• I believe the line of stars in the Oct 21st entry in the second preface are for elision rather than a thematic break. (I might could be talked out of that, but that's what it looks like to me.) If so, then we should use a vertical elision character rather than a thematic break. The character (see 7.5.7) is ⋮ (U+22EE), and it should be centered. You can just put it in a <p> and then target it in CSS with an "nth-child" or "nth-of-type".
• We only use 2em dashes for occlusion, not for interruption (8.7.7.7.2). The ones in chapters 5 and 40 should be single em-dashes.
• The first two "Modernize" commits should be re-worded (and could be merged). The results of se modernize-spelling is doing both hyphenation and spelling (see the commit message at the end of step 14 of the Step-by-Step), so the first message should match that message. The second commit is merely the "manual" part of modernizing spelling and doesn't need a separate commit (again, per step 14). It's not a big deal that it is separate, but since it's not impacting hyphenation at all, and it is impacting spelling, the message should be "…Modernize spelling" if it stays separate. The third Modernize commit should have the same rewording as the first one.
• We have asked Alex about "worth while" multiple times, and he's always said leave it. If we ever change it in the future, we'll have to update the entire corpus, as it's all over the place. That commit should be removed or reverted.

There are several italic issues. In general, SE's approach to italics (found in the very long 8.2 section of the manual):

• We don't italicize something just because it's italicized in the source.
• If a word is in M-W's standard dictionary, we don't italicize it.
• We only use emphasis (<em>) for emphasis. If something is italicized in the source for other reasons, it may or not be italicized in SE, but if so it's not <em>.
• If it is italicized, in the vast majority of cases, it should have semantics (book, magazine, ship, non-English word, etc.). So unsemanticated italics are almost never used. There are a few explicit exceptions (sounds), and a few others that we encounter from time to time. (See below.)

Now, when we get off the tried-and-true path, i.e. what is explicitly stated in the manual, things get a little fuzzy, where the reviewers might come to different decisions, and Alex yet a different one. :)

• There are a few tagged words that are in M-W and so should not be (italicized/tagged): debris, canaille, porte(s) cochère(s).
• The porte(s) cochere(s) should also be standardized: some of them have dashes, some spaces. It doesn't have a dash in M-W, so that's the way I would lean.
• The context makes "psitt" sound like a sound, so unsemanticated italics is OK (8.2.6), but pshaw is a word, so if you think that's emphasis it should have <em>, otherwise it doesn't need italics.
• The second preface has a dateline semantic on "not he"; that should be emphasis.
• The reverse is true also; some of the emphasized words don't look like emphasis: bonjour, bonsoir, Amphitheatre (which is <q> in the preface but <em> in the book), List of Prize Winners…. If they should be italicized at all (the last is a radio program, so should be quoted, not italicized), they should probably be <i>. (Chick is another one, but I'm going to ask Alex about that one as I've run into this as well, i.e. using italics for something that should really be in quotes.) In the case of bon[js]oir, those are in M-W but labeled as French nouns, so I think it's OK to tag them as French.
• Halle Centrale is tagged, but is a proper name, which we don't italicize and only tag with a span if it's likely its pronounciation would suffer in English (8.2.9.2). Thus, if it's tagged at all, it should be with a <span>, not an <i>. And that means the other two standalone Halle's should not be italicized, either. And, in any case, none of these are emphasis.
• I don't believe we use <q> for signs, only for spoken words/thoughts. (I checked the corpus just now to make sure.) So, e.g., the sign in the second preface, the transcription in 51 and the sign in 48 should probably be <i>. (These are one of the few exceptions to the unsemanticated italics rule. Although in a lot of books these are small-caps, so these are a little unusual.)