This question was asked a long time ago, and never got an answer, but did receive some useful comments. The short answer is there are hundreds of journals that would publish such a paper. In some fields of mathematics, 50 pages is a totally reasonable length. For example, I've seen papers of that length in Annals, Transactions, Advances, Topology, Geometry & Topology, AGT, JPAA, HHA, New York Journal of Math, Journal of Algebra, Expositiones, Applied Categorical Structures, TAC, Compositio, Math Z, Journal of K-theory, etc. The point is, in homotopy theory, papers of this length are normal.

As others have pointed out, some papers do have limits (e.g., Proceedings of the AMS only allows papers under 15 pages, and Memoirs of the AMS only allows longer papers), but they are in the minority by far. I estimate fewer than 10% of journals have a page limit.

That said, it's also true that, all things being equal, longer papers are slightly harder to publish than shorter papers, for two reasons. First, they take longer to referee. Second, they are held to a slightly higher standard by the editors, because they take up more pages in the journal. In my area, it's common for young authors to write longer papers, and eventually learn how to write shorter papers, either by splitting off pieces or by relying more on earlier work and streamlining technical arguments (rather than spelling out every single detail). Folks who want to make the literature more friendly (and really write proofs where they really spell out all the details) often write books. It's good to get into the practice of discerning when a paper can be split in two, or when a bit of work you've done is already publishable and can be submitted, even if you plan to keep investigating and follow it up with a sequel.