Fixes #5541.
This time for sure.

R=golang-dev, minux.ma, iant
CC=golang-dev
https://golang.org/cl/9668043

If a union contains a pointer, it will mess up the garbage collector, causing memory corruption.

R=golang-dev, dave, nightlyone, adg, dvyukov, bradfitz, minux.ma, r, iant
CC=golang-dev
https://golang.org/cl/8469043

This is needed for preemptive scheduler, because the goroutine
can be preempted at surprising points.

R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/9376043

When cgo is used, runtime creates an additional M to handle callbacks on threads not created by Go.
This effectively disabled deadlock detection, which is a right thing, because Go program can be blocked
and only serve callbacks on external threads.
This also disables deadlock detection under race detector, because it happens to use cgo.
With this change the additional M is created lazily on first cgo call. So deadlock detector
works for programs that import "C", "net" or "net/http/pprof" but do not use them in fact.
Also fixes deadlock detector under race detector.
It should be fine to create the M later, because C code can not call into Go before first cgo call,
because C code does not know when Go initialization has completed. So a Go program need to call into C
first either to create an external thread, or notify a thread created in global ctor that Go
initialization has completed.
Fixes #4973.
Fixes #5475.

R=golang-dev, minux.ma, iant
CC=golang-dev
https://golang.org/cl/9303046

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/9662043

The original code was correct. The count returned must be the length
of the input slice, not the length of the formatted message.

««« original CL description
log/syslog: report errors from Fprintf
Thanks to chiparus for identifying this.

Fixes #5541.

R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/9658043
»»»

R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/9644044

Currently per-sizeclass stats are lost for destroyed MCache's. This patch fixes this.
Also, only update mstats.heap_alloc on heap operations, because that's the only
stat that needs to be promptly updated. Everything else needs to be up-to-date only in ReadMemStats().

R=golang-dev, remyoudompheng, dave, iant
CC=golang-dev
https://golang.org/cl/9207047

Thanks to chiparus for identifying this.

Fixes #5541.

R=golang-dev, iant
CC=golang-dev
https://golang.org/cl/9658043

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/9392049

The nlistmin/size thresholds are copied from tcmalloc,
but are unnecesary for Go malloc. We do not do explicit
frees into MCache. For sparse cases when we do (mainly hashmap),
simpler logic will do.

R=rsc, dave, iant
CC=gobot, golang-dev, r, remyoudompheng
https://golang.org/cl/9373043

Also delete the special case for exp, which isn't necessary any more.

Fixes #5529.

R=rsc, nightlyone
CC=golang-dev
https://golang.org/cl/9611048

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/9564044

R=dsymonds
CC=gobot, golang-dev
https://golang.org/cl/9570043

A bufio.Writer.Flush marks the usual end of a Writer's
life. Recycle its internal buffer on those explicit flushes,
but not on normal, as-needed internal flushes.

benchmark               old ns/op    new ns/op    delta
BenchmarkWriterEmpty         1959          727  -62.89%

benchmark              old allocs   new allocs    delta
BenchmarkWriterEmpty            2            1  -50.00%

benchmark               old bytes    new bytes    delta
BenchmarkWriterEmpty         4215           83  -98.03%

R=gri, iant
CC=gobot, golang-dev, voidlogic7
https://golang.org/cl/9459044

Works around a bug in nginx: http://trac.nginx.org/nginx/ticket/358

Fixes #5522

R=iant
CC=gobot, golang-dev
https://golang.org/cl/9627043

This should have been removed in 45c12efb4635. Not a correctness
issue, but unnecessary work.

This CL also adds paranoia checks in removeDep so this doesn't
happen again.

Fixes #5502

R=adg
CC=gobot, golang-dev, google
https://golang.org/cl/9543043

undo CL 8478044 / 0d28fd55e721

Lack of consensus.

««« original CL description
time: add Time.FormatAppend

This is a version of Time.Format that doesn't require allocation.

Fixes #5192
Update #5195

R=r
CC=gobot, golang-dev
https://golang.org/cl/8478044
»»»

R=r
CC=golang-dev
https://golang.org/cl/9462049

Fixes #5172.

R=golang-dev, bradfitz, r
CC=golang-dev
https://golang.org/cl/9614044

Fixes #5479.

R=golang-dev, dave
CC=golang-dev
https://golang.org/cl/9416047

R=minux.ma
CC=golang-dev
https://golang.org/cl/9459047

R=golang-dev, gri
CC=golang-dev
https://golang.org/cl/9459046

This isn't clearly a bug on Go's part, but it triggers a bug in Firefox
which means that crypto/tls and net/http cannot be wired up together
unless NextProtos includes "http/1.1". When net/http sets up the
tls.Config, it does this and so works fine. But anyone setting up the
tls.Config themselves will hit the Firefox bug.

Fixes #5445.

R=golang-dev, bradfitz, r
CC=golang-dev
https://golang.org/cl/9539045

as OpenBSD lacks 4 errno constants, api/go1.txt is updated so that
api check won't fail.

R=golang-dev, iant, bradfitz, r
CC=golang-dev
https://golang.org/cl/9149045

Several places used io.WriteString unnecessarily when the
static type already implemented WriteString. No need to
check for it at runtime.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/9608043

It contains the LHS of the range clause and gets
instrumented by racewalk, but it doesn't have any meaning.

Fixes #5446.

R=golang-dev, dvyukov, daniel.morsing, r
CC=golang-dev
https://golang.org/cl/9560044

R=golang-dev, daniel.morsing, r
CC=golang-dev
https://golang.org/cl/9358043

Several old TODOs are either resolved now (e.g. when is a return
needed), or are from a time the language wasn't frozen (^ for uints
only). Consolidated the others.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/9599044

Fixes #5434.

