Merge pull request #480 from Ana06/py3-only

.github/workflows/publish.yml

@@ -15,7 +15,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v2
         with:
-          python-version: '2.7'
+          python-version: '3.6'
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip
@@ -27,3 +27,4 @@ jobs:
         run: |
           python setup.py sdist bdist_wheel
           twine upload --skip-existing dist/*
+

.github/workflows/tests.yml

@@ -52,8 +52,6 @@ jobs:
         python-version: [3.6, 3.9]
         include:
           # on Ubuntu run these as well
-          - os: ubuntu-20.04
-            python-version: 2.7
           - os: ubuntu-20.04
             python-version: 3.7
           - os: ubuntu-20.04

CHANGELOG.md

@@ -2,6 +2,8 @@
 
 ## master (unreleased)
 
+The first Python 3 ONLY capa version.
+
 ### New Features
 
 ### New Rules
@@ -10,6 +12,8 @@
 
 ### Changes
 
+- py3: drop Python 2 support #480 @Ana06
+
 ### Development
 
 ### Raw diffs

capa/features/__init__.py

@@ -27,10 +27,7 @@ VALID_ARCH = (ARCH_X32, ARCH_X64)
 
 
 def bytes_to_str(b):
-    if sys.version_info[0] >= 3:
-        return str(codecs.encode(b, "hex").decode("utf-8"))
-    else:
-        return codecs.encode(b, "hex")
+    return str(codecs.encode(b, "hex").decode("utf-8"))
 
 
 def hex_string(h):

capa/features/extractors/helpers.py

@@ -16,10 +16,7 @@ MIN_STACKSTRING_LEN = 8
 
 
 def xor_static(data, i):
-    if sys.version_info >= (3, 0):
-        return bytes(c ^ i for c in data)
-    else:
-        return "".join(chr(ord(c) ^ i) for c in data)
+    return bytes(c ^ i for c in data)
 
 
 def is_aw_function(symbol):

capa/features/extractors/ida/__init__.py

@@ -34,10 +34,7 @@ def add_ea_int_cast(o):
     this bit of skullduggery lets use cast viv-utils objects as ints.
     the correct way of doing this is to update viv-utils (or subclass the objects here).
     """
-    if sys.version_info[0] >= 3:
-        setattr(o, "__int__", types.MethodType(get_ea, o))
-    else:
-        setattr(o, "__int__", types.MethodType(get_ea, o, type(o)))
+    setattr(o, "__int__", types.MethodType(get_ea, o))
     return o
 
 

capa/features/extractors/ida/basicblock.py

@@ -39,18 +39,11 @@ def get_printable_len(op):
         raise ValueError("Unhandled operand data type 0x%x." % op.dtype)
 
     def is_printable_ascii(chars):
-        if sys.version_info[0] >= 3:
-            return all(c < 127 and chr(c) in string.printable for c in chars)
-        else:
-            return all(ord(c) < 127 and c in string.printable for c in chars)
+        return all(c < 127 and chr(c) in string.printable for c in chars)
 
     def is_printable_utf16le(chars):
-        if sys.version_info[0] >= 3:
-            if all(c == 0x00 for c in chars[1::2]):
-                return is_printable_ascii(chars[::2])
-        else:
-            if all(c == "\x00" for c in chars[1::2]):
-                return is_printable_ascii(chars[::2])
+        if all(c == 0x00 for c in chars[1::2]):
+            return is_printable_ascii(chars[::2])
 
     if is_printable_ascii(chars):
         return idaapi.get_dtype_size(op.dtype)

capa/features/extractors/ida/helpers.py

@@ -23,11 +23,7 @@ def find_byte_sequence(start, end, seq):
         end: max virtual address
         seq: bytes to search e.g. b"\x01\x03"
     """
-    if sys.version_info[0] >= 3:
-        seq = " ".join(["%02x" % b for b in seq])
-    else:
-        seq = " ".join(["%02x" % ord(b) for b in seq])
-
+    seq = " ".join(["%02x" % b for b in seq])
     while True:
         ea = idaapi.find_binary(start, end, seq, 0, idaapi.SEARCH_DOWN)
         if ea == idaapi.BADADDR:

