Do not output attribute information when unparsing TermLike

[geo2a]

When pretty-printing a TermLike, the term attributes are serialized into text and in-lined into the unparsed Kore term, leading to results like this one:

    Could not infer the equation requirement:
        /* Spa */
        \equals{SortBool{}, _}(
            /* T Fn D Spa */
            Lbl'Unds-LT-'Int'Unds'{}(
                /* T Fn D Spa */
                Lblf'LParUndsRParUnds'TEST'Unds'Int'Unds'Int{}(
                    /* T Fn D Sfa */ ConfigVarI:SortInt{}
                ),
                /* T Fn D Sfa Cl */ \dv{SortInt{}}("0")
            ),
            /* T Fn D Sfa Cl */ \dv{SortBool{}}("true")
        )

I'd argue that we do not care as much about the exact attributes of TermLikes anymore today. Therefore I propose to remove the attributes, leading to a much cleaner log output:

    Could not infer the equation requirement:
        \equals{SortBool{}, _}(
            Lbl'Unds-LT-'Int'Unds'{}(
                Lblf'LParUndsRParUnds'TEST'Unds'Int'Unds'Int{}(
                    ConfigVarI:SortInt{}
                ),
                \dv{SortInt{}}("0")
            ),
            \dv{SortBool{}}("true")
        )

The integration tests fail, but I will fix them once we have decided if we want this merged.

[goodlyrottenapple]

Looks good, tho you will have to likely update a lot of integration golden files

[jberthold]

Looks good, tho you will have to likely update a lot of integration golden files

➕ Yes. As suggested on slack:
13 hours ago
Maybe sed -i -e 's,/\*[^*]*\*/,,g' -e '/^ *$/d' <EXPECTED_OUTPUT_FILES> in the integration tests?