fix exponentation of tropical polynomials

[YueRen]

Solves a bug reported by @shelbycox via slack.

[thofma]

Do you have a reference for the characteristic of semirings? (Other than a stackexchange link)

[YueRen]

Do you have a reference for the characteristic of semirings? (Other than a stackexchange link)

Not at hand, but I'll ask around.

[YueRen]

It seems that I was mistaken. The correct characteristic of the tropical numbers (or any idempotent semiring) is 1. Maybe it's better to remove that function then, in case some other Oscar function implicitly assumes that characteristic is either 0 or prime.

[afkafkafk13]

It seems that I was mistaken. The correct characteristic of the tropical numbers (or any idempotent semiring) is 1. Maybe it's better to remove that function then, in case some other Oscar function implicitly assumes that characteristic is either 0 or prime.

It might also be useful to place a meaningful error pointing out that the characteristic of the tropical semiring is not available by choice and not by chance (as it is always 1 and might confuse ring functionality).

[afkafkafk13]

Looks fine now.

[YueRen]

It seems that I was mistaken. The correct characteristic of the tropical numbers (or any idempotent semiring) is 1. Maybe it's better to remove that function then, in case some other Oscar function implicitly assumes that characteristic is either 0 or prime.

It might also be useful to place a meaningful error pointing out that the characteristic of the tropical semiring is not available by choice and not by chance (as it is always 1 and might confuse ring functionality).

I have added an error that the characteristic of tropical semirings is not supported. I've thought of ways to make the error more verbose, but what elaboration makes sense strongly depends on how the user encounters the error.

[afkafkafk13]

It seems that I was mistaken. The correct characteristic of the tropical numbers (or any idempotent semiring) is 1. Maybe it's better to remove that function then, in case some other Oscar function implicitly assumes that characteristic is either 0 or prime.

It might also be useful to place a meaningful error pointing out that the characteristic of the tropical semiring is not available by choice and not by chance (as it is always 1 and might confuse ring functionality).

I have added an error that the characteristic of tropical semirings is not supported. I've thought of ways to make the error more verbose, but what elaboration makes sense strongly depends on how the user encounters the error.

I alreay saw it and approved it. The degree of elaboration in the error message is fine.