David Chase [Mon, 9 May 2016 18:59:25 +0000 (14:59 -0400)]
cmd/compile: reduce element size of arrays in sparse{map,set}
sparseSet and sparseMap only need 32 bit integers in their
arrays, since a sparseEntry key is also limited to 32 bits.
This appears to reduce the space allocated for at least
one pathological compilation by 1%, perhaps more.
Not necessarily for 1.7, but it saves a little and is very
low-risk.
Change-Id: Icf1185859e9f5fe1261a206b441e02c34f7d02fd
Reviewed-on: https://go-review.googlesource.com/22972
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Russ Cox <rsc@golang.org>
Cuihtlauac ALVARADO [Tue, 17 May 2016 07:27:00 +0000 (09:27 +0200)]
runtime: in tests, make sure gdb does not start with a shell
On some systems, gdb is set to: "startup-with-shell on". This
breaks runtime_test. This just make sure gdb does not start by
spawning a shell.
Fixes #15354
Change-Id: Ia040931c61dea22f4fdd79665ab9f84835ecaa70
Reviewed-on: https://go-review.googlesource.com/23142 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
which can leave g's temporary value on the stack as a uintptr, rather
than an unsafe.Pointer. Instead, recognize uintptr-to-unsafe.Pointer
conversions when reordering function calls to instead produce:
Mikio Hara [Wed, 18 May 2016 01:54:20 +0000 (10:54 +0900)]
net: deflake TestLookupPort on Android
Looks like some version of Android still fails with "servname not
supported for ai_socktype". It probably doesn't support
ai_socktype=SOCK_STREAM.
Updates #14576.
Change-Id: I77ecff147d5b759e3281b3798c60f150a4aab811
Reviewed-on: https://go-review.googlesource.com/23194 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Brad Fitzpatrick [Wed, 18 May 2016 01:00:32 +0000 (01:00 +0000)]
regexp/syntax: clarify that \Z means Perl's \Z
Fixes #14793
Change-Id: I408056d096cd6a999fa5e349704b5ea8e26d2e4e
Reviewed-on: https://go-review.googlesource.com/23201 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Ian Lance Taylor [Wed, 18 May 2016 01:16:47 +0000 (18:16 -0700)]
runtime: deflake TestSignalExitStatus
The signal might get delivered to a different thread, and that thread
might not run again before the currently running thread returns and
exits. Sleep to give the other thread time to pick up the signal and
crash.
Not tested for all cases, but, optimistically:
Fixes #14063.
Change-Id: Iff58669ac6185ad91cce85e0e86f17497a3659fd
Reviewed-on: https://go-review.googlesource.com/23203
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>
Brad Fitzpatrick [Wed, 18 May 2016 00:35:43 +0000 (00:35 +0000)]
net/http: add test confirming a connection reuse case
Verify that for a server doing chunked encoding, with the final data
and EOF arriving together, the client will reuse the connection even
if it closes the body without seeing an EOF. The server sends at least
one non-zero chunk and one zero chunk. This verifies that the client's
bufio reading reads ahead and notes the EOF, so even if the JSON
decoder doesn't read the EOF itself, as long as somebody sees it, a
close won't forcible tear down the connection. This was true at least
of https://golang.org/cl/21291
No code change. Test already passed (even with lots of runs, including
in race mode with randomized goroutine scheduling).
James Chacon [Fri, 6 May 2016 18:51:56 +0000 (11:51 -0700)]
runtime: prevent racefini from being invoked more than once
racefini calls __tsan_fini which is C code and at the end of it
invoked the standard C library exit(3) call. This has undefined
behavior if invoked more than once. Specifically in C++ programs
it caused static destructors to run twice. At least on glibc
impls it also means the at_exit handlers list (where those are
stored) also free's a list entry when it completes these. So invoking
twice results in a double free at exit which trips debug memory
allocation tracking.
Fix all of this by using an atomic as a boolean barrier around
calls to racefini being invoked > 1 time.
Alessandro Arzilli [Sat, 14 May 2016 05:54:02 +0000 (07:54 +0200)]
debug/gosym: parse remote packages correctly
Fixes #15675
Change-Id: I8bad220988e5d690f20804db970b2db037c81187
Reviewed-on: https://go-review.googlesource.com/23086 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Colin Cross [Tue, 17 May 2016 20:09:11 +0000 (13:09 -0700)]
reflect: remove out of date UTF-8 StructOf restriction
The initial implementation of reflect.StructOf in
https://golang.org/cl/9251 had a limitation that field names had to be
ASCII, which was later lifted by https://golang.org/cl/21777. Remove
the out-of-date documentation disallowing UTF-8 field names.
