...to avoid executing the supportAVX2 assembly function.
Also, combine all no-op intrinsics into a single function.
Change-Id: Ic65a80d3a6df52c3850c34406f034781057a0991
Reviewed-on: https://go-review.googlesource.com/28711
Reviewed-by: Robert Griesemer <gri@golang.org>
Both functions had documentation referring to BuildAll when the
function is now called Build.
Change-Id: I59cce397a0a72bf7fa36f9798e2b07bb6b1da726
Reviewed-on: https://go-review.googlesource.com/25084
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
...now that gc's export data records fewer dependencies.
Change-Id: Ie0f68f1f7b0825f6d2c100dfe189be017383db60
Reviewed-on: https://go-review.googlesource.com/22584
Reviewed-by: Robert Griesemer <gri@golang.org>
...which is used by $GOROOT/test/goprint.go.
Change-Id: I4626b8fae3f87d9c8dd8cdcd8c05036955a36262
Reviewed-on: https://go-review.googlesource.com/22560
Reviewed-by: Robert Griesemer <gri@golang.org>
Per https://groups.google.com/forum/#!topic/golang-announce/qu_rAphYdxY
this change deletes the packages
go/exact
go/gccgoimporter
go/gcimporter
go/importer
go/types
cmd/vet
from the x/tools repo and any files depending on those packages
building against Go 1.4.
x/tools packages depending on any of these libraries must use the
respective versions from the std lib or use vendored versions if
building against 1.4.
Remaining packages may or may not build against Go 1.4 anymore
and will not be supported against 1.4.
Change-Id: I1c655fc30aee49b6c7326ebd4eb1bb0836ac97e0
Reviewed-on: https://go-review.googlesource.com/20810
Reviewed-by: Alan Donovan <adonovan@google.com>
This fixes some print calls with wrong format directives. Additionally,
struct initialisers were changed to use keyed fields, purely to reduce
the amount of noise generated by go vet.
Change-Id: Ib9f6fd8f2dff7ce84826478de0ba83dda9746270
Reviewed-on: https://go-review.googlesource.com/21180
Reviewed-by: Ian Lance Taylor <iant@golang.org>
The BuilderModeFlag function does not work with flag packages (such as
Google's) that define their own FlagSet. This change exposes a type
that satisifies flag.Value insted of a function. That type
is *BuilderMode; there's no need for a separate wrapper type.
Change-Id: I8095b80de499e3c52a29a5c1996d1b1fe3799358
Reviewed-on: https://go-review.googlesource.com/20330
Reviewed-by: Michael Matloob <matloob@golang.org>
A few files have been forked and tagged "go1.5,!go1.6" to work around
minor API changes between the two types packages:
- constant.Value.String() in oracle/describe.go and its tests;
- constant.ToInt must now be called before constant.Int64Val.
- types.Config{Importer: importer.Default()} in a number of places
- go/types/typeutil/import_test.go uses lowercase names to avoid 'import "C"'.
Files in go/types/typesutil, missing from my previous CL, have been
tagged !go1.5; these files will be deleted in February.
All affected packages were tested using 1.4.1, 1.5, and ~1.6 (tip).
Change-Id: Iec7fd370e1434508149b378438fb37f65b8d2ba8
Reviewed-on: https://go-review.googlesource.com/18207
Reviewed-by: Robert Griesemer <gri@golang.org>
This change will ensure that the tree continues to work with go1.4.1.
All files continue to depend on golang.org/x/tools/go/types, but in a
follow-up change, I will switch the primary files to depend on the
standard go/types package. Another (smaller) set of files will be
forked and tagged, this time !1.6, due to API differences between the
two packages.
All tests pass using 1.4.1, 1.5, and ~1.6 (tip).
Change-Id: Ifd75a6330e120957d646be91693daaba1ce0e8c9
Reviewed-on: https://go-review.googlesource.com/18333
Reviewed-by: Robert Griesemer <gri@golang.org>
1. Added intrinsic for os.Pipe to avoid use of "unsafe". This makes
Example tests work in the interpreter, fixing the broken "unicode"
package (which recently sprouted some Examples) and allowing us to
enable tests for:
encoding/csv
flag
path
text/scanner
unicode/utf8
The implementation of os.Pipe may require tweaking for non-Linux platforms.
