This reverts commit b141c162ee.
Reason for revert: Breaking WebRTC rolls. See https://ci.chromium.org/ui/b/8832847811929676465 for an example failed build.
Original change's description:
> Take out listen support from AsyncPacketSocket
>
> Moved to new interface class AsyncListenSocket.
>
> Bug: webrtc:13065
> Change-Id: Ib96ce154ba19979360ecd8144981d947ff5b8b18
> Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/232607
> Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
> Reviewed-by: Harald Alvestrand <hta@webrtc.org>
> Commit-Queue: Niels Moller <nisse@webrtc.org>
> Cr-Commit-Position: refs/heads/main@{#35234}
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: webrtc:13065
Change-Id: Id5d5b35cb21704ca4e3006caf1636906df062609
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/235824
Reviewed-by: Evan Shrubsole <eshr@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Evan Shrubsole <eshr@webrtc.org>
Cr-Commit-Position: refs/heads/main@{#35249}
The AsyncTCPSocket is an AsyncPacketSocket which means it
emulates UDP-like (packet) semantics via a TCP stream. When
sending, if the entire packet could not be written then the
packet socket should indicate it wrote the whole thing and
flush out the remaining later when the socket is available.
The WriteEvent signal was already wired up but was not getting
fired (at least with the virtual sockets) since it would not
call Send() enough on the underlying socket to get an
EWOULDBLOCK that would register the async event.
This changes AsyncTCPSocket to repeatedly call Send() on the
underlying socket until the entire packet has been written
or EWOULDBLOCK was returned.
Bug: webrtc:6655
Change-Id: I41e81e0c106c9b3e712a8a0f792d28745d93f2d2
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/168083
Reviewed-by: Qingsi Wang <qingsi@webrtc.org>
Commit-Queue: Steve Anton <steveanton@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30449}
