- 15 Sep, 2016 13 commits
-
-
Dave Cheney authored
Fix conflict between CL 29213 and 29134. Change-Id: Ie58bd7195893d7e634f1b257ee0bdd3250cd23c2 Reviewed-on: https://go-review.googlesource.com/29137 TryBot-Result: Gobot Gobot <gobot@golang.org> Run-TryBot: Dave Cheney <dave@cheney.net> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
-
Josh Bleecher Snyder authored
staticassign unwraps all CONVNOPs. However, in the included test, we need the CONVNOP for everything to typecheck. Stop unwrapping unnecessarily. The code we generate for this example is suboptimal, but that's not new; see #17113. Fixes #17111. Change-Id: I29532787a074a6fe19a5cc53271eb9c84bf1b576 Reviewed-on: https://go-review.googlesource.com/29213 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
-
Dave Cheney authored
After the removal of the old backend many types are no longer referenced outside internal/gc. Make these functions private so that tools like honnef.co/go/unused can spot when they become dead code. In doing so this CL identified several previously public helpers which are no longer used, so removes them. This should be the last of the public functions. Change-Id: I7e9c4e72f86f391b428b9dddb6f0d516529706c3 Reviewed-on: https://go-review.googlesource.com/29134 Run-TryBot: Dave Cheney <dave@cheney.net> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Josh Bleecher Snyder authored
Instead of dumping information for the use to copy/paste into the issue tracker, open the issue tracker directly with a pre-filled template. Change-Id: I370d0063b609200497014ccda35244fa4314a662 Reviewed-on: https://go-review.googlesource.com/29210 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Robert Griesemer <gri@golang.org> Reviewed-by: Minux Ma <minux@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Matthew Dempsky authored
Change-Id: If8d27cf1cd8d650ed0ba332448d3174d80b6b0ca Reviewed-on: https://go-review.googlesource.com/29217 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Matthew Dempsky authored
Change-Id: Ifa7a882b020f7b0c9602c28c3e5567e5d4c39e73 Reviewed-on: https://go-review.googlesource.com/29216 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
-
Matthew Dempsky authored
Consistently use int16 for [ft]reg and int64 for [ft]offset. Change-Id: I7d279bb6e4fb735105429234a949074bf1cefb29 Reviewed-on: https://go-review.googlesource.com/29215 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
-
Matthew Dempsky authored
Change-Id: Ib05a8e149db8accdb1474703cd7b7004243d91d4 Reviewed-on: https://go-review.googlesource.com/29214 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Dave Cheney <dave@cheney.net>
-
Matthew Dempsky authored
These are no longer reachable as gins dispatches to ginscon for all arch-specific instructions anyway. Change-Id: I7f34883c16058308d8afa0f960dcf554af31bfe4 Reviewed-on: https://go-review.googlesource.com/29211Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Dave Cheney authored
After the removal of the old backend many types are no longer referenced outside internal/gc. Make these functions private so that tools like honnef.co/go/unused can spot when they become dead code. In doing so this CL identified several previously public helpers which are no longer used, so removes them. Change-Id: Idc2d485f493206de9d661bd3cb0ecb4684177b32 Reviewed-on: https://go-review.googlesource.com/29133 Run-TryBot: Dave Cheney <dave@cheney.net> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Dave Cheney authored