2. Added intrinsic for strings.Index to avoid assembly dependency.
3. Added dumb intrinsic for testing۰runExample to work around a bug in
go/ssa/testmain.go that requires an invasive fix
(Until then, the output of Example functions will not be checked.)
Change-Id: I6374c9c47aa802275b7cdc98525e057f5db0615a
Reviewed-on: https://go-review.googlesource.com/18191
Reviewed-by: Alan Donovan <adonovan@google.com>
Since Go 1.5, if a key expression in a map literal is itself a composite
literal, the type may be omitted. An & operation may be implied,
but was not generated, resulting in a type mismatch crash.
Added regression test.
Also, added missing intrinsic math.hasSSE4 to interpreter.
Fixes issue 13341
Change-Id: I6a5e560460cdcac295a500a3fc3738cbe8aa5853
Reviewed-on: https://go-review.googlesource.com/17123
Reviewed-by: Robert Griesemer <gri@golang.org>
This would cause oracle callee lookup failures when the RHS is an
invoke-style call.
Fixesgolang/go#12999
Change-Id: Ifd561c4e7bf26f57ace5f62afac746b926c70993
Reviewed-on: https://go-review.googlesource.com/16210
Reviewed-by: Alan Donovan <adonovan@google.com>
This adds a missing memory barrier at the end of Package.Build.
Change-Id: Ife35d5ad5a48ba121f35656fef682863d4f2aef6
Reviewed-on: https://go-review.googlesource.com/14761
Reviewed-by: Robert Griesemer <gri@golang.org>
until vendoring support is added (soon).
Change-Id: I954514b5c6b82622c0c831a289ccdcc16dbd37f8
Reviewed-on: https://go-review.googlesource.com/14130
Reviewed-by: Alan Donovan <adonovan@google.com>
The optional Qualifier function determines what prefix to attach to
package-level names, enabling clients to qualify packages in different
ways, for example, using only the package name instead of its complete
path, or using the locally appropriate name for package given a set of
(possibly renaming) imports.
Prior to this change, clients wanting this behavior had to copy
hundreds of lines of complex printing logic.
Fun fact: (*types.Package).Path and (*types.Package).Name are valid
Qualifier functions.
We provide the RelativeTo helper function to create Qualifiers so that
the old behavior remains a one-liner.
Fixesgolang/go#11133
Change-Id: Ibd63f639c7b3aa1738826d6165f2d810efeb8293
Reviewed-on: https://go-review.googlesource.com/11692
Reviewed-by: Robert Griesemer <gri@golang.org>
since we can't interpret os.Pipe yet.
Change-Id: If27dadc18532274ce97ad7e7557e8614dd15279e
Reviewed-on: https://go-review.googlesource.com/11712
Reviewed-by: Robert Griesemer <gri@golang.org>
since they use quick.Check, which requires the unimplemented
(reflect.Value).SetString.
Also:
- Add reflect.Type.{In,NumIn} methods, whose absence
was only the proximate cause of the failed test.
- Delete bodies of reflect.Value methods so that it's obvious a
function that should be intrinsic is missing.
Change-Id: Ib64b8f4953a913f4ead90e376bda70419adb87cb
Reviewed-on: https://go-review.googlesource.com/10796
Reviewed-by: Robert Griesemer <gri@golang.org>
Not sure how it ended up there...
Change-Id: I0d48025fd6595714b0c52cadf305b3e06e2b9b84
Reviewed-on: https://go-review.googlesource.com/9171
Reviewed-by: David Crawshaw <crawshaw@golang.org>
The loader package now loads, parses, and type-checks a whole program
from source, and that is all.
Also:
- simplified loader logic
- ssa.Create is gone; use ssautil.CreateProgram.
- ssautil.LoadPackage renamed to BuildPackage.
It is now independent of go/types' Import hook and the Packages map.
- ssadump: -importbin flag removed.
The value of this flag was that it caused the tool to print IR
for only a single package; this is now the normal behaviour.
