TOMLDecodeError: Invalid initial character for a key part

[marxin]

I noticed tomli is stricter than toml when it comes to whitespaces:

$ cat sample.toml
digests = [
  {
    key = "value",
    key2 = "value2"
  }
]
$ In [4]: toml.load('sample.toml')
Out[4]: {'digests': [{'key': 'value', 'key2': 'value2'}]}
$ In [5]: tomli.load(open('sample.toml', 'rb'))
...
    403     return parse_one_line_basic_str(src, pos)
--> 404 raise suffixed_err(src, pos, "Invalid initial character for a key part")

TOMLDecodeError: Invalid initial character for a key part (at line 2, column 4)

Is it something that can be relaxed, please?

[hukkin]

Hi there!

Tomli implements the TOML spec. This can only be relaxed if the spec changes. Quoting https://toml.io/en/v1.0.0#inline-table

No newlines are allowed between the curly braces unless they are valid within a value.

I'm closing this in favor of toml-lang/toml#516

[marxin]

Thanks, appreciate your fast reply.

[jakirkham]

Thank you both for your discussion of this issue and raising this with the upstream spec! 🙏

It looks like upstream has resolved the issue: toml-lang/toml#516

In particular they implemented this behavior of multiline tables in the spec with PR: toml-lang/toml#904

Here is the current text covering multine tables and an associated example. Inlining a succinct version of this text below for simplicity:

Inline tables can have multiple key/value pairs on the same line, or they can be put on different lines. A terminating comma (also called trailing comma) is permitted after the last key/value pair.

...

contact = {
    personal = {
        name = "Donald Duck",
        email = "donald@duckburg.com",
    },
    work = {
        name = "Coin cleaner",
        email = "donald@ScroogeCorp.com",
    },
}

Given the upstream spec has accepted this change, would we be willing to reconsider this change?

Thank you for your help! 🙂