Remove unused functions spotted by honnef.co/go/unused. Change-Id: Iabf3b201215ce21e420a60f4ef2679b36231eda7 Reviewed-on: https://go-review.googlesource.com/29132 Run-TryBot: Dave Cheney <dave@cheney.net> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Keith Randall authored
It's not everything, but it is a good start. I tried to make the CL delete only. goimports forced a few exceptions to that rule. Update #16357 Change-Id: I041925cb2fe68bb7ae1617af862b22c48da649c1 Reviewed-on: https://go-review.googlesource.com/29168 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Martin Möhrmann <martisch@uos.de>
-
Keith Randall authored
Get rid of BlockCheck. Josh goaded me into it, and I went down a rabbithole making it happen. NilCheck now panics if the pointer is nil and returns void, as before. BlockCheck is gone, and NilCheck is no longer a Control value for any block. It just exists (and deadcode knows not to throw it away). I rewrote the nilcheckelim pass to handle this case. In particular, there can now be multiple NilCheck ops per block. I moved all of the arch-dependent nil check elimination done as part of ssaGenValue into its own proper pass, so we don't have to duplicate that code for every architecture. Making the arch-dependent nil check its own pass means I needed to add a bunch of flags to the opcode table so I could write the code without arch-dependent ops everywhere. Change-Id: I419f891ac9b0de313033ff09115c374163416a9f Reviewed-on: https://go-review.googlesource.com/29120 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
-
- 14 Sep, 2016 24 commits
-
-
Josh Bleecher Snyder authored
We will soon add dedicated builders for running vet/all. Their name will end with "-vetall". On those builders, run vet/all and nothing else. On all other builders, including local all.bash, don't run vet/all at all, because it is slow. This will probably be refined more over time. Change-Id: Ib1d0337adda84353931a325515c132068d4320cd Reviewed-on: https://go-review.googlesource.com/28962 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Matthew Dempsky authored
Passes toolstash -cmp. Fixes #16870. Change-Id: I70dc3bbb3cd3031826e5a54b96ba1ea603c282d1 Reviewed-on: https://go-review.googlesource.com/27910 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
-
Martin Möhrmann authored
when not instrumenting: - Intrinsify uses of slicebytetostringtmp within the runtime package in the ssa backend. - Pass OARRAYBYTESTRTMP nodes to the compiler backends for lowering instead of generating calls to slicebytetostringtmp. name old time/op new time/op delta ConcatStringAndBytes-4 27.9ns ± 2% 24.7ns ± 2% -11.52% (p=0.000 n=43+43) Fixes #17044 Change-Id: I51ce9c3b93284ce526edd0234f094e98580faf2d Reviewed-on: https://go-review.googlesource.com/29017 Run-TryBot: Martin Möhrmann <martisch@uos.de> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
-
Josh Bleecher Snyder authored
Change-Id: I63e7fd7f62aa80e1252b0c5b6c472439aa66da73 Reviewed-on: https://go-review.googlesource.com/29169Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Joe Tsai authored
It is unlikely that the value of UnixNano overflow in most use cases. However, the max date of 2262 is also within the range where it may be of concern to some users. Rather than have each person recompute when this overflows to validate if its okay for their use case, we just document it as within the years 1678 and 2262, for user's convenience. Fixes #16977 Change-Id: I4988738c147f4a6d30f8b8735c3941b75113bb16 Reviewed-on: https://go-review.googlesource.com/28478Reviewed-by: Andrew Gerrand <adg@golang.org>
-
Matthew Dempsky authored
