sly-eval-print: Don't hardcode comment size

[aartaka]

I'm often using sly-eval-print for a kind of literate programming workflow of evaluating forms one after the other and printing results inline. One particular thing that makes this workflow uncomfortable is that sly-eval-print prints the result with a single colon. This single colon is:

• Often improperly indented by SLY itself.
• Is against the convention that one-semicolon comments are strictly the inline ones.
  • Block comments (two semicolons) make more sense for result printing with line break.
• And then it's hard-coded, which makes it much less customizable by user. comment-region has a lot of customizations, so the result of sly-eval-print might be really far from what one might expect in their code comments.

This PR removes the hard-coded prefix arg, so that sly-eval-print comments are more customizable, consistent, and semantically correct.