When a client receives a HEADER frame with a END_STREAM flag,
clientConn.streamByID closes the stream before processing the headers
which may contain a full non-error response. This causes the request's
bodyWriter cancelation to race with the response.

Closing the stream after processing headers allows the response to be
available before the bodyWriter is canceled.

Updates golang/go#20521

Change-Id: I70740e88f75240836e922163a54a6cd89535f7b3
Reviewed-on: https://go-review.googlesource.com/70510
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
Reviewed-by: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

On android, runtime.GOOS returns "android" instead of "linux".

Fixes golang/go#22115.

Change-Id: Idd24d5602f8fa9be1fca0b64bd8966849ea2b5e7
Reviewed-on: https://go-review.googlesource.com/67770Reviewed-by: Ian Lance Taylor <iant@golang.org>

Fixes golang/go#22117.

Change-Id: I0d9c3e126aaf97cd297c84e064e9a521ddac626f
Reviewed-on: https://go-review.googlesource.com/67750Reviewed-by: Ian Lance Taylor <iant@golang.org>

Move the README to README.md so Gerrit can render it; currently
Gerrit only renders files named exactly "README.md" (for example at
https://go.googlesource.com/go).

Add more links to the README explaining how to file issues,
how to submit code changes, where to download the code to and
how to get it. Hopefully this should help people who go to
https://go.googlesource.com/net or https://github.com/golang/net
figure out how to get started with development.

Change-Id: I4eb54f2b210e4f028c9f00955ff19b55643315e6
Reviewed-on: https://go-review.googlesource.com/49850Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>

Some elements and attributes have been removed from the spec,
but are kept here for backwards compatibility.

Also, this makes gen.go support go:generate and write contents into
target files directly.

Finally, this makes `go generate` work on an entire directory.

Change-Id: I8d41840eec69eec1ef08527d8d71786612a3121d
Reviewed-on: https://go-review.googlesource.com/65152
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Tom Bergan <tombergan@google.com>

RFC 1929 describes the SOCKS5 username/password authentication method.
The proxy package implements this method, but doesn't mention RFC 1929.
This change adds mention of RFC 1929.

Change-Id: I30487fb41c1baa16b6ee8a99210168a597e5cb65
Reviewed-on: https://go-review.googlesource.com/60870Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>
Reviewed-by: Avelino <t@avelino.xxx>
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Also, the isindex element has been remove from specification,
so this commit adds a comment about it.

Change-Id: I79a9b1eb9dae8274e2ca498ab73b2e73521d54e9
Reviewed-on: https://go-review.googlesource.com/64230Reviewed-by: Tom Bergan <tombergan@google.com>
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

The number of children (494) got pretty close to the available maximum
number of children (511) in the table. This number increased over the
last year by roughly 65 indicating that we would run out of space in
the next three months. The other fields still have enough room left.
The following table show the growth over the last 1.5 years:

    # Commit Date Children TextOffset Length High Low
    b05061f 2017-09-18 494 28750 36 8407 8402
    859d1a86 2017-09-06 494 28750 36 8407 8402
    ddf80d09 2017-06-14 479 28411 36 8262 8257
    61557ac0 2017-01-26 466 28023 36 8121 8120
    5695785b 2016-10-20 434 27930 36 8135 8134
    07b51741 2016-08-11 424 27866 36 8062 8051
    7864c9ee 2016-07-07 421 27811 36 8049 8038
    3f122ce3 2016-06-09 417 27680 36 8029 8018
    d58ca661 2016-03-04 409 27059 36 7887 7886
    6c581b96 2016-02-01 406 26999 36 7868 7867
    78e1654e 2016-01-20 405 26986 36 7863 7862

Given this rate of grow of max text offset it will overflow in 2021.
Thus use the last of the available 32 bits to encode more children.

Change-Id: I04db02100b202f220a0b4ee509f868db031fd8ab
Reviewed-on: https://go-review.googlesource.com/64330Reviewed-by: Nigel Tao <nigeltao@golang.org>

Change-Id: I557b34aeabf0dc838f9c574966c2d12f02e41a19
Reviewed-on: https://go-review.googlesource.com/57332Reviewed-by: Tom Bergan <tombergan@google.com>

a -> an

Change-Id: I7d0413396e51f16e0a83732a07a5536548e03506
Reviewed-on: https://go-review.googlesource.com/63992Reviewed-by: Ian Lance Taylor <iant@golang.org>

Update the list to revision 38b238d6324042f2c2e6270459d1f4ccfe789fba
(2017-08-28T20:09:01Z).