capa/features/extractors/smda/__init__.py

@@ -15,8 +15,6 @@ from capa.features.extractors import FeatureExtractor
 class SmdaFeatureExtractor(FeatureExtractor):
     def __init__(self, smda_report: SmdaReport, path):
         super(SmdaFeatureExtractor, self).__init__()
-        if sys.version_info < (3, 0):
-            raise UnsupportedRuntimeError("SMDA should only be used with Python 3.")
         self.smda_report = smda_report
         self.path = path
 

capa/features/freeze.py

@@ -264,15 +264,14 @@ def main(argv=None):
     parser.add_argument(
         "-f", "--format", choices=[f[0] for f in formats], default="auto", help="Select sample format, %s" % format_help
     )
-    if sys.version_info >= (3, 0):
-        parser.add_argument(
-            "-b",
-            "--backend",
-            type=str,
-            help="select the backend to use",
-            choices=(capa.main.BACKEND_VIV, capa.main.BACKEND_SMDA),
-            default=capa.main.BACKEND_VIV,
-        )
+    parser.add_argument(
+        "-b",
+        "--backend",
+        type=str,
+        help="select the backend to use",
+        choices=(capa.main.BACKEND_VIV, capa.main.BACKEND_SMDA),
+        default=capa.main.BACKEND_VIV,
+    )
     args = parser.parse_args(args=argv)
 
     if args.quiet:
@@ -285,8 +284,7 @@ def main(argv=None):
         logging.basicConfig(level=logging.INFO)
         logging.getLogger().setLevel(logging.INFO)
 
-    backend = args.backend if sys.version_info > (3, 0) else capa.main.BACKEND_VIV
-    extractor = capa.main.get_extractor(args.sample, args.format, backend)
+    extractor = capa.main.get_extractor(args.sample, args.format, args.backend)
     with open(args.output, "wb") as f:
         f.write(dump(extractor))
 

capa/helpers.py

@@ -12,9 +12,7 @@ _hex = hex
 
 
 def hex(i):
-    # under py2.7, long integers get formatted with a trailing `L`
-    # and this is not pretty. so strip it out.
-    return _hex(oint(i)).rstrip("L")
+    return _hex(oint(i))
 
 
 def oint(i):

capa/ida/helpers/__init__.py

@@ -10,7 +10,6 @@ import logging
 import datetime
 
 import idc
-import six
 import idaapi
 import idautils
 
@@ -85,7 +84,7 @@ def get_func_start_ea(ea):
 def get_file_md5():
     """ """
     md5 = idautils.GetInputFileMD5()
-    if not isinstance(md5, six.string_types):
+    if not isinstance(md5, str):
         md5 = capa.features.bytes_to_str(md5)
     return md5
 
@@ -93,7 +92,7 @@ def get_file_md5():
 def get_file_sha256():
     """ """
     sha256 = idaapi.retrieve_input_file_sha256()
-    if not isinstance(sha256, six.string_types):
+    if not isinstance(sha256, str):
         sha256 = capa.features.bytes_to_str(sha256)
     return sha256
 

capa/ida/plugin/README.md

@@ -36,7 +36,7 @@ For more information on the FLARE team's open-source framework, capa, check out
 
 capa explorer supports the following IDA setups:
 
-* IDA Pro 7.4+ with Python 2.7 or Python 3.
+* IDA Pro 7.4+ with Python >= 3.6.
 
 If you encounter issues with your specific setup, please open a new [Issue](https://github.com/fireeye/capa/issues).
 

capa/ida/plugin/item.py

@@ -328,14 +328,10 @@ class CapaExplorerByteViewItem(CapaExplorerFeatureItem):
         """
         byte_snap = idaapi.get_bytes(location, 32)
 