R=golang-dev, dsymonds
CC=golang-dev
https://golang.org/cl/9595044

R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/9495045

Failures caused by errors like invalid signatures or missing hash
functions cause rather generic, unhelpful error messages because no
trust chain can be constructed: "x509: certificate signed by unknown
authority."

With this change, authority errors may contain the reason why an
arbitary candidate step in the chain was rejected. For example, in the
event of a missing hash function the error looks like:

x509: certificate signed by unknown authority (possibly because of
"crypto/x509: cannot verify signature: algorithm unimplemented" while
trying to verify candidate authority certificate 'Thawte SGC CA')

Fixes 5058.

R=golang-dev, r
CC=golang-dev
https://golang.org/cl/9104051

Currently most of the 32-bit builder are broken.
Fixes #5516.

R=golang-dev, dave, iant
CC=golang-dev
https://golang.org/cl/9573043

benchmark                                   old ns/op    new ns/op    delta
BenchmarkServerFakeConnNoKeepAlive              14431        14247   -1.28%
BenchmarkServerFakeConnWithKeepAlive            11618        11357   -2.25%
BenchmarkServerFakeConnWithKeepAliveLite         6735         6427   -4.57%
BenchmarkServerHandlerTypeLen                    8842         8740   -1.15%
BenchmarkServerHandlerNoLen                      8001         7828   -2.16%
BenchmarkServerHandlerNoType                     8270         8227   -0.52%
BenchmarkServerHandlerNoHeader                   6148         5920   -3.71%

benchmark                                  old allocs   new allocs    delta
BenchmarkServerFakeConnNoKeepAlive                 30           29   -3.33%
BenchmarkServerFakeConnWithKeepAlive               25           24   -4.00%
BenchmarkServerFakeConnWithKeepAliveLite           10            9  -10.00%
BenchmarkServerHandlerTypeLen                      18           17   -5.56%
BenchmarkServerHandlerNoLen                        15           14   -6.67%
BenchmarkServerHandlerNoType                       16           15   -6.25%
BenchmarkServerHandlerNoHeader                     10            9  -10.00%

benchmark                                   old bytes    new bytes    delta
BenchmarkServerFakeConnNoKeepAlive               2557         2492   -2.54%
BenchmarkServerFakeConnWithKeepAlive             2260         2194   -2.92%
BenchmarkServerFakeConnWithKeepAliveLite         1092         1026   -6.04%
BenchmarkServerHandlerTypeLen                    1941         1875   -3.40%
BenchmarkServerHandlerNoLen                      1898         1832   -3.48%
BenchmarkServerHandlerNoType                     1906         1840   -3.46%
BenchmarkServerHandlerNoHeader                   1092         1026   -6.04%

Update #5195

R=golang-dev, daniel.morsing
CC=golang-dev
https://golang.org/cl/9492044

R=golang-dev, dave, iant
CC=golang-dev
https://golang.org/cl/9575043

Update #5199

R=golang-dev, dvyukov
CC=golang-dev
https://golang.org/cl/8825043

R=golang-dev, bradfitz
CC=dvyukov, golang-dev, mikioh.mikioh
https://golang.org/cl/9495044

Fixes #5493.

R=golang-dev, minux.ma, iant
CC=golang-dev
https://golang.org/cl/9557043

Don't allocate for the Date or Content-Length headers.
A custom Date header formatter replaces use of time.Format.

benchmark                                   old ns/op    new ns/op    delta
BenchmarkClientServer                           67791        64424   -4.97%
BenchmarkClientServerParallel4                  62956        58533   -7.03%
BenchmarkClientServerParallel64                 62043        54789  -11.69%
BenchmarkServer                                254609       229060  -10.03%
BenchmarkServerFakeConnNoKeepAlive              17038        16316   -4.24%
BenchmarkServerFakeConnWithKeepAlive            14184        13226   -6.75%
BenchmarkServerFakeConnWithKeepAliveLite         8591         7532  -12.33%
BenchmarkServerHandlerTypeLen                   10750         9961   -7.34%
BenchmarkServerHandlerNoLen                      9535         8935   -6.29%
BenchmarkServerHandlerNoType                     9858         9362   -5.03%
BenchmarkServerHandlerNoHeader                   7754         6856  -11.58%

benchmark                                  old allocs   new allocs    delta
BenchmarkClientServer                              68           66   -2.94%
BenchmarkClientServerParallel4                     68           66   -2.94%
BenchmarkClientServerParallel64                    68           66   -2.94%
BenchmarkServer                                    21           19   -9.52%
BenchmarkServerFakeConnNoKeepAlive                 32           30   -6.25%
BenchmarkServerFakeConnWithKeepAlive               27           25   -7.41%
BenchmarkServerFakeConnWithKeepAliveLite           12           10  -16.67%
BenchmarkServerHandlerTypeLen                      19           18   -5.26%
BenchmarkServerHandlerNoLen                        17           15  -11.76%
BenchmarkServerHandlerNoType                       17           16   -5.88%
BenchmarkServerHandlerNoHeader                     12           10  -16.67%

Update #5195

R=nigeltao
CC=golang-dev
https://golang.org/cl/9432046

This is in preparation for netpoll windows version.

R=golang-dev, bradfitz
CC=dvyukov, golang-dev, mikioh.mikioh
https://golang.org/cl/9569043

R=golang-dev, r, dave
CC=golang-dev
https://golang.org/cl/9307045

R=r, adg, dsymonds
CC=golang-dev
https://golang.org/cl/9535044