>>28134

Glad everything works again.

I'm now even archiving all notable files, even when it's small pictures. My archiver is trying to catch up, but has tons of things to do.

btw. it seems qresear.ch anon got doxxed, and only after a day or so the thread was removed.

I'm also not sure about qmap.pub. That site had misleading titles for Qposts all over. That always looked fishy to me.

Can't you parse threads in a way that threads that are marked for manual review are skipped at first, so that at least other breads are shown?

>>28193

Is it such a problem if there is a duplicate thread, or a thread is missing?

My current code is ignoring bread numbers. It simply marks posts independently as notables (but remembers which thread marked them as such) and when I get a bun later via GUI, the code is actually selecting the marked posts instead of getting the notable list.

I mean let's assume a shill creates a fake bread, and we capture notables from there, this may still happen even when you try to verify everything, because there is no way to verify that a real baker created the bread (because tripcodes are disabled for everyone except Q and maybe BO).

And I mean just in the last few days we had a few "Q Research #"-breads, instead of calling them "Q Research General #" and that was the fault of real bakers. I think it's more likely that an actual baker simply made an error.

and I assume that fake breads are removed within a few minutes or hours.

>>28243

a way to get around such fake breads would be to delay parsing threads until the notable list is at least let's say 4 hours old. That way if something invalid goes through, it's the fault of BO / BVs.

Over the last months I have never really seen that happening (of course maybe I just missed it).

>>28244

another method would be using a small delay like 30-60 minutes, at the same time checking for XX hours that this thread stays up, and if it's deleted removing those notables.

That way it would run automatically and through the actions of BO/BVs it could even fix itself.

>>28246

I would (and I do) use the thread number, not the bread # for saving notables.

I'm using the bread title, but only for viewing.

btw. just this bread right now has a wrong # once again.

>>28267

>The problem with waiting for BO/BV actions is the delay it would introduce

As I said, you wouldn't have to.

You could simply recheck if the bread you took it from gets deleted or not and use it straight away. That would be a good idea regardless, because who knows what could happen.

Oh and I actually got a reply from isitwetyet, weeks after I sent the email. They replied they would take care of the JSON data issues, and for the last 1-2 weeks I haven't seen anything like that happening again.