Fixes#9955
Change-Id: I4571118258ab1a46dccece3241b7dc51401a3acc
Reviewed-on: https://go-review.googlesource.com/8953
Reviewed-by: Robert Griesemer <gri@golang.org>
Remove all dependencies from non-test code in go/ssa to go/loader,
except the deprecated Create function which will be eliminated in
favor of ssautil.CreateProgram in a mechnanical followup.
Add Examples of two main use cases of SSA construction:
loading a complete program from source; and
building a single package, loading its dependencies from import data.
Add tests to ssautil of the two load functions.
Suggestions welcome for better names.
Planned follow-ups:
- replace all references to ssa.Create with ssautil.CreateProgram and eliminate it.
- eliminate support in go/loader for the ImportBinary flag, and the
PackageCreated hook which is no longer needed since clients can
create the package themselves (see Example).
Step 1 to fixing issue 9955.
Change-Id: I4e64df67fcd5b7f0c0388047e06cea247fddfec5
Reviewed-on: https://go-review.googlesource.com/8669
Reviewed-by: Robert Griesemer <gri@golang.org>
...otherwise the interpreted program can make the whole test suite
slow. Just ignore the argument and return the current GOMAXPROCS
value.
Change-Id: Ife2ad6c53e6fdf9feea1d1b231d8d796b3db3a24
Also: add missing intrinsic for os.runtime_beforeExit.
Reviewed-on: https://go-review.googlesource.com/8591
Reviewed-by: Robert Griesemer <gri@golang.org>
Composite literals are initialized in place where possible, but in
cases the initializer expression refers to the variable that
is being updated
x = T{a: x.a}
we must ensure that the RHS is fully evaluated before we execute any
stores to x. This means we need to record the sequence of stores in a
"store buffer" and execute it only once the entire composite literal
has been evaluated.
Fixes issue #10127
Change-Id: If94e3b179beb25feea5b298ed43de6a199aaf347
Reviewed-on: https://go-review.googlesource.com/7533
Reviewed-by: Robert Griesemer <gri@golang.org>
Since all SSA values are immutable, no value copying is required for
any operations except those that load from or store to a variable;
those operations must do an aggregate copy, i.e., descend into struct
and array elements. All other calls to copyVal have been removed;
they were pieces of duct tape, as I had long suspected.
The descent must be based on the static type information, not the
"shape" of the dynamic value, since two reflect.Value structs may have
different internal shapes. We clobber the true definition of
reflect.Value's underlying type, replacing it with struct{interface{},
interface{}}, which is close enough to make the load/store functions
work.
+ Test
Change-Id: I5e239d91ed0cb2a669a9f75766024fe1f9a5c347
Reviewed-on: https://go-review.googlesource.com/7532
Reviewed-by: Robert Griesemer <gri@golang.org>
Before this change, declared init functions were not package members;
this choice dates from when go/types did not create Func objects for them.
Now, they have an Object. They appear in Members, keyed by "init#%d"
(sequence number) for uniqueness. They can be enumerated. They can
be looked up from a *types.Func via (*Program).FuncValue.
Caveat: fn.Object.Name() no longer equals fn.Name() in all cases.
NB: incompatible API change! (Your build will not break though.)
Change-Id: I2de873079fd57329e6c2f55a282940f6699a77a1
Reviewed-on: https://go-review.googlesource.com/6950
Reviewed-by: Robert Griesemer <gri@golang.org>
Reviewed-by: Peter Collingbourne <pcc@google.com>
...since the zero value is more useful by far.
This is a breaking API change, obviously. (One or two tests in this
CL have intentional been left using the zero value, i.e., they now
load source.)
Change-Id: I42287bfcdb1afef8ee84e5eac12534dd0a1fd5d2
Reviewed-on: https://go-review.googlesource.com/5653
Reviewed-by: Robert Griesemer <gri@golang.org>
ssadump's -build=G option is now spelled -import; it was never related to ssa.
Change-Id: Ic21cd8b6990c0ffd25651c17a842a63bfa5019cf
Reviewed-on: https://go-review.googlesource.com/5172
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Before this change, many kinds of error would cause the loader to stop.
making it brittle when analyzing large codebases, as in "godoc -analysis".
