Fix duplicate keys preservation of JSON data (#1163)

* Fix duplicate keys preservation of JSON data

* Update issue number

* Fix type annotations

* Changes after review

* Rewording

CHANGELOG.md

@@ -5,6 +5,7 @@ This project adheres to [Semantic Versioning](https://semver.org/).
 
 ## [2.6.0.dev0](https://github.com/httpie/httpie/compare/2.5.0...master) (unreleased)
 
+- Fixed duplicate keys preservation of JSON data. ([#1163](https://github.com/httpie/httpie/issues/1163))
 - Added support for formatting & coloring of JSON bodies preceded by non-JSON data (e.g., an XXSI prefix). ([#1130](https://github.com/httpie/httpie/issues/1130))
 
 ## [2.5.0](https://github.com/httpie/httpie/compare/2.4.0...2.5.0) (2021-09-06)

httpie/cli/requestitems.py

@@ -15,7 +15,7 @@ from .dicts import (
     RequestQueryParamsDict,
 )
 from .exceptions import ParseError
-from ..utils import get_content_type, load_json_preserve_order
+from ..utils import get_content_type, load_json_preserve_order_and_dupe_keys
 
 
 class RequestItems:
@@ -150,6 +150,6 @@ def load_text_file(item: KeyValueArg) -> str:
 
 def load_json(arg: KeyValueArg, contents: str) -> JSONType:
     try:
-        return load_json_preserve_order(contents)
+        return load_json_preserve_order_and_dupe_keys(contents)
     except ValueError as e:
         raise ParseError(f'{arg.orig!r}: {e}')

httpie/output/utils.py

@@ -2,6 +2,7 @@ import json
 import re
 from typing import Tuple
 
+from ..utils import load_json_preserve_order_and_dupe_keys
 from .lexers.json import PREFIX_REGEX
 
 
@@ -11,14 +12,14 @@ def load_prefixed_json(data: str) -> Tuple[str, json.JSONDecoder]:
     """
     # First, the full data.
     try:
-        return '', json.loads(data)
+        return '', load_json_preserve_order_and_dupe_keys(data)
     except ValueError:
         pass
 
     # Then, try to find the start of the actual body.
     data_prefix, body = parse_prefixed_json(data)
     try:
-        return data_prefix, json.loads(body)
+        return data_prefix, load_json_preserve_order_and_dupe_keys(body)
     except ValueError:
         raise ValueError('Invalid JSON')
 

httpie/utils.py

@@ -1,19 +1,69 @@
 import json
 import mimetypes
+import re
+import sys
 import time
 from collections import OrderedDict
 from http.cookiejar import parse_ns_headers
 from pprint import pformat
-from typing import List, Optional, Tuple
+from typing import Any, List, Optional, Tuple
-import re
 
 import requests.auth
 
 RE_COOKIE_SPLIT = re.compile(r', (?=[^ ;]+=)')
+Item = Tuple[str, Any]
+Items = List[Item]
 
 
-def load_json_preserve_order(s):
+class JsonDictPreservingDuplicateKeys(OrderedDict):
-    return json.loads(s, object_pairs_hook=OrderedDict)
+    """A specialized JSON dict preserving duplicate keys.
+
+    """
+
+    # Python versions prior to 3.8 suffer from an issue with multiple keys with the same name.
+    # `json.dumps(obj, indent=N, sort_keys=True)` will output sorted keys when they are unique, and
+    # duplicate keys will be outputted as they were defined in the original data.
+    # See <https://bugs.python.org/issue23493#msg400929> for the behavior change between Python versions.
+    SUPPORTS_SORTING = sys.version_info >= (3, 8)
+
+    def __init__(self, items: Items):
+        self._items = items
+        self._ensure_items_used()
+
+    def _ensure_items_used(self) -> None:
+        """HACK: Force `json.dumps()` to use `self.items()` instead of an empty dict.
+
+        Two JSON encoders are available on CPython: pure-Python (1) and C (2) implementations.
+
+        (1) The pure-python implementation will do a simple `if not dict: return '{}'`,
+        and we could fake that check by implementing the `__bool__()` method.
+        Source:
+            - <https://github.com/python/cpython/blob/9d318ad/Lib/json/encoder.py#L334-L336>
+
+        (2) On the other hand, the C implementation will do a check on the number of
+        items contained inside the dict, using a verification on `dict->ma_used`, which
+        is updated only when an item is added/removed from the dict. For that case,
+        there is no workaround but to add an item into the dict.
+        Sources:
+            - <https://github.com/python/cpython/blob/9d318ad/Modules/_json.c#L1581-L1582>
+            - <https://github.com/python/cpython/blob/9d318ad/Include/cpython/dictobject.h#L53>
+            - <https://github.com/python/cpython/blob/9d318ad/Include/cpython/dictobject.h#L17-L18>
+
+        To please both implementations, we simply add one item to the dict.
+
+        """
+        if self._items:
+            self['__hack__'] = '__hack__'
+
+    def items(self) -> Items:
+        """Return all items, duplicate ones included.
+
+        """
+        return self._items
+
+
+def load_json_preserve_order_and_dupe_keys(s):
+    return json.loads(s, object_pairs_hook=JsonDictPreservingDuplicateKeys)
 
 
 def repr_dict(d: dict) -> str:

tests/fixtures/__init__.py

@@ -16,6 +16,7 @@ def patharg(path):
 FIXTURES_ROOT = Path(__file__).parent
 FILE_PATH = FIXTURES_ROOT / 'test.txt'
 JSON_FILE_PATH = FIXTURES_ROOT / 'test.json'
+JSON_WITH_DUPE_KEYS_FILE_PATH = FIXTURES_ROOT / 'test_with_dupe_keys.json'
 BIN_FILE_PATH = FIXTURES_ROOT / 'test.bin'
 XML_FILES_PATH = FIXTURES_ROOT / 'xmldata'
 XML_FILES_VALID = list((XML_FILES_PATH / 'valid').glob('*_raw.xml'))