+        details = ""
         if byte_snap:
             byte_snap = codecs.encode(byte_snap, "hex").upper()
-            if sys.version_info >= (3, 0):
-                details = " ".join([byte_snap[i : i + 2].decode() for i in range(0, len(byte_snap), 2)])
-            else:
-                details = " ".join([byte_snap[i : i + 2] for i in range(0, len(byte_snap), 2)])
-        else:
-            details = ""
+            details = " ".join([byte_snap[i : i + 2].decode() for i in range(0, len(byte_snap), 2)])
 
         super(CapaExplorerByteViewItem, self).__init__(parent, display, location=location, details=details)
         self.ida_highlight = idc.get_color(location, idc.CIC_ITEM)

capa/ida/plugin/proxy.py

@@ -5,7 +5,6 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-import six
 from PyQt5 import QtCore
 from PyQt5.QtCore import Qt
 
@@ -208,7 +207,7 @@ class CapaExplorerSearchProxyModel(QtCore.QSortFilterProxyModel):
             if not data:
                 continue
 
-            if not isinstance(data, six.string_types):
+            if not isinstance(data, str):
                 # sanity check: should already be a string, but double check
                 continue
 

capa/main.py

@@ -1,4 +1,4 @@
-#!/usr/bin/env python2
+#!/usr/bin/env python3
 """
 Copyright (C) 2020 FireEye, Inc. All Rights Reserved.
 Licensed under the Apache License, Version 2.0 (the "License");
@@ -288,26 +288,15 @@ def get_workspace(path, format, should_save=True):
     return vw
 
 
-def get_extractor_py2(path, format, disable_progress=False):
-    import capa.features.extractors.viv
-
-    with halo.Halo(text="analyzing program", spinner="simpleDots", stream=sys.stderr, enabled=not disable_progress):
-        vw = get_workspace(path, format, should_save=False)
-
-        try:
-            vw.saveWorkspace()
-        except IOError:
-            # see #168 for discussion around how to handle non-writable directories
-            logger.info("source directory is not writable, won't save intermediate workspace")
-
-    return capa.features.extractors.viv.VivisectFeatureExtractor(vw, path)
-
-
 class UnsupportedRuntimeError(RuntimeError):
     pass
 
 
-def get_extractor_py3(path, format, backend, disable_progress=False):
+def get_extractor(path, format, backend, disable_progress=False):
+    """
+    raises:
+      UnsupportedFormatError:
+    """
     if backend == "smda":
         from smda.SmdaConfig import SmdaConfig
         from smda.Disassembler import Disassembler
@@ -337,17 +326,6 @@ def get_extractor_py3(path, format, backend, disable_progress=False):
         return capa.features.extractors.viv.VivisectFeatureExtractor(vw, path)
 
 
-def get_extractor(path, format, backend, disable_progress=False):
-    """
-    raises:
-      UnsupportedFormatError:
-    """
-    if sys.version_info >= (3, 0):
-        return get_extractor_py3(path, format, backend, disable_progress=disable_progress)
-    else:
-        return get_extractor_py2(path, format, disable_progress=disable_progress)
-
-
 def is_nursery_rule_path(path):
     """
     The nursery is a spot for rules that have not yet been fully polished.
@@ -460,7 +438,7 @@ def install_common_args(parser, wanted=None):
       wanted (Set[str]): collection of arguments to opt-into, including:
         - "sample": required positional argument to input file.
         - "format": flag to override file format.
-        - "backend": flag to override analysis backend under py3.
+        - "backend": flag to override analysis backend.
         - "rules": flag to override path to capa rules.
         - "tag": flag to override/specify which rules to match.
     """
@@ -498,22 +476,11 @@ def install_common_args(parser, wanted=None):
     #
 
     if "sample" in wanted:
-        if sys.version_info >= (3, 0):
-            parser.add_argument(
-                # Python 3 str handles non-ASCII arguments correctly
-                "sample",
-                type=str,
-                help="path to sample to analyze",
-            )
-        else:
-            parser.add_argument(
-                # in #328 we noticed that the sample path is not handled correctly if it contains non-ASCII characters
-                # https://stackoverflow.com/a/22947334/ offers a solution and decoding using getfilesystemencoding works
-                # in our testing, however other sources suggest `sys.stdin.encoding` (https://stackoverflow.com/q/4012571/)
-                "sample",
-                type=lambda s: s.decode(sys.getfilesystemencoding()),
-                help="path to sample to analyze",
-            )
+        parser.add_argument(
+            "sample",
+            type=str,
+            help="path to sample to analyze",
+        )
 
     if "format" in wanted:
         formats = [
@@ -532,15 +499,15 @@ def install_common_args(parser, wanted=None):
             help="select sample format, %s" % format_help,
         )
 
