The current way of saving the figure as a byte string is by using None as the file name, correct? The original PR wanted to add a new method (to_bytes or write_bytes) to the fig object to make this conversion more explicit? I'm new to the community, so I would be grateful if you could give me some more guidance.

No thats backwards. None was the proposed way to do that. The current way is to make a BytesIO buffer and write to that instead of a disk buffer..

No thats backwards. None was the proposed way to do that. The current way is to make a BytesIO buffer and write to that instead of a disk buffer..

Ok, so instead of making a new method on the fig object, you're proposing that whenever None is entered as the file name, it savefig should return a bytes string. Is that correct?

Thats my understanding, but of course you should check the PR details.

@tacaswell is of the opinion that we shouldn't change the return type of the savefig method simply based on an input type or a kwarg (refer to this comment). He suggested that we could add it as a separate function rather than adding another method to Figure.

... keep reading 😉

Oh alright, @tacaswell in this comment (#6909 (comment)) said that he thinks using None as the file name should return a bytestring. I will try to re-implement what was done in the PR then, if that's alright?

@hs2361 I think that would be fine. I don't personally need this feature, but sounds like @anntzer thinks it would be useful?

@hs2361 I think that would be fine. I don't personally need this feature, but sounds like @anntzer thinks it would be useful?

@anntzer also mentioned simplifying the tests. @efiring mentioned in the comments for the PR that the tests were unnecessarily complex for the code being added in.