Interesting! Thanks for reporting. Looks like these were sent to your replies in those threads as well as to your original post. That's why they're missing the original post's target URL, like you mention. I think the separate webmentions to the original post did go through though. For example, here are the corresponding webmentions for the last two you mention, these source pages do contain the davidcel.is target URL:

• https://brid.gy/comment/mastodon/@[email protected]/109776142147798890/109778842648443627
• https://brid.gy/comment/mastodon/@[email protected]/109776142147798890/109776826643749474

So, sounds like you hopefully got valid wms for every reply, and these were invalid extras. I'll keep this open to track fixing and not sending those!

Ohh snap! Yup, you seem to be right; I do think I got all of the proper mentions and now I'm remembering that I thought I saw something duplicated on my end before tracking these few failures. Thank you for helping me look into it and for the future fix!

@snarfed Hey again! I just noticed some failed webmentions on my end that follow the same pattern, where the target URL wasn't included in the source HTML. One example:

https://brid.gy/comment/mastodon/@[email protected]/112073199019621267/112074944454786511

In this case, I don't have any other webmentions I've received on my end that were a duplicate but which were valid. This was the only one I received for the given interaction, but I ended up processing it as invalid because of the URL not being included in the source. I think I may have noticed something similar with another post of mine after a thread of replies got deep enough, so I'm wondering if there's an issue that crops up once nested replies reach a certain depth. In this case, one person and I had a back-and-forth and, after a depth of four replies to my original post, I stopped getting successful webmentions for anything else in that thread.

Hmm! Interesting. Sorry for the trouble! We don't have an explicit depth limit; we use https://docs.joinmastodon.org/methods/statuses/#context (authenticated) to get Mastodon replies, and this is far below its threshold, so I'm not sure what's going on yet.

No worries! There probably isn't harm in me just skipping the target URL verification step for webmentions coming from brid.gy for now, but I'll be curious to see if you're able to figure out why the original post's URL is missing. Thanks for taking a look!