This change moves operations that used to occur during
configuration---(*build.Context).Import, loading, and parsing of
initial packages---into the Load call, and ensures that all failures
during Loading are reported at the end so that the maximum amount of
progress is made.
Also: redesign the tests and add many new cases.
Change-Id: Ia8cd99416af7c5d4a5fe133908adfa83676d401f
Reviewed-on: https://go-review.googlesource.com/3626
Reviewed-by: Robert Griesemer <gri@golang.org>
And log its value in godoc -analysis.
Related to issue 8968
Change-Id: I96a96922a3fa5c434c69e0faff1cc8ec4686b6f2
Reviewed-on: https://go-review.googlesource.com/3154
Reviewed-by: Robert Griesemer <gri@golang.org>
Details:
- rename (*Program).TypesWithMethodSets() to RuntimeTypes()
- delete (*Package).TypesWithMethodSets() method and simplify
- move code to methods.go
- update test to use
1-2% improvement in space and time (though I barely trust this data
because the GC at tip is in such terrible state).
Change-Id: I38eab78b11e0ad0ff16e0530e775b6ff6a2ab246
Reviewed-on: https://go-review.googlesource.com/3148
Reviewed-by: Robert Griesemer <gri@golang.org>
...to avoid namespace conflicts.
Also make its name "main", since it defines func main().
And fix 2 typos.
Change-Id: I7cf7894d6bed134907b3d2742255e5a82426071b
Reviewed-on: https://go-review.googlesource.com/3150
Reviewed-by: Robert Griesemer <gri@golang.org>
+ Tests:
Extend function name uniqueness check to exported wrappers.
Check that expected synthetic functions are created, reachable, and not duplicated.
Change-Id: I0e87ebb2712e33e1f49da3fa9a9dde0085bf3850
Reviewed-on: https://go-review.googlesource.com/2013
Reviewed-by: Robert Griesemer <gri@golang.org>
+ test
Change-Id: Ie37835577ffcdd764cf6a0b611e02f04386755cf
Reviewed-on: https://go-review.googlesource.com/1580
Reviewed-by: Robert Griesemer <gri@golang.org>
...like we do for "runtime" functions, so that they fail informatively
if executed. They all need intrinsics, but only some are yet defined.
Also:
- added test for issue 9462
- "BUG" in test output is now a failure in all tests (not just $GOROOT tests)
- added intrinsic for reflect.SliceOf
- show dynamic type of panic value
Fixes issue 9462
Change-Id: I3a504c7faeed81e922fedc7dd59222717f3a7e95
Reviewed-on: https://go-review.googlesource.com/2145
Reviewed-by: Robert Griesemer <gri@golang.org>
Rewrite performed with this command:
sed -i '' 's_code.google.com/p/go\._golang.org/x/_g' \
$(grep -lr 'code.google.com/p/go.' *)
LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/170920043
The needMethods cache logic was wrong: it would treat any
previous call as a cache hit, even if 'skip' was true for that
call. As a result it could fail to generate methods for some
'skip' types, i.e. anonymous structs.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/144750043
This CL is the first refactoring automated by "gorename". :)
Though I had to update the comments and run 'hg gofmt'. :(
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/142930043
It now has a main() function, which was confusing the logic to find the entry point.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/142860044
(godoc is excluded from this CL since it will continue to use
/src/pkg in its URL namespace, making the necessary cleanup
more subtle.)
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/141770043
Recent changes in the runtime caused the interpreter to call 'getg',
and it should never have gotten that far.
Also, delete bodies of "runtime" functions, since they're too magical.
This makes missing intrinsics cause very obvious failures.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/135330043
LookupFieldOrMethod now also decides whether a found
method is actually in the method set. Simplifies call
sites. Added corresponding API tests.
TODO (separate CL): Decide what the correct value for
the indirect result should be (as required for code
generation). For now, the result value for indirect
is unchanged from before if a field/method is found.
Fixesgolang/go#8584.
LGTM=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/132260043
(1) support Example functions defined in programs that don't
import "testing". We emit code to testmain.main() to call
them directly, since we can't call testing.Main.