-    if "backend" in wanted and sys.version_info >= (3, 0):
-        parser.add_argument(
-            "-b",
-            "--backend",
-            type=str,
-            help="select the backend to use",
-            choices=(BACKEND_VIV, BACKEND_SMDA),
-            default=BACKEND_VIV,
-        )
+        if "backend" in wanted:
+            parser.add_argument(
+                "-b",
+                "--backend",
+                type=str,
+                help="select the backend to use",
+                choices=(BACKEND_VIV, BACKEND_SMDA),
+                default=BACKEND_VIV,
+            )
 
     if "rules" in wanted:
         parser.add_argument(
@@ -576,10 +543,9 @@ def handle_common_args(args):
     # disable vivisect-related logging, it's verbose and not relevant for capa users
     set_vivisect_log_level(logging.CRITICAL)
 
-    # py2 doesn't know about cp65001, which is a variant of utf-8 on windows
-    # tqdm bails when trying to render the progress bar in this setup.
-    # because cp65001 is utf-8, we just map that codepage to the utf-8 codec.
-    # see #380 and: https://stackoverflow.com/a/3259271/87207
+    # Since Python 3.8 cp65001 is an alias to utf_8, but not for Pyhton < 3.8
+    # TODO: remove this code when only supporting Python 3.8+
+    # https://stackoverflow.com/a/3259271/87207
     import codecs
 
     codecs.register(lambda name: codecs.lookup("utf-8") if name == "cp65001" else None)
@@ -599,6 +565,9 @@ def handle_common_args(args):
 
 
 def main(argv=None):
+    if sys.version_info < (3, 6):
+        raise UnsupportedRuntimeError("This version of capa can only be used with Python 3.6+")
+
     if argv is None:
         argv = sys.argv[1:]
 
@@ -703,8 +672,7 @@ def main(argv=None):
     else:
         format = args.format
         try:
-            backend = args.backend if sys.version_info > (3, 0) else BACKEND_VIV
-            extractor = get_extractor(args.sample, args.format, backend, disable_progress=args.quiet)
+            extractor = get_extractor(args.sample, args.format, args.backend, disable_progress=args.quiet)
         except UnsupportedFormatError:
             logger.error("-" * 80)
             logger.error(" Input file does not appear to be a PE file.")
@@ -715,16 +683,6 @@ def main(argv=None):
             logger.error(" If you don't know the input file type, you can try using the `file` utility to guess it.")
             logger.error("-" * 80)
             return -1
-        except UnsupportedRuntimeError:
-            logger.error("-" * 80)
-            logger.error(" Unsupported runtime or Python interpreter.")
-            logger.error(" ")
-            logger.error(" capa supports running under Python 2.7 using Vivisect for binary analysis.")
-            logger.error(" It can also run within IDA Pro, using either Python 2.7 or 3.5+.")
-            logger.error(" ")
-            logger.error(" If you're seeing this message on the command line, please ensure you're running Python 2.7.")
-            logger.error("-" * 80)
-            return -1
 
     meta = collect_metadata(argv, args.sample, args.rules, format, extractor)
 

capa/render/__init__.py

@@ -8,8 +8,6 @@
 
 import json
 
-import six
-
 import capa.rules
 import capa.engine
 
@@ -249,7 +247,7 @@ class CapaJsonObjectEncoder(json.JSONEncoder):
     """JSON encoder that emits Python sets as sorted lists"""
 
     def default(self, obj):
-        if isinstance(obj, (list, dict, int, float, bool, type(None))) or isinstance(obj, six.string_types):
+        if isinstance(obj, (list, dict, int, float, bool, type(None))) or isinstance(obj, str):
             return json.JSONEncoder.default(self, obj)
         elif isinstance(obj, set):
             return list(sorted(obj))

capa/render/default.py

@@ -8,7 +8,6 @@
 
 import collections
 
-import six
 import tabulate
 
 import capa.render.utils as rutils

capa/render/utils.py

@@ -6,7 +6,8 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 
-import six
+import io
+
 import termcolor
 