Prepared with gofmt -r. Change-Id: Ib9f224cc20353acd9c5850dead1a2d32ca5427d3 Reviewed-on: https://go-review.googlesource.com/29165 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
Brad Fitzpatrick authored
No need to test so many sizes in race mode, especially for a package which doesn't use goroutines. Reduces test time from 2.5 minutes to 25 seconds. Updates #17104 Change-Id: I7065b39273f82edece385c0d67b3f2d83d4934b8 Reviewed-on: https://go-review.googlesource.com/29163Reviewed-by: David Crawshaw <crawshaw@golang.org>
-
Brad Fitzpatrick authored
No coverage is gained by running the 1e6 versions of the test over the 1e4 versions. It just adds 140 seconds of race overhead time. Updates #17104 Change-Id: I41408aedae34a8b1a148eebdda20269cdefffba3 Reviewed-on: https://go-review.googlesource.com/29159 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
-
Josh Bleecher Snyder authored
Found by vet. Change-Id: I9dbc6208ddbb5b407f4ddd20efbc166aac852cf7 Reviewed-on: https://go-review.googlesource.com/29162 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Josh Bleecher Snyder authored
CL 29110 brought the fix into the main tree. Change-Id: I7bf02670d40f22d35c63e05173419fdee9f93462 Reviewed-on: https://go-review.googlesource.com/29161 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Josh Bleecher Snyder authored
The standard sort swap method func (t T) Swap(i, j int) { t[i], t[j] = t[j], t[i] } uses no stack space on architectures for which FixedFrameSize == 0, currently 386 and amd64. Nevertheless, we insert a stack check prologue. This is because it contains a call to runtime.panicindex. However, for a few common runtime functions, we know at compile time that they require no arguments. Allow them to pass unnoticed. Triggers for 380 functions during make.bash. Cuts 4k off cmd/go. encoding/binary benchmarks: ReadSlice1000Int32s-8 9.49µs ± 3% 9.41µs ± 5% ~ (p=0.075 n=29+27) ReadStruct-8 1.50µs ± 3% 1.48µs ± 2% -1.49% (p=0.000 n=30+28) ReadInts-8 599ns ± 3% 600ns ± 3% ~ (p=0.471 n=30+29) WriteInts-8 836ns ± 4% 841ns ± 3% ~ (p=0.371 n=30+29) WriteSlice1000Int32s-8 8.84µs ± 3% 8.69µs ± 5% -1.71% (p=0.001 n=30+30) PutUvarint32-8 29.6ns ± 1% 28.1ns ± 3% -5.21% (p=0.000 n=28+28) PutUvarint64-8 82.6ns ± 5% 82.3ns ±10% -0.43% (p=0.014 n=27+30) Swap assembly before: "".T.Swap t=1 size=74 args=0x28 locals=0x0 0x0000 00000 (swap.go:5) TEXT "".T.Swap(SB), $0-40 0x0000 00000 (swap.go:5) MOVQ (TLS), CX 0x0009 00009 (swap.go:5) CMPQ SP, 16(CX) 0x000d 00013 (swap.go:5) JLS 67 0x000f 00015 (swap.go:5) FUNCDATA $0, gclocals·3cadd97b66f25a3a642be35e9362338f(SB) 0x000f 00015 (swap.go:5) FUNCDATA $1, gclocals·69c1753bd5f81501d95132d08af04464(SB) 0x000f 00015 (swap.go:5) MOVQ "".i+32(FP), AX 0x0014 00020 (swap.go:5) MOVQ "".t+16(FP), CX 0x0019 00025 (swap.go:5) CMPQ AX, CX 0x001c 00028 (swap.go:5) JCC $0, 60 0x001e 00030 (swap.go:5) MOVQ "".t+8(FP), DX 0x0023 00035 (swap.go:5) MOVBLZX (DX)(AX*1), BX 0x0027 00039 (swap.go:5) MOVQ "".j+40(FP), SI 0x002c 00044 (swap.go:5) CMPQ SI, CX 0x002f 00047 (swap.go:5) JCC $0, 60 0x0031 00049 (swap.go:5) MOVBLZX (DX)(SI*1), CX 0x0035 00053 (swap.go:5) MOVB CL, (DX)(AX*1) 0x0038 00056 (swap.go:5) MOVB BL, (DX)(SI*1) 0x003b 00059 (swap.go:5) RET 0x003c 00060 (swap.go:5) PCDATA $0, $1 0x003c 00060 (swap.go:5) CALL runtime.panicindex(SB) 0x0041 00065 (swap.go:5) UNDEF 0x0043 00067 (swap.go:5) NOP 0x0043 00067 (swap.go:5) CALL runtime.morestack_noctxt(SB) 0x0048 00072 (swap.go:5) JMP 0 Swap assembly after: "".T.Swap t=1 size=52 args=0x28 locals=0x0 0x0000 00000 (swap.go:5) TEXT "".T.Swap(SB), $0-40 0x0000 00000 (swap.go:5) FUNCDATA $0, gclocals·3cadd97b66f25a3a642be35e9362338f(SB) 0x0000 00000 (swap.go:5) FUNCDATA $1, gclocals·69c1753bd5f81501d95132d08af04464(SB) 0x0000 00000 (swap.go:5) MOVQ "".i+32(FP), AX 0x0005 00005 (swap.go:5) MOVQ "".t+16(FP), CX 0x000a 00010 (swap.go:5) CMPQ AX, CX 0x000d 00013 (swap.go:5) JCC $0, 45 0x000f 00015 (swap.go:5) MOVQ "".t+8(FP), DX 0x0014 00020 (swap.go:5) MOVBLZX (DX)(AX*1), BX 0x0018 00024 (swap.go:5) MOVQ "".j+40(FP), SI 0x001d 00029 (swap.go:5) CMPQ SI, CX 0x0020 00032 (swap.go:5) JCC $0, 45 0x0022 00034 (swap.go:5) MOVBLZX (DX)(SI*1), CX 0x0026 00038 (swap.go:5) MOVB CL, (DX)(AX*1) 0x0029 00041 (swap.go:5) MOVB BL, (DX)(SI*1) 0x002c 00044 (swap.go:5) RET 0x002d 00045 (swap.go:5) PCDATA $0, $1 0x002d 00045 (swap.go:5) CALL runtime.panicindex(SB) 0x0032 00050 (swap.go:5) UNDEF Change-Id: I57dad14af8aaa5e6112deac407cfadc2bfaf1f54 Reviewed-on: https://go-review.googlesource.com/24814 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
-
Josh Bleecher Snyder authored
cmd/cover, cmd/trace, and cmd/pprof all open browsers. 'go bug' will soon also open a browser. It is time to unify the browser-handling code. Change-Id: Iee6b443e21e938aeaaac366a1aefb1afbc7d9b2c Reviewed-on: https://go-review.googlesource.com/29160 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Brad Fitzpatrick authored
Don't benchmark so many sizes during the race builder's benchmark run. This package doesn't even use goroutines. Cuts off 10 seconds. Updates #17104 Change-Id: Ibb2c7272c18b9014a775949c656a5b930f197cd4 Reviewed-on: https://go-review.googlesource.com/29158Reviewed-by: David Crawshaw <crawshaw@golang.org>
-
Brad Fitzpatrick authored
Shave 6.5 minutes off the *-race build time. The *-race builders run: go test -short -race -run=^$ -benchtime=.1s -cpu=4 $PKG ... for each package with benchmarks. The point isn't to measure the speed of the packages, but rather to see if there are any races. (which is why a benchtime of 0.1 seconds is used) But running in race mode makes things slower and our benchmarks aren't all very fast to begin with. The regexp benchmarks in race were taking over 6.5 minutes. With this CL, it's now 8 seconds. Updates #17104 Change-Id: I054528d09b1568d37aac9f9b515d6ed90a5cf5b0 Reviewed-on: https://go-review.googlesource.com/29156 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Crawshaw <crawshaw@golang.org>
-
Keith Randall authored
Rip out the code that allows SSA to be used conditionally. No longer exists: ssa=0 flag GOSSAHASH GOSSAPKG SSATEST GOSSAFUNC now only controls the printing of the IR/html. Still need to rip out all of the old backend. It should no longer be callable after this CL. Update #16357 Change-Id: Ib30cc18fba6ca52232c41689ba610b0a94aa74f5 Reviewed-on: https://go-review.googlesource.com/29155 Run-TryBot: Keith Randall <khr@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org> Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
-
David Chase authored
It passed tests once, if anything's wrong, better to fail sooner than later. Change-Id: Ibb1c5db3f4c5535a4ff4681fd157db77082c5041 Reviewed-on: https://go-review.googlesource.com/28982 Run-TryBot: David Chase <drchase@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
-
Brad Fitzpatrick authored