Updates: #5748
Updates: #15064
Change-Id: I2c5bfea46bfd682449c6e847fc972a1a131f51b7
Reviewed-on: https://go-review.googlesource.com/23170 Reviewed-by: David Crawshaw <crawshaw@golang.org>
Brad Fitzpatrick [Mon, 16 May 2016 22:45:18 +0000 (15:45 -0700)]
A+C: add Andrew Werner (corporate CLA for Upthere, Inc)
Change-Id: I7627e480d5d2366cba223fd81635c4115649f752
Reviewed-on: https://go-review.googlesource.com/23154 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Ian Lance Taylor [Mon, 16 May 2016 21:11:01 +0000 (14:11 -0700)]
runtime: yield after raising signal that should kill process
Issue #15613 points out that the darwin builders have been getting
regular failures in which a process that should exit with a SIGPIPE
signal is instead exiting with exit status 2. The code calls
runtime.raise. On most systems runtime.raise is the equivalent of
pthread_kill(gettid(), sig); that is, it kills the thread with the
signal, which should ensure that the program does not keep going. On
darwin, however, runtime.raise is actually kill(getpid(), sig); that is,
it sends a signal to the entire process. If the process decides to
deliver the signal to a different thread, then it is possible that in
some cases the thread that calls raise is able to execute the next
system call before the signal is actually delivered. That would cause
the observed error.
I have not been able to recreate the problem myself, so I don't know
whether this actually fixes it. But, optimistically:
Fixed #15613.
Change-Id: I60c0a9912aae2f46143ca1388fd85e9c3fa9df1f
Reviewed-on: https://go-review.googlesource.com/23152
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
David Chase [Thu, 21 Apr 2016 17:24:58 +0000 (13:24 -0400)]
cmd/compile: use sparse algorithm for phis in large program
This adds a sparse method for locating nearest ancestors
in a dominator tree, and checks blocks with more than one
predecessor for differences and inserts phi functions where
there are.
Uses reversed post order to cut number of passes, running
it from first def to last use ("last use" for paramout and
mem is end-of-program; last use for a phi input from a
backedge is the source of the back edge)
Includes a cutover from old algorithm to new to avoid paying
large constant factor for small programs. This keeps normal
builds running at about the same time, while not running
over-long on large machine-generated inputs.
Add "phase" flags for ssa/build -- ssa/build/stats prints
number of blocks, values (before and after linking references
and inserting phis, so expansion can be measured), and their
product; the product governs the cutover, where a good value
seems to be somewhere between 1 and 5 million.
Among the files compiled by make.bash, this is the shape of
the tail of the distribution for #blocks, #vars, and their
product:
The old algorithm is indeed usually fastest, for 99%ile
values of usually.
The fix to LookupVarOutgoing
( https://go-review.googlesource.com/#/c/22790/ )
deals with some of the same problems addressed by this CL,
but on at least one bug ( #15537 ) this change is still
a significant help.
With this CL:
/tmp/gopath$ rm -rf pkg bin
/tmp/gopath$ time go get -v -gcflags -memprofile=y.mprof \
github.com/gogo/protobuf/test/theproto3/combos/...
...
real 4m35.200s
user 13m16.644s
sys 0m36.712s
and pprof reports 3.4GB allocated in one of the larger profiles
With tip:
/tmp/gopath$ rm -rf pkg bin
/tmp/gopath$ time go get -v -gcflags -memprofile=y.mprof \
github.com/gogo/protobuf/test/theproto3/combos/...
...
real 10m36.569s
user 25m52.286s
sys 4m3.696s
and pprof reports 8.3GB allocated in the same larger profile
With this CL, most of the compilation time on the benchmarked
input is spent in register/stack allocation (cumulative 53%)
and in the sparse lookup algorithm itself (cumulative 20%).
Fixes #15537.
Change-Id: Ia0299dda6a291534d8b08e5f9883216ded677a00
Reviewed-on: https://go-review.googlesource.com/22342 Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Austin Clements [Mon, 9 May 2016 15:29:34 +0000 (11:29 -0400)]
runtime: use conventional shift style for gcBitsChunkBytes
The convention for writing something like "64 kB" is 64<<10, since
this is easier to read than 1<<16. Update gcBitsChunkBytes to follow
this convention.