 
@@ -49,7 +50,7 @@ def capability_rules(doc):
         yield rule
 
 
-class StringIO(six.StringIO):
+class StringIO(io.StringIO):
     def writeln(self, s):
         self.write(s)
         self.write("\n")

capa/rules.py

@@ -18,7 +18,8 @@ try:
 except ImportError:
     from backports.functools_lru_cache import lru_cache
 
-import six
+import io
+
 import yaml
 import ruamel.yaml
 
@@ -244,7 +245,7 @@ def parse_description(s, value_type, description=None):
     """
     s can be an int or a string
     """
-    if value_type != "string" and isinstance(s, six.string_types) and DESCRIPTION_SEPARATOR in s:
+    if value_type != "string" and isinstance(s, str) and DESCRIPTION_SEPARATOR in s:
         if description:
             raise InvalidRule(
                 'unexpected value: "%s", only one description allowed (inline description with `%s`)'
@@ -256,12 +257,11 @@ def parse_description(s, value_type, description=None):
     else:
         value = s
 
-    if isinstance(value, six.string_types):
+    if isinstance(value, str):
         if value_type == "bytes":
             try:
                 value = codecs.decode(value.replace(" ", ""), "hex")
-            # TODO: Remove TypeError when Python2 is not used anymore
-            except (TypeError, binascii.Error):
+            except binascii.Error:
                 raise InvalidRule('unexpected bytes value: "%s", must be a valid hex sequence' % value)
 
             if len(value) > MAX_BYTES_FEATURE_SIZE:
@@ -406,7 +406,7 @@ def build_statements(d, scope):
             return Range(feature, min=min, max=max, description=description)
         else:
             raise InvalidRule("unexpected range: %s" % (count))
-    elif key == "string" and not isinstance(d[key], six.string_types):
+    elif key == "string" and not isinstance(d[key], str):
         raise InvalidRule("ambiguous string value %s, must be defined as explicit string" % d[key])
     else:
         Feature = parse_feature(key)
@@ -699,7 +699,7 @@ class Rule(object):
         for key in hidden_meta.keys():
             del meta[key]
 
-        ostream = six.BytesIO()
+        ostream = io.BytesIO()
         self._get_ruamel_yaml_parser().dump(definition, ostream)
 
         for key, value in hidden_meta.items():
@@ -938,7 +938,7 @@ class RuleSet(object):
         rules_filtered = set([])
         for rule in rules:
             for k, v in rule.meta.items():
-                if isinstance(v, six.string_types) and tag in v:
+                if isinstance(v, str) and tag in v:
                     logger.debug('using rule "%s" and dependencies, found tag in meta.%s: %s', rule.name, k, v)
                     rules_filtered.update(set(capa.rules.get_rules_and_dependencies(rules, rule.name)))
                     break

doc/installation.md

@@ -59,6 +59,25 @@ Use `pip` to install the source code in "editable" mode. This means that Python
 
 You'll find that the `capa.exe` (Windows) or `capa` (Linux/MacOS) executables in your path now invoke the capa binary from this directory.
 
+#### Development
+
+##### venv [optional]
+
+For development, we recommend to use [venv](https://docs.python.org/3/tutorial/venv.html). It allows you to create a virtual environment: a self-contained directory tree that contains a Python installation for a particular version of Python, plus a number of additional packages. This approach avoids conflicts between the requirements of different applications on your computer. It also ensures that you don't overlook to add a new requirement to `setup.up` using a library already installed on your system.
+
+To create an environment (in the parent directory, to avoid commiting it by accident or messing with the linters), run:
+`$ python3 -m venv ../capa-env`
+
+To activate `capa-env` in Linux or MacOS, run:
+`$ source ../capa-env/bin/activate`
+
+To activate `capa-env` in Windows, run:
+`$ ..\capa-env\Scripts\activate.bat`
+
+For more details about creating and using virtual environments, check out the [venv documentation](https://docs.python.org/3/tutorial/venv.html).
+
+##### Install development dependencies
+
 We use the following tools to ensure consistent code style and formatting:
   - [black](https://github.com/psf/black) code formatter, with `-l 120`
   - [isort 5](https://pypi.org/project/isort/) code formatter, with `--profile black --length-sort --line-width 120`
@@ -69,11 +88,21 @@ To install these development dependencies, run:
 
