Closed fiatjaf closed 11 months ago
Yes, this was exactly my idea. I will implement it.
Found an issue. A kind 1 event showed up with a single e
tag and a single a
tag.
Since the e
tag value was an unknown event, I showed the warning notice (assuming this value could have been a replaced 30023 event).
It turns out that e
tag value was a pointer to a highlight (kind 9802) that was not loaded, probably as it was outside of the current relay set.
We will likely encounter false positives like these. We need a different language than "in reply to a different version of this article" as that may simply not be true.
So someone made a highlight Y referencing an event X and someone replied to the highlight and included both X and Y as tags?
@fiatjaf Yes, exactly. The e
tag has no marker so it's difficult to say what it's referencing. The highlight eventually appeared so this stopped being a problem, but it could happen again
I'm solving it like this for now:
These events do exist and are of type 9801. Who knows wth that kind is.
Addressed and significantly improved in https://github.com/fr4nzap/zapthreads/commit/f82d58ec8c64f7a8c4b4e018687c9f8c89b358b8
Pseudocode:
Something like this (obviously just an idea):
I was trying to implement it, but I figured I would make a big mess in the code, so I decided to just open this issue instead.