Change-Id: I5b5a3f726dcf482051ba5b1814db247ff3b8bb2f
Reviewed-on: https://go-review.googlesource.com/23132 Reviewed-by: Rick Hudson <rlh@golang.org>
Austin Clements [Thu, 12 May 2016 22:10:03 +0000 (18:10 -0400)]
runtime: improve heapBitsSetType documentation
Currently the heapBitsSetType documentation says that there are no
races on the heap bitmap, but that isn't exactly true. There are no
*write-write* races, but there are read-write races. Expand the
documentation to explain this and why it's okay.
Change-Id: Ibd92b69bcd6524a40a9dd4ec82422b50831071ed
Reviewed-on: https://go-review.googlesource.com/23092 Reviewed-by: Rick Hudson <rlh@golang.org>
Austin Clements [Wed, 11 May 2016 18:57:33 +0000 (14:57 -0400)]
runtime: perform publication barrier even for noscan objects
Currently we only execute a publication barrier for scan objects (and
skip it for noscan objects). This used to be okay because GC would
never consult the object itself (so it wouldn't observe uninitialized
memory even if it found a pointer to a noscan object), and the heap
bitmap was pre-initialized to noscan.
However, now we explicitly initialize the heap bitmap for noscan
objects when we allocate them. While the GC will still never consult
the contents of a noscan object, it does need to see the initialized
heap bitmap. Hence, we need to execute a publication barrier to make
the bitmap visible before user code can expose a pointer to the newly
allocated object even for noscan objects.
Tom Bergan [Fri, 13 May 2016 05:03:46 +0000 (22:03 -0700)]
net, net/http: don't trace DNS dials
This fixes change https://go-review.googlesource.com/#/c/23069/, which
assumes all DNS requests are UDP. This is not true -- DNS requests can
be TCP in some cases. See:
https://tip.golang.org/src/net/dnsclient_unix.go#L154
https://en.wikipedia.org/wiki/Domain_Name_System#Protocol_transport
Also, the test code added by the above change doesn't actually test
anything because the test uses a faked DNS resolver that doesn't
actually make any DNS queries. I fixed that by adding another test
that uses the system DNS resolver.
David Crawshaw [Fri, 13 May 2016 16:33:27 +0000 (12:33 -0400)]
reflect: fix method type string
By picking up a spurious tFlagExtraStar, the method type was printing
as unc instead of func.
Updates #15673
Change-Id: I0c2c189b99bdd4caeb393693be7520b8e3f342bf
Reviewed-on: https://go-review.googlesource.com/23103 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The commit had a wrong prefix in the description line, probably
misreconginized something. As a result it broke golang.org/x/tools/godoc
and golang.org/x/tools/cmd/godoc like the following:
Brad Fitzpatrick [Thu, 12 May 2016 22:12:11 +0000 (22:12 +0000)]
net, net/http: don't trace UDP dials
The httptrace.ConnectStart and ConnectDone hooks are just about the
post-DNS connection to the host. We were accidentally also firing on
the UDP dials to DNS. Exclude those for now. We can add them back
later as separate hooks if desired. (but they'd only work for pure Go
DNS)
This wasn't noticed earlier because I was developing on a Mac at the
time, which always uses cgo for DNS. When running other tests on
Linux, I started seeing UDP dials.
Emmanuel Odeke [Wed, 11 May 2016 04:28:30 +0000 (21:28 -0700)]
cmd/build: reject non-existant directories in ImportDir
Re-apply @adg's CL https://golang.org/cl/7129048 that was
previously disabled in https://golang.org/cl/7235052 because
it broke `godoc net/http` for go1.1.
Currently `godoc net/http` seems to work fine with this CL.
Fixes #3428.
Change-Id: I7df06df02fd62dededac6ec60bea62561be59cf1
Reviewed-on: https://go-review.googlesource.com/23013
Run-TryBot: Andrew Gerrand <adg@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Andrew Gerrand <adg@golang.org>
Dmitry Vyukov [Tue, 3 May 2016 14:44:25 +0000 (16:44 +0200)]
cmd/trace: split large traces into parts
Trace viewer cannot handle traces larger than 256MB (limit on js string size):
https://github.com/catapult-project/catapult/issues/627
And even that is problematic (chrome hangs and crashes).
Split large traces into 100MB parts. Somewhat clumsy, but I don't see any other
solution (other than rewriting trace viewer). At least it works reliably now.
Fixes #15482
Change-Id: I993b5f43d22072c6f5bd041ab5888ce176f272b2
Reviewed-on: https://go-review.googlesource.com/22731 Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
Joel Sing [Sun, 8 May 2016 16:13:03 +0000 (02:13 +1000)]
runtime: stop using sigreturn on openbsd/386
In future releases of OpenBSD, the sigreturn syscall will no longer
exist. As such, stop using sigreturn on openbsd/386 and just return
from the signal trampoline (as we already do for openbsd/amd64 and
openbsd/arm).