Change-Id: I80aca3aa21255176e0b6a4c3f8b1b755d0ee2c8e
Reviewed-on: https://go-review.googlesource.com/61890Reviewed-by: Nigel Tao <nigeltao@golang.org>
Run-TryBot: Nigel Tao <nigeltao@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Change-Id: I1011371968c2a548378c7490212a39989a9de75e
Reviewed-on: https://go-review.googlesource.com/38284
Run-TryBot: Marcel van Lohuizen <mpvl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Marcel van Lohuizen <mpvl@golang.org>

Add a new peerMaxHeaderListSize member to ClientConn which records the
SETTINGS_MAX_HEADER_LIST_SIZE requested by the client's peer, and
respect this limit in (*ClientConn) encodeHeaders / encodeTrailers.

Attempting to send more than peerMaxHeaderListSize bytes of headers or
trailers will result in RoundTrip returning errRequestHeaderListSize.

Updates golang/go#13959

Change-Id: Ic707179782acdf8ae543429ea1af7f4f30e67e59
Reviewed-on: https://go-review.googlesource.com/29243
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Tom Bergan <tombergan@google.com>

If the server returns a DATA frame before a HEADERS frame, the client
must close the connection with a PROTOCOL_ERROR as describe in the
section 8.1.2.6.

The current implementation does not reject such sequence and panics
trying to write on the non-initialized bufPipe.

Fixes golang/go#21466

Change-Id: I55755dba259d026529a031e9ff82c915f5e4afae
Reviewed-on: https://go-review.googlesource.com/56770Reviewed-by: Tom Bergan <tombergan@google.com>
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Currently if the http2.Transport hits SettingsMaxConcurrentStreams for a
server, it just makes a new TCP connection and creates the stream on the
new connection. This CL updates that behavior to instead block RoundTrip
until a new stream is available.

I also fixed a second bug, which was necessary to make some tests pass:
Previously, a stream was removed from cc.streams only if either (a) we
received END_STREAM from the server, or (b) we received RST_STREAM from
the server. This CL removes a stream from cc.streams if the request was
cancelled (via ctx.Close, req.Cancel, or resp.Body.Close) before
receiving END_STREAM or RST_STREAM from the server.

Updates golang/go#13774
Updates golang/go#20985
Updates golang/go#21229

Change-Id: I660ffd724c4c513e0f1cc587b404bedb8aff80be
Reviewed-on: https://go-review.googlesource.com/53250
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I35049c37558ddedb9056278c9878f3e3fd6f7f76
Reviewed-on: https://go-review.googlesource.com/53640Reviewed-by: Avelino <t@avelino.xxx>
Reviewed-by: Tom Bergan <tombergan@google.com>

We still support Go 1.6 for a few more days. (it'd normally be dropped
after Go 1.9 final is out)

And maybe we'll need to make a special case for supporting it longer
than normal if gRPC needs to.

Change-Id: I78675f1ef26aa09436a70d0f8aa3a0958768dd14
Reviewed-on: https://go-review.googlesource.com/53641Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

RoundTrip will retry a request if it receives REFUSED_STREAM. To guard
against servers that use REFUSED_STREAM to encourage rate limiting, or
servers that return REFUSED_STREAM deterministically for some requests,
we retry after an exponential backoff and we cap the number of retries.

The exponential backoff starts on the second retry, with a backoff
sequence of 1s, 2s, 4s, etc, with 10% random jitter.

The retry cap was set to 6, somewhat arbitrarily. Rationale: this is
what Firefox does.

Updates golang/go#20985

Change-Id: I4dcac4392ac4a3220d6d839f28bf943fe6b3fea7
Reviewed-on: https://go-review.googlesource.com/50471
Run-TryBot: Tom Bergan <tombergan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Also makes it possible to capture RTM_DESYNC messages when you have a
routing protocol implementation that requires to be aware of the
health condition of kernel RIB on OpenBSD.

Change-Id: Idd0c8c5e8f5ea72a4d56c9a46c137786bcda6354
Reviewed-on: https://go-review.googlesource.com/50191
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Change-Id: I9fb463a5b26e79073a3efc06eea13fd8fcdbf021
Reviewed-on: https://go-review.googlesource.com/50190
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: Id891991e2b8371ad5744e7a0ec392d9d9bbeefdc
Reviewed-on: https://go-review.googlesource.com/50130Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Updates golang/go#16218

Change-Id: I613cf0a65d40a9823c83e0a547fbd48b2236d16e
Reviewed-on: https://go-review.googlesource.com/49630Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>

Change-Id: I38980c8463b7806c129024967713125acf663b82
Reviewed-on: https://go-review.googlesource.com/49791
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Change-Id: Ia3dbad890e3574b34370034fd4cb4bf96323bab3
Reviewed-on: https://go-review.googlesource.com/49790
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Updates golang/go#17613

Change-Id: I0b272c19c263f20084a936b5122bb3cfebacc752
Reviewed-on: https://go-review.googlesource.com/49750Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Since Go 1.7 started including the context package in the core
distribution, this package has been forwarding to the core package's
implementation to the best of its ability. Since type aliases are due
to be included in the Go 1.9 release, use them to forward more
completely.