(2) expose a FindTests function which reports the set of
Test, Example and Benchmark functions it finds.
Certain clients need this.
Added test for logic in FindTests.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/115290048
- Break out parts of coverage.go into more specific files.
- Re-enable test of nil interface-to-interface conversion.
- Update initorder test to reflect spec ambiguity and gc vs go/types variance.
- Re-enable test dependent on now-fixed bug 8189 ("value,ok" yields an untyped bool)
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/119530043
And serialize the printing of each item with a mutex.
It is the formatted output of this tool, after all.
Also: minor doc tweaks.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/114620044
Examples:
- "foo$1" becomes "pkg.foo$1"
- "init$1" (meaning the first declared "init" function) becomes "init#1",
to distinguish it from "init$1" (meaning the first anonymous function
within the synthetic "init" function that initializes package-level vars).
It is now an invariant that all source-level (non-synthetic)
functions have distinct names, and that all names include the
enclosing package. Added test for this.
+ updated various clients.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/122750043
It was missing from the four conversions and the Make* instructions.
(Experiments with pure bytes.Buffer-based printing were not faster; various TODOs removed.)
LGTM=crawshaw
R=gri, crawshaw
CC=golang-codereviews
https://golang.org/cl/58040043
Users need only add an extra file to the package to specify
additional imports and initialization steps in testmain, to
match their build system.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/120090043
If an expression is addressable, we compute its address then load, rather than
extracting the value of subelements. For aggregates this avoids large copies.
Example:
var x [2]struct{y [3]int}
print(x[1].y[2])
Was:
t0 = local [3]struct{x [5]int} (x) *[3]struct{x [5]int}
t1 = *t0 [3]struct{x [5]int}
t2 = t1[1:int] struct{x [5]int}
t3 = t2.x [#0] [5]int
t4 = t3[2:int] int
Now:
t1 = &t0[1:int] *struct{x [5]int}
t2 = &t1.x [#0] *[5]int
t3 = &t2[2:int] *int
t4 = *t3 int
Also:
- make emitFieldSelections responsible for calling emitDebugRef, as
one of its two calls was forgetting to do it.
- relax the specification of (*Program).VarValue because not all
subexpressions are materalized as values now.
- fix up the objlookup.go test expectations to match.
go/ssa/interp test runs 10% faster.
Thanks to Peter Collingbourne for pointing this out.
LGTM=pcc
R=pcc, gri
CC=golang-codereviews
https://golang.org/cl/109710043
PackageInfo:
- deleted IsType
- inlined + deleted: ValueOf, TypeCaseVar, ImportSpecPkg
- on failure, TypeOf accessor now returns nil (was: panic)
go/ssa: avoid extra map lookups by using Uses or Defs directly when safe to do so,
and keeping the TypeAndValue around in expr0().
LGTM=gri
R=gri, pcc
CC=golang-codereviews
https://golang.org/cl/107650043
This extends the sanity checker to identify and report referrers
which do not appear in the function's instruction lists, and fixes two
bugs in the lifting algorithm which were caught by the sanity check.
LGTM=adonovan
R=adonovan
CC=axwalk, golang-codereviews
https://golang.org/cl/110210045
Bare init functions omit calls to dependent init functions and the
use of an init guard. They are useful in cases where the client uses
a different calling convention for init functions, or cases where
it is easier for a client to analyze bare init functions.
LGTM=adonovan
R=adonovan
CC=golang-codereviews, iant
https://golang.org/cl/78780043
Previously, statements such as:
type T struct { a, b int }
[...]
x = T{}
x = T{b: 1}
would only affect the aggregate members mentioned in the composite
literal and leave the other members unchanged. This change causes us
to write a zero value to the target in cases where the target is not
already known to hold a zero value and the number of initializers in
the composite literal differs from the number of elements in its type.
Author: Peter Collingbourne. (hg clpatch got confused)
LGTM=pcc
R=pcc
CC=golang-codereviews
https://golang.org/cl/107980045
Also, define ssa:wrapnilchk intrinsic to check and gracefully
fail when a T method is dynamically invoked via a nil *T receiver.
+ Test.
A follow-up CL will add another intrinsic, ssa:memclr.
+ minor cleanups.
