Release notes: cover the non-blocking tls-write change and the fd leak

Changed

 RELEASE_NOTES_v0.16.5.md | 23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)

Diff

RELEASE_NOTES_v0.16.5.mdmodified

@@ -28,10 +28,25 @@ two new procedures report why a connection attempt failed.

and then went silent; `"handshake-failed"` means the peer refused us. Those

are different facts and callers retry them differently, which a single `#f`

cannot express.

−31

- Fixed a file descriptor leak on the `tls-upgrade` failure path. The socket

−32

object is marked closed before the handshake runs, so nothing else owned

−33

that descriptor. It was hard to reach before; a bounded handshake makes

−34

timeouts an ordinary outcome and the leak reachable.

+31

- `tls-write` on a non-blocking connection returns the bytes it actually

+32

committed, possibly zero, instead of spinning. It previously retried

+33

`WANT_WRITE` in a hot loop that never returned and never yielded, so a peer

+34

that stopped reading could pin a CPU indefinitely and no caller could bound

+35

a write at all. Blocking connections are unchanged. `tls-write` also gains

+36

optional start and end byte offsets, mirroring `socket-write`, so a write

+37

loop resending the tail of a large body stays O(n) rather than O(n squared).

+38

+39

## Fixed

+40

+41

- A file descriptor leak on the `tls-upgrade` failure path. The socket object

+42

is marked closed before the handshake runs, so nothing else owned that

+43

descriptor and nothing would ever close it.

+44

+45

This one matters more than its size suggests. The leak was hard to reach

+46

while handshake failures were rare, and bounding the handshake makes

+47

timeouts an ordinary outcome. Shipping the bound without this would have

+48

traded a hang for descriptor exhaustion weeks later, by which time the

+49

connection between the two would have been invisible.

## Upgrade notes