Change-Id: I97b92cd34b3f4426f565bb3d115f119d3c83a1dd
Reviewed-on: https://go-review.googlesource.com/49020
Run-TryBot: Emmanuel Odeke <emm.odeke@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Fix two format argument warnings found by vet.

Change-Id: I01d55197c110c2a7486b1bd5dbd0afd74812e535
Reviewed-on: https://go-review.googlesource.com/48090Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Fixes golang/go#13456

Change-Id: I0b938f824c47b29ac2026eff83e61c2f227a6cc1
Reviewed-on: https://go-review.googlesource.com/47530
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I248e5e5215960bceaba357d3d4a6e72191f955b1
Reviewed-on: https://go-review.googlesource.com/47372
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>

Change-Id: I588bc65b0052efb12b3b7bc2267deedcdad029ce
Reviewed-on: https://go-review.googlesource.com/47371
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>

Change-Id: I17444d1cc3ce03b6947aaae18d4cecdf9c36193f
Reviewed-on: https://go-review.googlesource.com/47110Reviewed-by: Ian Gudger <igudger@google.com>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

CL 47095 applied a change to ExampleWithTimeout to resolve test
flakiness. This change deviated ExampleWithTimeout further away
from the ExampleWithTimeout in standard library's context package.

There was a minor issue spotted in code review. Instead of attempting
to resolve it here, it's better to use the version of ExampleWithTimeout
from standard library that doesn't have the issue, and reduce
maintenance.

Change-Id: I3da9d7f7c35b33b52635df4c61058cb0b8ae95d9
Reviewed-on: https://go-review.googlesource.com/47212
Run-TryBot: Dmitri Shuralyov <shurcool@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Got a trybot failure in
https://storage.googleapis.com/go-build-log/8aee0b8b/linux-amd64_38d0be3e.log

Change-Id: I867a658968a0088c9f3f76ff235816e4179f0bec
Reviewed-on: https://go-review.googlesource.com/47095Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>

If the transport had previously sent a RST_STREAM but had not yet
deleted the stream from its list of active streams, we should refund
connection-level flow control for any DATA frame received as such
DATA frames will never be read.

We already refund connection-level flow control if a stream closes with
unread data in bufPipe. However, when we receive a DATA frame after
reset, we don't bother writing it to bufPipe, so we have to refund the
flow control separately.

Updates golang/go#20469

Change-Id: I5a9810a5d6b1bd7e291173af53646246545a6665
Reviewed-on: https://go-review.googlesource.com/46591
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I51969f70d4fc69829fd5a8bcd8a34b3be15b9db0
Reviewed-on: https://go-review.googlesource.com/46930
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Also renames the existing example function to make it visible with
godoc command and godoc.org.

Change-Id: I0e5bef22f2a95930c59c6e79f010ba61651012cd
Reviewed-on: https://go-review.googlesource.com/46791
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
Reviewed-by: Ian Gudger <igudger@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

Change-Id: I288f01233defe6c673ef9dd5ec79353d3663a9af
Reviewed-on: https://go-review.googlesource.com/46790
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Gudger <igudger@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Fixes golang/go#19821

Change-Id: Iee5d43dfeaaf580c39ca38a7a3a74cf2a8b347b3
Reviewed-on: https://go-review.googlesource.com/44381
Run-TryBot: Marcel van Lohuizen <mpvl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>

Updates golang/go#16218

Change-Id: Ice2821e873807ac4222825f0ca6f47b6a53f90b9
Reviewed-on: https://go-review.googlesource.com/46630Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>

Parsing and building can now be done 100% allocation free for most
types. Non-name variable-length fields such as TXT records still
require allocations. Building with compression still causes allocations
due to the use of a map. Both parsing and building will perform
allocations in error paths.

Updates golang/go#16218

Change-Id: I644a7776bd0bf4ad20def521cee44eeadc5e9c06
Reviewed-on: https://go-review.googlesource.com/41145
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>