers\" after running "
+                  "\"tests/scripts/list-identifiers.sh --internal\")")
         )
         parser.add_argument(
             "-b", "--brief", action="store_true",
@@ -371,6 +398,7 @@
             version="old",
             repository=abi_args.old_repo,
             revision=abi_args.old_rev,
+            commit=None,
             crypto_repository=abi_args.old_crypto_repo,
             crypto_revision=abi_args.old_crypto_rev,
             abi_dumps={},
@@ -380,6 +408,7 @@
             version="new",
             repository=abi_args.new_repo,
             revision=abi_args.new_rev,
+            commit=None,
             crypto_repository=abi_args.new_crypto_repo,
             crypto_revision=abi_args.new_crypto_rev,
             abi_dumps={},
diff --git a/tests/compat.sh b/tests/compat.sh
index 0eae1ea..80c2d31 100755
--- a/tests/compat.sh
+++ b/tests/compat.sh
@@ -15,6 +15,10 @@
 
 set -u
 
+# Limit the size of each log to 10 GiB, in case of failures with this script
+# where it may output seemingly unlimited length error logs.
+ulimit -f 20971520
+
 # initialise counters
 TESTS=0
 FAILED=0
diff --git a/tests/scripts/check-files.py b/tests/scripts/check-files.py
index 00fd0ed..255bed8 100755
--- a/tests/scripts/check-files.py
+++ b/tests/scripts/check-files.py
@@ -1,14 +1,12 @@
 #!/usr/bin/env python3
+
+# This file is part of Mbed TLS (https://tls.mbed.org)
+# Copyright (c) 2018, Arm Limited, All Rights Reserved
+
 """
-This file is part of Mbed TLS (https://tls.mbed.org)
-
-Copyright (c) 2018, Arm Limited, All Rights Reserved
-
-Purpose
-
 This script checks the current state of the source code for minor issues,
 including incorrect file permissions, presence of tabs, non-Unix line endings,
-trailing whitespace, presence of UTF-8 BOM, and TODO comments.
+trailing whitespace, and presence of UTF-8 BOM.
 Note: requires python 3, must be run from Mbed TLS root.
 """
 
@@ -170,19 +168,6 @@
             return True
         return False
 
-class TodoIssueTracker(LineIssueTracker):
-    """Track lines containing ``TODO``."""
-
-    heading = "TODO present:"
-    files_exemptions = frozenset([
-        os.path.basename(__file__),
-        "benchmark.c",
-        "pull_request_template.md",
-    ])
-
-    def issue_with_line(self, line, _filepath):
-        return b"todo" in line.lower()
-
 
 class IntegrityChecker(object):
     """Sanity-check files under the current directory."""
@@ -211,7 +196,6 @@
             TrailingWhitespaceIssueTracker(),
             TabIssueTracker(),
             MergeArtifactIssueTracker(),
-            TodoIssueTracker(),
         ]
 
     @staticmethod
@@ -257,15 +241,7 @@
 
 
 def run_main():
-    parser = argparse.ArgumentParser(
-        description=(
-            "This script checks the current state of the source code for "
-            "minor issues, including incorrect file permissions, "
-            "presence of tabs, non-Unix line endings, trailing whitespace, "
-            "presence of UTF-8 BOM, and TODO comments. "
-            "Note: requires python 3, must be run from Mbed TLS root."
-        )
-    )
+    parser = argparse.ArgumentParser(description=__doc__)
     parser.add_argument(
         "-l", "--log_file", type=str, help="path to optional output log",
     )
diff --git a/tests/scripts/mbedtls_test.py b/tests/scripts/mbedtls_test.py
index ac2912d..6ac68a4 100755
--- a/tests/scripts/mbedtls_test.py
+++ b/tests/scripts/mbedtls_test.py
@@ -79,7 +79,7 @@
         split_colon_fn = lambda x: re.sub(r'\\' + split_char, split_char, x)
         if len(split_char) > 1:
             raise ValueError('Expected split character. Found string!')
-        out = map(split_colon_fn, re.split(r'(?<!\\)' + split_char, inp_str))
+        out = list(map(split_colon_fn, re.split(r'(?<!\\)' + split_char, inp_str)))
         out = [x for x in out if x]
         return out
 
@@ -99,11 +99,11 @@
 
             # Check dependencies
             dependencies = []
-            line = data_f.next().strip()
+            line = next(data_f).strip()
             match = re.search('depends_on:(.*)', line)
             if match:
                 dependencies = [int(x) for x in match.group(1).split(':')]
-                line = data_f.next().strip()
+                line = next(data_f).strip()
 