The go_test_bench:* tests run: go test -short -race -run=^$ -benchtime=.1s -cpu=4 $PKG ... on each discovered package with any tests. (The same set used for the "go_test:*" tests) That set was 168 packages: $ go tool dist test -list | grep go_test: | wc -l 168 But only 76 of those have a "func Benchmark", and running each "go_test_bench:" test and compiling it in race mode, just to do nothing took 1-2 seconds each. So stop doing that and filter out the useless packages earlier. Now: $ go tool dist test -list -race | grep go_test_bench: | wc -l 76 Should save 90-180 seconds. (or maybe 45 seconds for trybots, since they're sharded) Updates #17104 Change-Id: I08ccb072a0dc0454ea425540ee8e74b59f83b773 Reviewed-on: https://go-review.googlesource.com/29153 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
Josh Bleecher Snyder authored
CL 26668 exposed #17065. Skip the cgo race tests on darwin for now. Updates #17065 Change-Id: I0ad0ce2ff1af6d515b8ce6184ddeabc49806950f Reviewed-on: https://go-review.googlesource.com/29077 Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
Jan Mercl authored
The change corrects the values of the largest float32 value (f1) and the value of the halfway point between f1 and the next, overflow value (f2). Fixes #17012 Change-Id: Idaf9997b69d61fafbffdb980d751c9857732e14d Reviewed-on: https://go-review.googlesource.com/29171Reviewed-by: Robert Griesemer <gri@golang.org>
-
David Crawshaw authored
For #17068 Change-Id: I4e3ab166f08100292b779b651a9acfbfb44a55cd Reviewed-on: https://go-review.googlesource.com/29119 Run-TryBot: David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
David Crawshaw authored
This gets -buildmode=pie -ldflags=-linkmode=internal working on Ubuntu 16.04. Fixes #17068 Change-Id: Ice5036199005fb528cc58279a7f057170dc6b73d Reviewed-on: https://go-review.googlesource.com/29118 Run-TryBot: David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com> Reviewed-by: Ian Lance Taylor <iant@golang.org>
-
David Stainton authored
Fixes #16681 Change-Id: I6ff7ec81fe48ab06be3aae5b7ff92e9dc70960c3 Reviewed-on: https://go-review.googlesource.com/26990 Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>
-
Emmanuel Odeke authored
Updates #16360. Adds examples for: + Chmod + Chtimes + FileMode Change-Id: I1b61ee0392fa3774593a7f36aaf0fa1e484c778b Reviewed-on: https://go-review.googlesource.com/28963 Run-TryBot: Andrew Gerrand <adg@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Andrew Gerrand <adg@golang.org>
-
Aaron Jacobs authored
While you could argue the previous wording technically said that -1 is an acceptable way to indicate "unknown" on the client, it could be read as ambiguous. Now it's clear that both 0 and -1 mean unknown. Change-Id: I3bc5a3fd5afd1999e487296ec121eb548415e6b0 Reviewed-on: https://go-review.googlesource.com/29130Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-
- 13 Sep, 2016 3 commits
-
-
David Crawshaw authored
Created by running 'go generate'. That made debugging fun today. Change-Id: I9ffe00877851f2b198275220ad6058b9005daa72 Reviewed-on: https://go-review.googlesource.com/29117Reviewed-by: Ian Lance Taylor <iant@golang.org> Run-TryBot: Ian Lance Taylor <iant@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
David Crawshaw authored
Change-Id: I11899096c71ee0e24e902c87914601fcd7ffd7a9 Reviewed-on: https://go-review.googlesource.com/28967Reviewed-by: Ian Lance Taylor <iant@golang.org> Run-TryBot: Ian Lance Taylor <iant@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
-
Matthew Dempsky authored
Exported is no longer used since removing the text-format exporter, and Safe is only used within importfile so it can be made into a local variable. Change-Id: I92986f704d7952759c79d9243620a22c24602333 Reviewed-on: https://go-review.googlesource.com/29115 Run-TryBot: Matthew Dempsky <mdempsky@google.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
-