The important distinction here is that the results, while made public, were not peer-reviewed. From IEEE guidelines:
(...) Authors should only submit original work that has neither appeared elsewhere for publication, nor which is under review for another refereed publication.
While they don't make it unambiguously clear, the second part of this sentence is a strong suggestion that published only applies to papers that have been published as a refereed (i.e. peer-reviewed) publication. Guidelines from other publishers are similar.
So, if, as others suggest, it is not possible to add an appendix or supplementary material to your accepted publication, as others suggest, I would suggest following the Editor's advice with one small (and not very tangible) change. Publish the full results, not as a pre-print, but as a technical report under a different title. Remove the old pre-print and have only a matching pre-print sharing the title of your accepted paper. Refer to your technical report as you would any other work, including the URL in the citation (as this is a "web-resource", and not peer-reviewed).
Since you may not resubmit the work only if it has been published as a refereed publication, which your technical report will not be, you will still be free to re-use those results when trying to publish in a refereed journal or conference. You should, in that case, still cite your technical report. This not only encourages consistency, but additionally, publishing all the results as a technical report also "time-stamps your name" on it, removing the possibility (or at least, giving you a well-documented trail) of your work or results being scooped.
This practice is something I've occasionally seen in my field. The last example I've seen was of a group that developed a new theory, and then also applied that theory to a number specific problems that were still of fairly wide interest to the community. The theory, with selected examples that validate it and demonstrate it were published in a high-ranking journal. The application of the theory to several specific problems was "published" only as a technical report (on their university/institute web pages) which was cited in the journal paper. It would also allow the authors to reuse any of those results (from the technical report) in their subsequent publications, e.g. they might want to more deeply explore the implications of, or process the, results of applying the technique to one of the specific problems covered in the technical report, in which case it would be perfectly fine to include those results (previously published in a technical report) in such a new paper.