             # Read test vectors
             line = line.replace('\\n', '\n')
@@ -115,7 +115,7 @@
                 err_str_fmt = "Number of test arguments({}) should be even: {}"
                 raise TestDataParserError(err_str_fmt.format(args_count, line))
             grouped_args = [(args[i * 2], args[(i * 2) + 1])
-                            for i in range(len(args)/2)]
+                            for i in range(int(len(args)/2))]
             self.tests.append((name, function_name, dependencies,
                                grouped_args))
 
@@ -261,21 +261,21 @@
         data_bytes += bytearray([function_id, len(parameters)])
         for typ, param in parameters:
             if typ == 'int' or typ == 'exp':
-                i = int(param)
-                data_bytes += 'I' if typ == 'int' else 'E'
+                i = int(param, 0)
+                data_bytes += b'I' if typ == 'int' else b'E'
                 self.align_32bit(data_bytes)
                 data_bytes += self.int32_to_big_endian_bytes(i)
             elif typ == 'char*':
                 param = param.strip('"')
                 i = len(param) + 1  # + 1 for null termination
-                data_bytes += 'S'
+                data_bytes += b'S'
                 self.align_32bit(data_bytes)
                 data_bytes += self.int32_to_big_endian_bytes(i)
-                data_bytes += bytearray(list(param))
-                data_bytes += '\0'   # Null terminate
+                data_bytes += bytearray(param, encoding='ascii')
+                data_bytes += b'\0'   # Null terminate
             elif typ == 'hex':
                 binary_data = self.hex_str_bytes(param)
-                data_bytes += 'H'
+                data_bytes += b'H'
                 self.align_32bit(data_bytes)
                 i = len(binary_data)
                 data_bytes += self.int32_to_big_endian_bytes(i)
@@ -310,7 +310,7 @@
 
         param_bytes, length = self.test_vector_to_bytes(function_id,
                                                         dependencies, args)
-        self.send_kv(length, param_bytes)
+        self.send_kv(''.join('{:02x}'.format(x) for x in length), ''.join('{:02x}'.format(x) for x in param_bytes))
 
     @staticmethod
     def get_result(value):
diff --git a/tests/ssl-opt.sh b/tests/ssl-opt.sh
index fc0e1d0..238ed97 100755
--- a/tests/ssl-opt.sh
+++ b/tests/ssl-opt.sh
@@ -21,6 +21,10 @@
 
 set -u
 
+# Limit the size of each log to 10 GiB, in case of failures with this script
+# where it may output seemingly unlimited length error logs.
+ulimit -f 20971520
+
 if cd $( dirname $0 ); then :; else
     echo "cd $( dirname $0 ) failed" >&2
     exit 1
diff --git a/tests/suites/helpers.function b/tests/suites/helpers.function
index a33cc03..ccd4d42 100644
--- a/tests/suites/helpers.function
+++ b/tests/suites/helpers.function
@@ -108,6 +108,21 @@
        }                                                    \
     } while( 0 )
 
+/**
+ * \brief   This macro tests the expression passed to it and skips the
+ *          running test if it doesn't evaluate to 'true'.
+ *
+ * \param   TEST    The test expression to be tested.
+ */
+#define TEST_ASSUME( TEST )                         \
+    do {                                            \
+        if( ! (TEST) )                              \
+        {                                           \
+            test_skip( #TEST, __LINE__, __FILE__ ); \
+            goto exit;                              \
+        }                                           \
+    } while( 0 )
+
 #if defined(MBEDTLS_CHECK_PARAMS) && !defined(MBEDTLS_PARAM_FAILED_ALT)
 /**
  * \brief   This macro tests the statement passed to it as a test step or
@@ -241,10 +256,17 @@
 /*----------------------------------------------------------------------------*/
 /* Global variables */
 
+typedef enum
+{
+    TEST_RESULT_SUCCESS = 0,
+    TEST_RESULT_FAILED,
+    TEST_RESULT_SKIPPED
+} test_result_t;
+
 static struct
 {
     paramfail_test_state_t paramfail_test_state;
-    int failed;
+    test_result_t result;
     const char *test;
     const char *filename;
     int line_no;
@@ -280,7 +302,15 @@
 
 void test_fail( const char *test, int line_no, const char* filename )
 {
-    test_info.failed = 1;
+    test_info.result = TEST_RESULT_FAILED;
+    test_info.test = test;
+    test_info.line_no = line_no;
+    test_info.filename = filename;
+}
+
+void test_skip( const char *test, int line_no, const char* filename )
+{
+    test_info.result = TEST_RESULT_SKIPPED;
     test_info.test = test;
     test_info.line_no = line_no;
     test_info.filename = filename;
@@ -319,7 +349,7 @@
         /* Record the location of the failure, but not as a failure yet, in case
          * it was part of the test */
         test_fail( failure_condition, line, file );
-        test_info.failed = 0;
+        test_info.result = TEST_RESULT_SUCCESS;
 
