My advisor has been sitting on revisions to a brief report for over four months now. Last February, we discovered a mistake in calculations in a published paper, that crossed the level of significance. Results that had been acceptably significant could now only be called "marginally" significant. We found the error in the process of replicating the results, so the effect is probably real, but the editor should still be informed.
First I had to argue my advisor and her co-author into agreeing that the mistake should be reported at all, on the grounds that it isn't up to us (them the original authors, me from the same lab). Then I dragged them through drafting a letter to the editor. Then my advisor decided that it should be a brief report instead, because the post-hoc explanation for why the results were only marginal was "interesting". And then she sat on the revisions for months. And then when I started pushing her to get me the draft back, she said that the delays weren't due to her new baby, but because she didn't consider the project a priority. This was right after she updated our lab's manuscripts page, with this revision right at the bottom, after even the revisions with a deadline in March.
I have said repeatedly that I want revisions by the end of December. I gave her several weeks' warning on this. I have explained that I will write my own letter to the editor if I don't get them, because her concerns about the ethics of cutting out the original authors on this are outweighed by my concerns about the ethics of going almost a year knowing that there is an error in the results without informing the editor, or anyone outside our lab.
I'm preparing to negotiate. I just can't sit back and have this always be the lowest priority, because new projects and manuscripts are always being added that will be higher priority, by virtue of having an actual deadline, but I'm not quite so anal as to demand revisions by the end of the month, as reasonable as that seemed when I first set the date. Perhaps I should be more laid-back about an error when the overall idea is probably correct, and this really should be my advisor's lowest priority. But I can't convince myself of that. It's a relatively minor ethical issue, but I can't help wondering if this is how she would react if we didn't have some reason to argue that the results didn't change that much.