 `$ pip install -e /local/path/to/src[dev]`
 
-Note that some development dependencies (including the black code formatter) require Python 3.
-
 To check the code style, formatting and run the tests you can run the script `scripts/ci.sh`.
 You can run it with the argument `no_tests` to skip the tests and only run the code style and formatting: `scripts/ci.sh no_tests`
 
+##### Setup hooks [optional]
+
+If you plan to contribute to capa, you may want to setup the hooks.
+Run `scripts/setup-hooks.sh` to set the following hooks up:
+- The `pre-commit` hook runs checks before every `git commit`.
+  It runs `scripts/ci.sh no_tests` aborting the commit if there are code style or rule linter offenses you need to fix.
+- The `pre-push` hook runs checks before every `git push`.
+  It runs `scripts/ci.sh` aborting the push if there are code style or rule linter offenses or if the tests fail.
+  This way you can ensure everything is alright before sending a pull request.
+
+You can skip the checks by using the `--no-verify` git option.
+
 ### 3. Compile binary using PyInstaller
 We compile capa standalone binaries using PyInstaller. To reproduce the build process check out the source code as described above and follow these steps.
 
@@ -86,13 +115,3 @@ For Python 3: `$ pip install 'pyinstaller`
 `$ pyinstaller .github/pyinstaller/pyinstaller.spec`
 
 You can find the compiled binary in the created directory `dist/`.
-
-### 4. Setup hooks [optional]
-If you plan to contribute to capa, you may want to setup the hooks.
-Run `scripts/setup-hooks.sh` to set the following hooks up:
-- The `pre-commit` hook runs checks before every `git commit`.
-  It runs `scripts/ci.sh no_tests` aborting the commit if there are code style or rule linter offenses you need to fix.
-  You can skip this check by using the `--no-verify` git option.
-- The `pre-push` hook runs checks before every `git push`.
-  It runs `scripts/ci.sh` aborting the push if there are code style or rule linter offenses or if the tests fail.
-  This way you can ensure everything is alright before sending a pull request.

scripts/ci.sh

@@ -9,6 +9,7 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 # Use a console with emojis support for a better experience
+# Use venv to ensure that `python` calls the correct python version
 
 # Stash uncommited changes
 MSG="pre-push-$(date +%s)";
@@ -25,17 +26,8 @@ restore_stashed() {
   fi
 }
 
-python_3() {
-  case "$(uname -s)" in
-   CYGWIN*|MINGW32*|MSYS*|MINGW*)
-     py -3 -m $1 > $2 2>&1;;
-   *)
-     python3 -m $1 > $2 2>&1;;
-  esac
-}
-
 # Run isort and print state
-python_3 'isort --profile black --length-sort --line-width 120 -c .' 'isort-output.log';
+python -m isort --profile black --length-sort --line-width 120 -c . > isort-output.log 2>&1;
 if [ $? == 0 ]; then
   echo 'isort succeeded!! 💖';
 else
@@ -46,7 +38,7 @@ else
 fi
 
 # Run black and print state
-python_3 'black -l 120 --check .' 'black-output.log';
+python -m black -l 120 --check . > black-output.log 2>&1;
 if [ $? == 0 ]; then
   echo 'black succeeded!! 💝';
 else
@@ -70,7 +62,7 @@ fi
 # Run tests except if first argument is no_tests
 if [ "$1" != 'no_tests' ]; then
   echo 'Running tests, please wait ⌛';
-  pytest tests/ --maxfail=1;
+  python -m pytest tests/ --maxfail=1;
   if [ $? == 0 ]; then
     echo 'Tests succeed!! 🎉';
   else

setup.py

@@ -12,7 +12,6 @@ import sys
 import setuptools
 
 requirements = [
-    "six==1.15.0",
     "tqdm==4.60.0",
     "pyyaml==5.4.1",
     "tabulate==0.8.9",
@@ -21,24 +20,13 @@ requirements = [
     "wcwidth==0.2.5",
     "ida-settings==2.1.0",
     "viv-utils==0.6.0",
+    "halo==0.0.31",
+    "networkx==2.5.1",
+    "ruamel.yaml==0.17.0",
+    "vivisect==1.0.1",
+    "smda==1.5.13",
 ]
 
