Saúl Ibarra Corretgé
Saúl Ibarra Corretgé
LGTM. Left a minor suggestion, not a blocker. Great work!
New CI: https://ci.nodejs.org/view/libuv/job/libuv-test-commit/1485/
I'd say writing to a hidden file should just work. That said, regarding this approach can't we just set the flag always?
I think something along the lines of 1) would be best. We have `uv_pipe_init_ex` which. takes a `flags` argument so there is a place where this could be turned on.
That's kinda what Unix does with having most of the stuff in `stream.c`, right? Makes sense to do the same here IMHO.
I think we need this logic for (at least) Linux too: https://github.com/torvalds/linux/blob/f1baf68e1383f6ed93eb9cff2866d46562607a43/net/ipv4/tcp.c#L3313 If you call `uv_tcp_keepalive(handle, 1, 0)` on a connected socket you get `UV_EINVAL` on Linux, not so on...
Looking good! Do we know what other platforms do when trying to set the delay to 0? We now know Windows resets it and Linux errors out.
I'll try to look at what the FreeBSD kernel does shortly.
The functionality of both looks too close not to try to unify them. How about this: - `uv_net_interfaces` (or any other better name) becomes the "new" API, which lists all...
> Some thoughts about the interface for this: > > ``` > I think it's fine to skip over X25, IPX etc. and restrict to eth/ip interfaces (inb4 uv_actually_all_interfaces). >...