The spec does not specify whether spaces must seperate the '=' in a key value pair.

[keyneston]

The spec makes it clear that key = "value" is legitimate but it doesn't make any claims if key="value" is legitimate.

At the very least the python library doesn't handle this case cleanly.

[laurent22]

I think white spaces before and after = should be ignored by parsers. It's a user-edited file, so it's easy to accidentally add or remove a space.

[hit9]

the whitespaces should be ignored , NOT only in this case ,but also everywhere. well ,my py parser toml.py noticed this.

[bootstraponline]

key="value" produces an error using the toml ruby gem. I think it should be valid.

[88Alex]

I think it should be valid.

It really should. Many languages (i.e. C++, C, Python, Java, among others) don't really care about whether you type key="value" or key = "value". TOML should be like that.

[88Alex]

@bootstraponline I submitted a pull request fixing that.

[parkr]

I agree: it shouldn't matter how many spaces there are in between the keys/values and the equals sign.

[BurntSushi]

I think the spec is clear on this. key=1 is valid TOML. Here's the relevant section:

Keys start with the first non-whitespace character and end with the last non-whitespace character before the equals sign.

In key=1, y is the last non-whitespace character before the equals sign. This implies that key names (but not table names) cannot have = in them. It is worded in this way so that, e.g., key =1 is also valid.