Change-Id: Ic4de1795bbfbfb062a685832aea0d597988c6985
Reviewed-on: https://go-review.googlesource.com/23024 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Robert Griesemer [Wed, 11 May 2016 19:40:17 +0000 (12:40 -0700)]
cmd/compile: use ONAME instead of OPACK in binary export format
This is addressing feedback given on golang.org/cl/23052;
we do it in a separate CL to separate the functional from
the rename change.
ONAME was not used in the export data, but it's the natural node op
where we used OPACK instead. Renamed.
Furthermore, OPACK and ONONAME nodes are replaced by the type checker
with ONAME nodes, so OPACK nodes cannot occur when exporting type-checked
code. Removed a special-case for OPACK nodes since they don't appear.
Change-Id: I78b01a1badbf60e9283eaadeca2578a65d28cbd2
Reviewed-on: https://go-review.googlesource.com/23053
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Ian Lance Taylor [Wed, 11 May 2016 16:37:46 +0000 (09:37 -0700)]
cmd/cgo: remove //extern for check functions in gccgo Go prologue
The //extern comments are incorrect and cause undefined symbol
errorswhen building cgo code with -compiler=gccgo. The code is already
designed to use weak references, and that support relies on the cgo
check functions being treated as local functions.
Change-Id: Ib38a640cc4ce6eba74cfbf41ba7147ec88769ec0
Reviewed-on: https://go-review.googlesource.com/23014
Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Mikio Hara [Wed, 11 May 2016 04:04:22 +0000 (13:04 +0900)]
net: fix nits found by vet
Change-Id: I323231f31c4e1e7415661ebd943a90b2f1e9da1c
Reviewed-on: https://go-review.googlesource.com/23020 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Robert Griesemer [Wed, 11 May 2016 00:06:19 +0000 (17:06 -0700)]
go/importer: use correct path when checking if package was already imported
The importer uses a global (shared) package map across multiple imports
to determine if a package was imported before. That package map is usually
indexed by package (import) path ('id' in this code). However, the binary
importer was using the incoming (possibly unclean) path.
Fixes #15517.
Change-Id: I0c32a708dfccf345e0353fbda20ad882121e437c
Reviewed-on: https://go-review.googlesource.com/23012
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Alex Brainman [Thu, 7 Apr 2016 05:12:32 +0000 (15:12 +1000)]
runtime: make mksyscall_windows.go flags do what they say they do
The -systemdll and -xsys flags generate broken code in some situations
(see issue for details). Fix all that.
This CL only fixes bugs in existing code, but I have more changes comming:
golang.org/x/sys/windows is not the only package that uses mksyscall_windows.go.
golang.org/x/exp/shiny and github.com/derekparker/delve do too. I also have
few personal packages that use mksyscall_windows.go. None of those packages
are aware of new -xsys flag. I would like to change mksyscall_windows.go, so
external packages do not need to use -xsys flag. I would love to get rid of
-xsys flag altogether, but I don't see how it is possible. So I will, probably,
replace -xsys with a flag that means opposite to -xsys, and use new flag
everywhere in standard libraries. Flag name suggestions are welcome.
-systemdll flag makes users code more "secure". I would like to make -systemdll
behaviour a default for all mksyscall_windows.go users. We use that already in
standard library. If we think "secure" is important, we should encourage it in
all users code. If mksyscall_windows.go user insist on using old code, provide
-use_old_loaddll (need good name here) flag for that. So -systemdll flag will
be replaced with -use_old_loaddll.
Ian Lance Taylor [Tue, 10 May 2016 21:27:32 +0000 (14:27 -0700)]
os/user: don't create C function mygetgrouplist
Instead of exporting the C function mygetgrouplist as a global symbol to
conflict with other symbols of the same name, use trivial Go code and a
static C function.
Change-Id: I98dd667814d0a0ed8f7b1d4cfc6483d5a6965b26
Reviewed-on: https://go-review.googlesource.com/23008
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
David du Colombier [Tue, 19 Apr 2016 18:59:03 +0000 (20:59 +0200)]
os: add TestReadAtOffset
In the Plan 9 kernel, there used to be a bug in the implementation of
the pread syscall, where the channel offset was erroneously updated after
calling pread on a file.
This test verifies that ReadAt is behaving as expected.
Fixes #14534.