-if sys.version_info >= (3, 0):
-    # py3
-    requirements.append("halo==0.0.31")
-    requirements.append("networkx==2.5.1")
-    requirements.append("ruamel.yaml==0.17.0")
-    requirements.append("vivisect==1.0.1")
-    requirements.append("smda==1.5.13")
-else:
-    # py2
-    requirements.append("enum34==1.1.6")  # v1.1.6 is needed by halo 0.0.30 / spinners 0.0.24
-    requirements.append("halo==0.0.30")  # halo==0.0.30 is the last version to support py2.7
-    requirements.append("vivisect==0.2.1")
-    requirements.append("networkx==2.2")  # v2.2 is last version supported by Python 2.7
-    requirements.append("ruamel.yaml==0.16.13")  # last version tested with Python 2.7
-    requirements.append("backports.functools-lru-cache==1.6.1")
-
 # this sets __version__
 # via: http://stackoverflow.com/a/7071358/87207
 # and: http://stackoverflow.com/a/2073599/87207
@@ -77,13 +65,13 @@ setuptools.setup(
     install_requires=requirements,
     extras_require={
         "dev": [
-            "pytest==4.6.11",  # TODO: Change to 6.2.3 when removing py2
+            "pytest==6.2.3",
             "pytest-sugar==0.9.4",
             "pytest-instafail==0.4.2",
             "pytest-cov==2.11.1",
             "pycodestyle==2.7.0",
-            "black==20.8b1 ; python_version>'3.0'",
-            "isort==4.3.21",  # TODO: Change to 5.8.0 when removing py2
+            "black==20.8b1",
+            "isort==5.8.0",
         ]
     },
     zip_safe=False,
@@ -94,8 +82,8 @@ setuptools.setup(
         "Intended Audience :: Information Technology",
         "License :: OSI Approved :: Apache Software License",
         "Natural Language :: English",
-        "Programming Language :: Python :: 2.7",
         "Programming Language :: Python :: 3",
         "Topic :: Security",
     ],
+    python_requires=">=3.6",
 )

tests/fixtures.py

@@ -8,7 +8,6 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 import os
-import sys
 import os.path
 import binascii
 import contextlib

tests/test_freeze.py

@@ -5,7 +5,6 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-import sys
 import textwrap
 
 import pytest
@@ -174,7 +173,6 @@ def test_serialize_features():
     roundtrip_feature(capa.features.file.Import("#11"))
 
 
-@pytest.mark.xfail(sys.version_info >= (3, 0), reason="vivsect only works on py2")
 def test_freeze_sample(tmpdir, z9324d_extractor):
     # tmpdir fixture handles cleanup
     o = tmpdir.mkdir("capa").join("test.frz").strpath
@@ -182,7 +180,6 @@ def test_freeze_sample(tmpdir, z9324d_extractor):
     assert capa.features.freeze.main([path, o, "-v"]) == 0
 
 
-@pytest.mark.xfail(sys.version_info >= (3, 0), reason="vivsect only works on py2")
 def test_freeze_load_sample(tmpdir, z9324d_extractor):
     o = tmpdir.mkdir("capa").join("test.frz")
 

tests/test_helpers.py

@@ -12,8 +12,6 @@ from capa.features.extractors import helpers
 
 
 def test_all_zeros():
-    # Python 2: <str>
-    # Python 3: <bytes>
     a = b"\x00\x00\x00\x00"
     b = codecs.decode("00000000", "hex")
     c = b"\x01\x00\x00\x00"

tests/test_main.py

@@ -6,7 +6,6 @@
 # Unless required by applicable law or agreed to in writing, software distributed under the License
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
-import sys
 import json
 import textwrap
 
