Olivier Langlois

On Fri, 2020-03-06 at 18:08 -0800, John Lindal wrote: > It is a performance edge case. With a "normal" source file with lots > of short lines, it's fast. But...

I'm going to wrap up my comments by saying this. calling memmove() has a cost. For very small copies, that cost will be higher than the speed-up gain that the...

1. John, your test program is very convincing 2. Yes memmove is needed. Overlap is certainly possible. dest points at the start of yy_ch_buf and source points at the end....

Some more improvements in the situation. I still think that a function declaration shouldn't be in the generated header file. I insisted having it because I was worried that with...

I think that I have a similar problem with 1.8.8: 2022-09-24T19:58:08.739382120-04:00 INFO::Sapling tree verification true 2022-09-24T19:58:08.739420648-04:00 INFO::Sync finished, doing post-processing 2022-09-24T19:58:08.788534640-04:00 INFO::Sync failed, not saving: Some("task 707908 panicked") 2022-09-24T20:02:48.979656535-04:00 INFO::Sapling...

What does shutdown() do to io_uring flying requests? I have been annoyed with this phenomenon when the remote TCP endpoint is either gracefully shutdown its side or a RST segment...

I have written code to try out... ``` /* * getPayloadBuffer() */ char *bio_iouring_params::getPayloadBuffer(bidEntry &entry) noexcept { auto *bidBuf{getBidBuffer(entry.bid)}; if (!(entry.read_flags & EV_IOURING_KTLS_READ)) return bidBuf; /* * the io_uring inline...

There is still some rework to do. oset should be private to every thread. thread B could overwrite oset while thread A has the mutex.

I have been using this patch for over a month with the bitcoin client and trickle has never been that stable. I am 100% confident that this patch is ok.

svenstaro, Since I know you, you have expressed this opinion several times. So I believe you when you say that you think that it is important! I am going to...