Change-Id: Ifc9fd40a1f94879ee7eb09b2ffc369aa2bec2926
Reviewed-on: https://go-review.googlesource.com/22244
Run-TryBot: David du Colombier <0intro@gmail.com> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This change makes encoding and decoding support integer types in map
keys, converting to/from JSON string keys.
JSON object keys are still sorted lexically, even though the keys may be
integer strings.
For backwards-compatibility, the existing Text(Un)Marshaler support for
map keys (added in CL 20356) does not take precedence over the default
encoding for string types. There is no such concern for integer types,
so integer map key encoding is only used as a fallback if the map key
type is not a Text(Un)Marshaler.
Tilman Dilo [Mon, 9 May 2016 21:37:07 +0000 (23:37 +0200)]
crypto/cipher: execute AES-GCM decryption example
The decryption example for AES-GCM was not executed, hiding the fact
that the provided ciphertext could not be authenticated.
This commit adds the required output comment, replaces the ciphertext
with a working example, and removes an unnecessary string conversion
along the way.
Brad Fitzpatrick [Mon, 9 May 2016 17:21:11 +0000 (17:21 +0000)]
time: don't depend on the io package
The time package has never depended on the io package until
a recent change during Go 1.7 to use the io.Seek* constants.
The go/build dependency check didn't catch this because "time" was
allowed to depend on meta package group "L0", which included "io".
Adding the "io" package broke one of Dmitry's tools. The tool is
fixable, but it's also not necessary for us to depend on "io" at all
for some constants. Mirror the constants instead, and change
deps_test.go to prevent an io dependency in the future.
Change-Id: I74325228565279a74fa4a2f419643f5710e3e09f
Reviewed-on: https://go-review.googlesource.com/22960
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Address two documentation issues:
1) Document that the GZIP and ZLIB footer is only verified when the
reader has been fully consumed.
2) The zlib reader is guaranteed to not read past the EOF if the
input io.Reader is also a io.ByteReader. This functionality was
documented in the flate and gzip packages but not on zlib.
David Chase [Sat, 7 May 2016 00:05:02 +0000 (17:05 -0700)]
cmd/compile: correct sparseSet probes in regalloc to avoid index error
In regalloc, a sparse map is preallocated for later use by
spill-in-loop sinking. However, variables (spills) are added
during register allocation before spill sinking, and a map
query involving any of these new variables will index out of
bounds in the map.
To fix:
1) fix the queries to use s.orig[v.ID].ID instead, to ensure
proper indexing. Note that s.orig will be nil for values
that are not eligible for spilling (like memory and flags).
2) add a test.
Fixes #15585.
Change-Id: I8f2caa93b132a0f2a9161d2178320d5550583075
Reviewed-on: https://go-review.googlesource.com/22911 Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
cmd/compile: add -linkobj flag to allow writing object file in two parts
This flag is experimental and the semantics may change
even after Go 1.7 is released. There are no changes to code
not using the flag.
The first part is for reading by future compiles.
The second part is for reading by the final link step.
Splitting the file this way allows distributed build systems
to ship the compile-input part only to compile steps and
the linker-input part only to linker steps.
The first part is basically just the export data,
and the second part is basically everything else.
The overall files still have the same broad structure,
so that existing tools will work with both halves.
It's just that various pieces are empty in the two halves.
This also copies the two bits of data the linker needed from
export data into the object header proper, so that the linker
doesn't need any export data at all. That eliminates a TODO
that was left for switching to the binary export data.
(Now the linker doesn't need to know about the switch.)
The default is still to write out a combined output file.
Nothing changes unless you pass -linkobj to the compiler.
There is no support in the go command for -linkobj,
since the go command doesn't copy objects around.
The expectation is that other build systems (like bazel, say)
might take advantage of this.
The header adjustment and the option for the split output
was intended as part of the zip archives, but the zip archives
have been cut from Go 1.7. Doing this to the current archives
both unblocks one step in the switch to binary export data
and enables alternate build systems to experiment with the
new flag using the Go 1.7 release.
Change-Id: I8b6eab25b8a22b0a266ba0ac6d31e594f3d117f3
Reviewed-on: https://go-review.googlesource.com/22500
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Dmitry Vyukov [Mon, 9 May 2016 13:11:24 +0000 (15:11 +0200)]
runtime/race: deflake test
The test sometimes fails on builders.
The test uses sleeps to establish the necessary goroutine
execution order. If sleeps undersleep/oversleep
the race is still reported, but it can be reported when the
main test goroutine returns. In such case test driver
can't match the race with the test and reports failure.
Wait for both test goroutines to ensure that the race
is reported in the test scope.