         longjmp( param_fail_jmp, 1 );
     }
diff --git a/tests/suites/host_test.function b/tests/suites/host_test.function
index fe6a2bc..0f98d23 100644
--- a/tests/suites/host_test.function
+++ b/tests/suites/host_test.function
@@ -498,7 +498,8 @@
 
             if( ( ret = get_line( file, buf, sizeof(buf) ) ) != 0 )
                 break;
-            mbedtls_fprintf( stdout, "%s%.66s", test_info.failed ? "\n" : "", buf );
+            mbedtls_fprintf( stdout, "%s%.66s",
+                    test_info.result == TEST_RESULT_FAILED ? "\n" : "", buf );
             mbedtls_fprintf( stdout, " " );
             for( i = strlen( buf ) + 1; i < 67; i++ )
                 mbedtls_fprintf( stdout, "." );
@@ -545,7 +546,7 @@
             // If there are no unmet dependencies execute the test
             if( unmet_dep_count == 0 )
             {
-                test_info.failed = 0;
+                test_info.result = TEST_RESULT_SUCCESS;
                 test_info.paramfail_test_state = PARAMFAIL_TESTSTATE_IDLE;
 
 #if defined(__unix__) || (defined(__APPLE__) && defined(__MACH__))
@@ -610,10 +611,15 @@
             }
             else if( ret == DISPATCH_TEST_SUCCESS )
             {
-                if( test_info.failed == 0 )
+                if( test_info.result == TEST_RESULT_SUCCESS )
                 {
                     mbedtls_fprintf( stdout, "PASS\n" );
                 }
+                else if( test_info.result == TEST_RESULT_SKIPPED )
+                {
+                    mbedtls_fprintf( stdout, "----\n" );
+                    total_skipped++;
+                }
                 else
                 {
                     total_errors++;
diff --git a/tests/suites/main_test.function b/tests/suites/main_test.function
index ca4783d..1f4180d 100644
--- a/tests/suites/main_test.function
+++ b/tests/suites/main_test.function
@@ -156,7 +156,7 @@
     else
     {
         /* Unexpected parameter validation error */
-        test_info.failed = 1;
+        test_info.result = TEST_RESULT_FAILED;
     }
 
     memset( param_fail_jmp, 0, sizeof(jmp_buf) );
diff --git a/tests/suites/target_test.function b/tests/suites/target_test.function
index e4c3e30..d430d9d 100644
--- a/tests/suites/target_test.function
+++ b/tests/suites/target_test.function
@@ -59,10 +59,29 @@
     return( DEPENDENCY_SUPPORTED );
 }
 
+/**
+ * \brief       Receives hex string on serial interface, and converts to a byte.
+ *
+ * \param none
+ *
+ * \return      unsigned int8
+ */
+uint8_t receive_byte()
+{
+    uint8_t byte;
+    uint8_t c[3];
+    char *endptr;
+    c[0] = greentea_getc();
+    c[1] = greentea_getc();
+    c[2] = '\0';
+
+    assert( unhexify( &byte, c ) != 2 );
+    return( byte );
+}
 
 /**
  * \brief       Receives unsigned integer on serial interface.
- *              Integers are encoded in network order.
+ *              Integers are encoded in network order, and sent as hex ascii string.
  *
  * \param none
  *
@@ -71,10 +90,17 @@
 uint32_t receive_uint32()
 {
     uint32_t value;
-    value =  (uint8_t)greentea_getc() << 24;
-    value |= (uint8_t)greentea_getc() << 16;
-    value |= (uint8_t)greentea_getc() << 8;
-    value |= (uint8_t)greentea_getc();
+    const uint8_t c[9] = { greentea_getc(),
+                           greentea_getc(),
+                           greentea_getc(),
+                           greentea_getc(),
+                           greentea_getc(),
+                           greentea_getc(),
+                           greentea_getc(),
+                           greentea_getc(),
+                           '\0'
+                         };
+    assert( unhexify( &value, c ) != 8 );
     return( (uint32_t)value );
 }
 
@@ -132,7 +158,7 @@
     greentea_getc(); // read ';' received after key i.e. *data_len
 
     for( i = 0; i < *data_len; i++ )
-        data[i] = greentea_getc();
+        data[i] = receive_byte();
 
     /* Read closing braces */
     for( i = 0; i < 2; i++ )