mixed types in array don't work

[vanyakosmos]

[config]
list = [
  { key = "value"},
  "foo",
]

this way file will be parsed but "foo" will be ignored: {config: {list: [{key: value}]}}

---

[config]
list = [
  "foo",
  { key = "value"},
]

After parsing toml file above parser will throw an error: IndexError: list index out of range.
Traceback:

  File "/path/to/lib/python3.7/site-packages/toml/decoder.py", line 456, in loads
    multibackslash)
  File "/path/to/lib/python3.7/site-packages/toml/decoder.py", line 725, in load_line
    value, vtype = self.load_value(pair[1], strictly_valid)
  File "/path/to/lib/python3.7/site-packages/toml/decoder.py", line 802, in load_value
    return (self.load_array(v), "array")
  File "/path/to/lib/python3.7/site-packages/toml/decoder.py", line 914, in load_array
    a[b] = a[b] + ',' + a[b + 1]
IndexError: list index out of range

[flokno]

This was only recently added to toml itself: toml-lang/toml#676

Edit: Wait, I think 2 different issues are going on here. The first is that indeed

[config]
list = [
  { key = "value"},
  "foo",
]

is parsed to 'config': {'list': [{'key': 'value'}]}} dropping the foo.

[config]
list = [
  "foo",
  { key = "value"},
]

throws the IndexError, but that is to be expected I think because of the trailing comma. The question is: Why is the unmatched trailing comma not detected in the first example?

Finally,

[config]
list = [
  "foo",
  { key = "value"}
]

i.e., without the trailing comma, raises toml.decoder.TomlDecodeError: Not a homogeneous array, which is not necessary anymore according to toml-lang/toml#676. I guess that is what the issue was relating to in the first place.

[dalonsoa]

I can still reproduce this error and, indeed, the result is rather inconsisten: it can dump arrays with mixed types, but it fails when loading that same output.

To reproduce:

• Python Python 3.8.3
• toml 0.10.1

import toml
data = {"a": [1, "cat"]}
data_encoded = toml.dumps(data)

The above works and results in data_encoded equal to 'a = [ 1, "cat",]\n' . Then:

data_back = toml.loads(data_encoded)

results in:

Traceback (most recent call last):
  File "XXXXXXX/python3.8/site-packages/toml/decoder.py", line 510, in loads
    ret = decoder.load_line(line, currentlevel, multikey,
  File "XXXXXXX/python3.8/site-packages/toml/decoder.py", line 777, in load_line
    value, vtype = self.load_value(pair[1], strictly_valid)
  File "XXXXXXX/python3.8/site-packages/toml/decoder.py", line 875, in load_value
    return (self.load_array(v), "array")
  File "XXXXXXX/python3.8/site-packages/toml/decoder.py", line 1024, in load_array
    raise ValueError("Not a homogeneous array")
ValueError: Not a homogeneous array
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "XXXXXXX/python3.8/site-packages/IPython/core/interactiveshell.py", line 3343, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-7-8f12f46e5bec>", line 1, in <module>
    data_back = toml.loads(data_encoded)
  File "XXXXXXX/python3.8/site-packages/toml/decoder.py", line 513, in loads
    raise TomlDecodeError(str(err), original, pos)
toml.decoder.TomlDecodeError: Not a homogeneous array (line 1 column 1 char 0)

The same applies when loading/dumping to a file rather than a string.

[arnauddevie]

Reproduced this error with mixed integer/float array

Python: 3.7.5
toml: 0.10.0, 0.10.1

Dump

import toml

mixed_types = [0, 1.1, 2.2, 3.3] # mixed integer and floats

toml.dumps({'array': mixed_types})

'array = [ 0, 1.1, 2.2, 3.3,]\n'

Load

toml.loads(toml.dumps({'array': mixed_types}))

---

ValueError Traceback (most recent call last)
~\Anaconda3\envs\modeling\lib\site-packages\toml\decoder.py in loads(s, _dict, decoder)
455 ret = decoder.load_line(line, currentlevel, multikey,
--> 456 multibackslash)
457 except ValueError as err:

~\Anaconda3\envs\modeling\lib\site-packages\toml\decoder.py in load_line(self, line, currentlevel, multikey, multibackslash)
724 else:
--> 725 value, vtype = self.load_value(pair[1], strictly_valid)
726 try:

~\Anaconda3\envs\modeling\lib\site-packages\toml\decoder.py in load_value(self, v, strictly_valid)
801 elif v[0] == '[':
--> 802 return (self.load_array(v), "array")
803 elif v[0] == '{':

~\Anaconda3\envs\modeling\lib\site-packages\toml\decoder.py in load_array(self, a)
940 if ntype != atype:
--> 941 raise ValueError("Not a homogeneous array")
942 else:

ValueError: Not a homogeneous array

During handling of the above exception, another exception occurred:

TomlDecodeError Traceback (most recent call last)
in
----> 1 toml.loads(toml.dumps({'array': mixed_types}))

~\Anaconda3\envs\modeling\lib\site-packages\toml\decoder.py in loads(s, _dict, decoder)
456 multibackslash)
457 except ValueError as err:
--> 458 raise TomlDecodeError(str(err), original, pos)
459 if ret is not None:
460 multikey, multilinestr, multibackslash = ret

TomlDecodeError: Not a homogeneous array (line 1 column 1 char 0)

Spec v1.0.0-rc.2 says:

Array

Arrays are square brackets with values inside. Whitespace is ignored. Elements are separated by commas. Arrays can contain values of the same data types as allowed in key/value pairs. Values of different types may be mixed.

# Mixed-type arrays are allowed
numbers = [ 0.1, 0.2, 0.5, 1, 2, 5 ]
contributors = [
  "Foo Bar <[email protected]>",
  { name = "Baz Qux", email = "[email protected]", url = "https://example.com/bazqux" }
]

[cykerway]

Still a problem on Python 3.9.9 with toml 0.10.2:

arr = [ 'a', 0 ]

Error:

toml.decoder.TomlDecodeError: Not a homogeneous array ...

Doc (https://toml.io/en/v1.0.0#array):

Values of different types may be mixed.

[RichardBruskiewich]

Still a problem 3 years later... See python-poetry/poetry#7094

[zimonitrome]

Yup like cykerway said, this should be allowed: https://toml.io/en/v1.0.0#array:~:text=%23%20Mixed%2Dtype%20arrays%20are%20allowed%0Anumbers%20%3D%20%5B%200.1%2C%200.2%2C%200.5%2C%201%2C%202%2C%205%20%5D

Still a problem 5 years later.

EDIT: Actually, this has been fixed in the master branch: 6d51c0a

But no new version of the package has been released since 2020. And the maintainer seems dead.

Maybe @arp242, who made the commit, knows anything about the pypi status?