Bronek Kozicki

> No, you don't need to do anything special about running out of memory. If rippled runs out of memory it bursts into flame. There's no special handling for low...

@ximinez do you mind checking again ? I added `throw std::bad_alloc()` where appropriate. There are no other error paths that I can see in in xxhash sources.

> Before I dive into the code, can you please share in an English comment (a) what problem this is meant to solve and (b) the general design / intuition...

@intelliot After 2.1 release is fine.

from the coverage report: > [src/ripple/app/misc/NetworkOPs.cpp](https://app.codecov.io/gh/XRPLF/rippled/pull/4833?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=XRPLF#diff-c3JjL3JpcHBsZS9hcHAvbWlzYy9OZXR3b3JrT1BzLmNwcA==) * old tests of that same part, which do not have (and did not have) full coverage > [src/ripple/protocol/MultiApiJson.h](https://app.codecov.io/gh/XRPLF/rippled/pull/4833?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=XRPLF#diff-c3JjL3JpcHBsZS9wcm90b2NvbC9NdWx0aUFwaUpzb24uaA==) * two `asserts` in new...

Indeed; as I was looking to answer this issue I found that this is enforced in preflight, so the check is very cheap to do. However I also noticed that...

Test added in https://github.com/XRPLF/rippled/pull/4860

> > You cannot send a Payment transaction to yourself. > > I think you can send a payment transaction to yourself as long as it crosses offers (currency conversion)...

No changes in regard to this issue are planned in the future.

It just occurred to me that, since Apr 2022 when the buggy code was deployed, we were more-or-less performing full book update every time (instead of the intended every 25600...