tests/fixtures/test_with_dupe_keys.json

@@ -0,0 +1 @@
+{"key":15,"key":15,"key":3,"key":7}

tests/test_cli.py

@@ -1,20 +1,21 @@
 """CLI argument parsing related tests."""
 import argparse
-import json
 
 import pytest
 from requests.exceptions import InvalidSchema
 
 import httpie.cli.argparser
-from .fixtures import (
-    FILE_CONTENT, FILE_PATH, FILE_PATH_ARG, JSON_FILE_CONTENT,
-    JSON_FILE_PATH_ARG,
-)
-from httpie.status import ExitStatus
 from httpie.cli import constants
 from httpie.cli.definition import parser
 from httpie.cli.argtypes import KeyValueArg, KeyValueArgType
 from httpie.cli.requestitems import RequestItems
+from httpie.status import ExitStatus
+from httpie.utils import load_json_preserve_order_and_dupe_keys
+
+from .fixtures import (
+    FILE_CONTENT, FILE_PATH, FILE_PATH_ARG, JSON_FILE_CONTENT,
+    JSON_FILE_PATH_ARG,
+)
 from .utils import HTTP_OK, MockEnvironment, StdinBytesIO, http
 
 
@@ -97,17 +98,15 @@ class TestItemParsing:
 
         # Parsed data
         raw_json_embed = items.data.pop('raw-json-embed')
-        assert raw_json_embed == json.loads(JSON_FILE_CONTENT)
+        assert raw_json_embed == load_json_preserve_order_and_dupe_keys(JSON_FILE_CONTENT)
         items.data['string-embed'] = items.data['string-embed'].strip()
         assert dict(items.data) == {
-            "ed": "",
+            'ed': '',
-            "string": "value",
+            'string': 'value',
-            "bool": True,
+            'bool': True,
-            "list": ["a", 1, {}, False],
+            'list': ['a', 1, {}, False],
-            "obj": {
+            'obj': load_json_preserve_order_and_dupe_keys('{"a": "b"}'),
-                "a": "b"
+            'string-embed': FILE_CONTENT,
-            },
-            "string-embed": FILE_CONTENT,
         }
 
         # Parsed query string parameters

tests/test_json.py

@@ -6,13 +6,29 @@ import responses
 from httpie.cli.constants import PRETTY_MAP
 from httpie.compat import is_windows
 from httpie.output.formatters.colors import ColorFormatter
+from httpie.utils import JsonDictPreservingDuplicateKeys
 
+from .fixtures import JSON_WITH_DUPE_KEYS_FILE_PATH
 from .utils import MockEnvironment, http, URL_EXAMPLE
 
 TEST_JSON_XXSI_PREFIXES = (r")]}',\n", ")]}',", 'while(1);', 'for(;;)', ')', ']', '}')
 TEST_JSON_VALUES = ({}, {'a': 0, 'b': 0}, [], ['a', 'b'], 'foo', True, False, None)  # FIX: missing int & float
 TEST_PREFIX_TOKEN_COLOR = '\x1b[38;5;15m' if is_windows else '\x1b[04m\x1b[91m'
 
+JSON_WITH_DUPES_RAW = '{"key": 15, "key": 15, "key": 3, "key": 7}'
+JSON_WITH_DUPES_FORMATTED_SORTED = '''{
+    "key": 3,
+    "key": 7,
+    "key": 15,
+    "key": 15
+}'''
+JSON_WITH_DUPES_FORMATTED_UNSORTED = '''{
+    "key": 15,
+    "key": 15,
+    "key": 3,
+    "key": 7
+}'''
+
 
 @pytest.mark.parametrize('data_prefix', TEST_JSON_XXSI_PREFIXES)
 @pytest.mark.parametrize('json_data', TEST_JSON_VALUES)
@@ -38,3 +54,35 @@ def test_json_formatter_with_body_preceded_by_non_json_data(data_prefix, json_da
         # meaning it was correctly handled as a whole.
         assert TEST_PREFIX_TOKEN_COLOR + data_prefix in expected_body, expected_body
     assert expected_body in r
+
+
+@responses.activate
+def test_duplicate_keys_support_from_response():
+    """JSON with duplicate keys should be handled correctly."""
+    responses.add(responses.GET, URL_EXAMPLE, body=JSON_WITH_DUPES_RAW,
+                  content_type='application/json')
+    args = ('--pretty', 'format', URL_EXAMPLE)
+
+    # Check implicit --sorted
+    if JsonDictPreservingDuplicateKeys.SUPPORTS_SORTING:
+        r = http(*args)
+        assert JSON_WITH_DUPES_FORMATTED_SORTED in r
+
+    # Check --unsorted
+    r = http(*args, '--unsorted')
+    assert JSON_WITH_DUPES_FORMATTED_UNSORTED in r
+
+
+def test_duplicate_keys_support_from_input_file():
+    """JSON file with duplicate keys should be handled correctly."""
+    args = ('--verbose', '--offline', URL_EXAMPLE,
+            f'@{JSON_WITH_DUPE_KEYS_FILE_PATH}')
+
+    # Check implicit --sorted
+    if JsonDictPreservingDuplicateKeys.SUPPORTS_SORTING:
+        r = http(*args)
+        assert JSON_WITH_DUPES_FORMATTED_SORTED in r
+
+    # Check --unsorted
+    r = http(*args, '--unsorted')
+    assert JSON_WITH_DUPES_FORMATTED_UNSORTED in r