Parsing fails on multiline string with comments and newlines

[ghost]

I can't put my finger on this exactly, but it seems that the parser gets confused by certain combinations of multiline strings with comments.

For example:

toml.loads('a = """#\n"""#')

returns {}
while it should return {"a": "#\n"}

May be related to #152.

[uiri]

Thank you for the bug report!

Your example appears to be equivalent to the following toml document:

a = """#
"""#

Is this correct?

This appears to be an issue of combining # and newlines in the same multiline string.

[ghost]

Yes I agree. That is the TOML document I'm talking about.

Indeed it looks like certain combinations do not work well inside the multiline string.
Note however that the second comment (immediately after the multiline string) is also necessary to trigger the problem. That puzzles me.

[rudy-6-4]

This issues is due to '# ' or any chracter (e.g. space )just after """. The check for """ is done on the non stripped version of the line.
'#' is replaced by a space during a preprocessing step and later the space is not removed when operating in multistring mode.