I’d actually wager the comments are cached, sent to the front end wrong (because of the bad cache), and then the front end posts against the wrong comment ID (maybe that’s what you mean to be fair :) ).
I had something different in mind, coming from Angular: There would be a list of comment objects associated with DOM nodes, then the comment list would get updated, and Angular would associate the DOM nodes with the wrong list entries.
How would a bad cache mess up the association between a comment and its ID?
I used to do AngularJS and I’ve done some react… maybe something like that could happen. I’d wager it’s unlikely though (bordering on Angular/Inferno itself having a bug).
I’ve seen some other things that seem like caching issues (e.g., seeing the wrong counts when switching between posts).
A cache could literally report the wrong ID for a comment to the front end in the JSON if the caching isn’t right (and bad input = bad output).
Granted, in both cases I’d wonder why we’re not seeing this all the time, it’s got to be something niche, possible something already fixed but not on all instances.
I’d actually wager the comments are cached, sent to the front end wrong (because of the bad cache), and then the front end posts against the wrong comment ID (maybe that’s what you mean to be fair :) ).
I had something different in mind, coming from Angular: There would be a list of comment objects associated with DOM nodes, then the comment list would get updated, and Angular would associate the DOM nodes with the wrong list entries.
How would a bad cache mess up the association between a comment and its ID?
I used to do AngularJS and I’ve done some react… maybe something like that could happen. I’d wager it’s unlikely though (bordering on Angular/Inferno itself having a bug).
I’ve seen some other things that seem like caching issues (e.g., seeing the wrong counts when switching between posts).
A cache could literally report the wrong ID for a comment to the front end in the JSON if the caching isn’t right (and bad input = bad output).
Granted, in both cases I’d wonder why we’re not seeing this all the time, it’s got to be something niche, possible something already fixed but not on all instances.