Benjamin Bannier

Upstream has a [PR](https://github.com/dcleblanc/SafeInt/issues/63) open to move unsafe bitfshifts from assertion to exception. If this lands soon a bump would fix our issue (would still be good to add some...

I pinged upstream on whether they could merge their fix. I also tried to customize `SAFEINT_ASSERT` like documented so we throw instead of abort, but this applies everywhere and is...

We discussed this offline and will try to implement the first approach (i.e., also considering the current field when looking for sync candidates). This will need at least a change...

> If the field is anonymous, one can't install a hook, no? Seems reasonable to to imply "cannot be accessed and imply skip" even if the unit is public, no?...

This has slight overlap with #118 which among other things also bumps node.

> Hmm - what's your ulimit -n. ```console $ ulimit -n 256 ``` > ... bumping the open filedescriptor limit to 100k seems an acceptable suggestion to avoid this. Is...

We don't have a good way to search when this last failed, but I don't see it fail recently either. Closing for now, we could reopen/recreate if it becomes an...

I reopened this since it still is an issue. In above case the test failed even during rerun where there should be much fewer noisy neighbors, but it still took...

A user reported that they see the search during recovery take considerable amount of time (80-90%) if there are a lot of gaps (here: high packet loss in Zeek). They...

Dropping this from TODO again since the ultimate issue is high packet loss, and even a more optimized implementation could eventually become too slow.