Added, thanks!
(A bit over-the-top for such a dinky little patch, but, I can hear Ace Ventura in my head, saying: "Yes. Yes. Oh, yeah. Can ya feel that, buddy? Huh? Huh? Huh?") What looked best IMO was to do the arrow and black color, but without the text-decoration:none or cursor: default.
So, style B + (unrevised) style C. Nice!
Also, while I was messing around with the topic HTML anyway, I decided to just completely remove the ad area. I don't foresee selling ads again anytime soon, and while maybe the factoids were providing some value, many of them were outdated.
I wonder if some users would appreciate the factoids (especially the older ones) being preserved/accessible for nostalgia's sake? Maybe a new endpoint (like
/factoid.php or
/index.php?action=factoid) that displays a random one each time? (I mean, I count 30
Wayback Machine snapshots of
/adrotate.php?adinfo from 2013 to 2023, so it's not strictly necessary, I suppose, but, yeah, something first-party might be nice...)
Something's wrong Yep, I noticed that, too. It's a long-standing SMF issue/quirk/design-decision, and wasn't caused by this patch (it's just easier to notice now).
It's quite complicated to explain, and most people would find the details boring AF, so, suffice it to say: it's okay if the message ID doesn't actually belong to the topic, as long as that message ID can be used to calculate the correct
offset into that topic (that is, as long as it lands you on the right topic-page, the
#new fragment takes care of putting you on the right message). A different way to think about it, is to consider what happens when you have a link to a deleted post: the message ID in the link is no longer part of that topic, but navigating to it will still land you in roughly the right place (topic-page-wise).
In the context of this particular patch, it
might have worked out nicer if SMF was more precise about message IDs in a few places, but it actually works out pretty neatly as-is, I think (that is, links with a
#new fragment don't really correspond to
one message, they correspond to potentially multiple new messages since the last one you read, so there's an argument to be made that it actually makes sense not to graphically single-out the first one).