@@ -58,10 +57,6 @@ def test_main_single_rule(z9324d_extractor, tmpdir):
 
 
 def test_main_non_ascii_filename(pingtaest_extractor, tmpdir, capsys):
-    # on py2.7, need to be careful about str (which can hold bytes)
-    #  vs unicode (which is only unicode characters).
-    # on py3, this should not be needed.
-    #
     # here we print a string with unicode characters in it
     # (specifically, a byte string with utf-8 bytes in it, see file encoding)
     assert capa.main.main(["-q", pingtaest_extractor.path]) == 0
@@ -69,20 +64,14 @@ def test_main_non_ascii_filename(pingtaest_extractor, tmpdir, capsys):
     std = capsys.readouterr()
     # but here, we have to use a unicode instance,
     # because capsys has decoded the output for us.
-    if sys.version_info >= (3, 0):
-        assert pingtaest_extractor.path in std.out
-    else:
-        assert pingtaest_extractor.path.decode("utf-8") in std.out
+    assert pingtaest_extractor.path in std.out
 
 
 def test_main_non_ascii_filename_nonexistent(tmpdir, caplog):
     NON_ASCII_FILENAME = "täst_not_there.exe"
     assert capa.main.main(["-q", NON_ASCII_FILENAME]) == -1
 
-    if sys.version_info >= (3, 0):
-        assert NON_ASCII_FILENAME in caplog.text
-    else:
-        assert NON_ASCII_FILENAME.decode("utf-8") in caplog.text
+    assert NON_ASCII_FILENAME in caplog.text
 
 
 def test_main_shellcode(z499c2_extractor):
@@ -370,16 +359,15 @@ def test_not_render_rules_also_matched(z9324d_extractor, capsys):
 
 # It tests main works with different backends
 def test_backend_option(capsys):
-    if sys.version_info > (3, 0):
-        path = get_data_path_by_name("pma16-01")
-        assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_VIV]) == 0
-        std = capsys.readouterr()
-        std_json = json.loads(std.out)
-        assert std_json["meta"]["analysis"]["extractor"] == "VivisectFeatureExtractor"
-        assert len(std_json["rules"]) > 0
+    path = get_data_path_by_name("pma16-01")
+    assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_VIV]) == 0
+    std = capsys.readouterr()
+    std_json = json.loads(std.out)
+    assert std_json["meta"]["analysis"]["extractor"] == "VivisectFeatureExtractor"
+    assert len(std_json["rules"]) > 0
 
-        assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_SMDA]) == 0
-        std = capsys.readouterr()
-        std_json = json.loads(std.out)
-        assert std_json["meta"]["analysis"]["extractor"] == "SmdaFeatureExtractor"
-        assert len(std_json["rules"]) > 0
+    assert capa.main.main([path, "-j", "-b", capa.main.BACKEND_SMDA]) == 0
+    std = capsys.readouterr()
+    std_json = json.loads(std.out)
+    assert std_json["meta"]["analysis"]["extractor"] == "SmdaFeatureExtractor"
+    assert len(std_json["rules"]) > 0

tests/test_smda_features.py

@@ -15,7 +15,6 @@ from fixtures import *
     FEATURE_PRESENCE_TESTS,
     indirect=["sample", "scope"],
 )
-@pytest.mark.xfail(sys.version_info < (3, 0), reason="SMDA only works on py3")
 @pytest.mark.xfail(sys.platform == "win32", reason="SMDA bug: https://github.com/danielplohmann/smda/issues/20")
 def test_smda_features(sample, scope, feature, expected):
     do_test_feature_presence(get_smda_extractor, sample, scope, feature, expected)
@@ -27,5 +26,4 @@ def test_smda_features(sample, scope, feature, expected):
     indirect=["sample", "scope"],
 )
 def test_smda_feature_counts(sample, scope, feature, expected):
-    with xfail(sys.version_info < (3, 0), reason="SMDA only works on py3"):
-        do_test_feature_count(get_smda_extractor, sample, scope, feature, expected)
+    do_test_feature_count(get_smda_extractor, sample, scope, feature, expected)