I've written an article which is 32 pages long, (including 1.5 page of references and 3 pages of appendices). I intend to submit it to a conference which accepts 20-page long articles including references and excluding appendices, with at most an additional 10 pages for appendices.
The obvious way to proceed is to move some parts of the original article to the appendices. Indeed, some proofs are very similar to proofs published in former articles (which are of course, cited throughout the proofs). Thus, it's no big deal if they were to be moved to the appendices, since the original proofs are, to my opinion, what makes the article interesting.
However, the article will still be 32 pages long. A possible fix is to simply remove the previous appendix of the article. This appendix is a proof written in a previous article (which I'm not affiliated with in any way). The goal of the appendix is to show that this proof can still be applied to the more general context the article deals with. The proof is essentially the same: at the exception of some trivial tricks in the equations, the only thing that differs between this proof and the original one are the equations.
However, I'm quite embarrassed to simply say "trust me, do the same proof and it'll work". A solution I've thought about, since the article is already an eprint available on Internet, is to cite this eprint in the conference version of the article. Something along the lines of "This proof can be found in the full version of this article [1]".
Is it ethical, and likely to be accepted by the editors?