LGTM=gri
R=gri
CC=golang-codereviews
https://golang.org/cl/101170044
It is easier for clients to recover from panics if the recover block
is always present. Otherwise, the client has to work around the lack
of a recover block by synthesizing a zero value return.
LGTM=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/87210044
Blocks dominated by "if false" should be retained in the
initial SSA form so they remain visible to subsequent source
code analysis tools.
In any case, true compilers already need a stronger version of
this optimization so they can simplify CFGs such as this:
const x, y = ...
switch x {case y:...}
where a branch is constant but the comparison of constants
does not occur within an expression.
LGTM=gri
R=gri
CC=golang-codereviews, pcc
https://golang.org/cl/101250043
The SSA builder shouldn't be in the business of
interprocedural optimization, especially in the presence of
concurrency.
This causes the instruction count to increase by 0.03%.
LGTM=gri
R=gri, pcc
CC=golang-codereviews
https://golang.org/cl/105020045
Also:
- extend Parent() to all Values and add to interface:
(Builtin/Const/Global => nil; Function => Enclosing)
- hide Function.Enclosing since it's now redundant wrt Parent()
- make (*Function).String robust for synthetics without pkg object
LGTM=gri
R=gri
CC=golang-codereviews, khr
https://golang.org/cl/87580044
Until now, the same Function was used to represent a method
(T)func() and the "method expression" function func(T) formed
from it. So the SSA code for this:
var buf bytes.Buffer
f := Buffer.Bytes
f(buf)
buf.Bytes()
would involve an implicit cast (ChangeType) on line 2.
However, compilers based on go/ssa may want to use different
calling conventions for them, like gccgo does (see issue
7839). This change decouples them by using an anonymous
function called a "thunk", rather like this:
f := func(r *bytes.Buffer) []byte { return r.Bytes() }
Thunks are similar to method wrappers; both are created by
makeWrapper.
"Interface method wrappers" were a special case of thunks for
direct calls (no indirection/fields) of interface methods.
They are now subsumed by thunks and have been deleted. Now
that only the needed thunks are built, we don't need to
populate the concrete method sets of interface types at all,
so (*Program).Method and LookupMethod return nil for them.
This results in a slight reduction in function count (>1%) and
instruction count (<<1%).
Details:
go/ssa:
- API: ChangeType no longer supports func/method conversions.
- API: (*Program).FuncValue now returns nil for abstract
(interface) methods.
- API: (*Function).RelString simplified.
"$bound" is now a suffix not a prefix, and the receiver
type is rendered package-relative.
- API: Function.Object is now defined for all wrappers too.
- API: (*Program).Method and LookupMethod return nil for
abstract methods.
- emitConv no longer permits (non-identical)
Signature->Signature conversions. Added assertion.
- add and use isInterface helper
- sanity: we check packages after Build, not Create, otherwise
cross-package refs might fail.
go/pointer:
- update tests for new function strings.
- pointer_test: don't add non-pointerlike probes to analysis.
(The error was checked, but too late, causing a panic.)
- fixed a minor bug: if a test probe print(x) was the sole
reference to x, no nodes were generated for x.
- (reflect.Type).MethodByName: updated due to ssa API changes.
Also, fixed incorrect testdata/funcreflect.go expectation
for MethodByName on interfaces.
oracle:
- fix for new FuncValue semantics.
- a "pointsto" query on an I.f thunk now returns an error.
Fixesgolang/go#7839
LGTM=gri
R=gri
CC=golang-codereviews, pcc
https://golang.org/cl/93780044
- Replaced check.initDependencies with check.initOrder;
this is the only semantic change, it affects only the
value of Info.InitOrder.
- Added additional init order test cases and adjusted
existing tests.
- Moved orderedSetObjects from resolver.go to ordering.go.
Fixesgolang/go#7964.
LGTM=adonovan
R=adonovan
CC=golang-codereviews
https://golang.org/cl/91450043
Alan Donovan
Tim Henderson
Robert Griesemer
Dominik Honnef
Brad Fitzpatrick
Daniel Morsing
Michal Cierniak
Peter Collingbourne
David Symonds
Andrew Gerrand