The seq serialization machinery is a historic artifact from when Go
mobile code had to run in a separate process. Now that Go code is running
in-process, replace the explicit serialization with direct calls and pass
arguments on the stack.
The benefits are a much smaller bind runtime, much less garbage (and, in
Java, fewer objects with finalizers), less argument copying, and faster
cross-language calls.
The cost is a more complex generator, because some of the work from the
bind runtime is moved to generated code. Generated code now handles
conversion between Go and Java/ObjC types, multiple return values and memory
management of byte slice and string arguments.
To overcome the lack of calling C code between Go packages, all bound
packages now end up in the same (fake) package, "gomobile_bind", instead of
separate packages (go_<pkgname>). To avoid name clashes, the package name is
added as a prefix to generated functions and types.
Also, don't copy byte arrays passed to Go, saving call time and
allowing read([]byte)-style interfaces to foreign callers (#12113).
Finally, add support for nil interfaces and struct pointers to objc.
This is a large CL, but most of the changes stem from changing testdata.
The full benchcmp output on the CL/20095 benchmarks on my Nexus 5 is
reproduced below. Note that the savings for the JavaSlice* benchmarks are
skewed because byte slices are no longer copied before passing them to Go.
benchmark old ns/op new ns/op delta
BenchmarkJavaEmpty 26.0 19.0 -26.92%
BenchmarkJavaEmptyDirect 23.0 22.0 -4.35%
BenchmarkJavaNoargs 7685 2339 -69.56%
BenchmarkJavaNoargsDirect 17405 8041 -53.80%
BenchmarkJavaOnearg 26887 2366 -91.20%
BenchmarkJavaOneargDirect 34266 7910 -76.92%
BenchmarkJavaOneret 38325 2245 -94.14%
BenchmarkJavaOneretDirect 46265 7708 -83.34%
BenchmarkJavaManyargs 41720 2535 -93.92%
BenchmarkJavaManyargsDirect 51026 8373 -83.59%
BenchmarkJavaRefjava 38139 21260 -44.26%
BenchmarkJavaRefjavaDirect 42706 28150 -34.08%
BenchmarkJavaRefgo 34403 6843 -80.11%
BenchmarkJavaRefgoDirect 40193 16582 -58.74%
BenchmarkJavaStringShort 32366 9323 -71.20%
BenchmarkJavaStringShortDirect 41973 19118 -54.45%
BenchmarkJavaStringLong 127879 94420 -26.16%
BenchmarkJavaStringLongDirect 133776 114760 -14.21%
BenchmarkJavaStringShortUnicode 32562 9221 -71.68%
BenchmarkJavaStringShortUnicodeDirect 41464 19094 -53.95%
BenchmarkJavaStringLongUnicode 131015 89401 -31.76%
BenchmarkJavaStringLongUnicodeDirect 134130 90786 -32.31%
BenchmarkJavaSliceShort 42462 7538 -82.25%
BenchmarkJavaSliceShortDirect 52940 17017 -67.86%
BenchmarkJavaSliceLong 138391 8466 -93.88%
BenchmarkJavaSliceLongDirect 205804 15666 -92.39%
BenchmarkGoEmpty 3.00 3.00 +0.00%
BenchmarkGoEmptyDirect 3.00 3.00 +0.00%
BenchmarkGoNoarg 40342 13716 -66.00%
BenchmarkGoNoargDirect 46691 13569 -70.94%
BenchmarkGoOnearg 43529 13757 -68.40%
BenchmarkGoOneargDirect 44867 14078 -68.62%
BenchmarkGoOneret 45456 13559 -70.17%
BenchmarkGoOneretDirect 44694 13442 -69.92%
BenchmarkGoRefjava 55111 28071 -49.06%
BenchmarkGoRefjavaDirect 60883 26872 -55.86%
BenchmarkGoRefgo 57038 29223 -48.77%
BenchmarkGoRefgoDirect 56153 27812 -50.47%
BenchmarkGoManyargs 67967 17398 -74.40%
BenchmarkGoManyargsDirect 60617 16998 -71.96%
BenchmarkGoStringShort 57538 22600 -60.72%
BenchmarkGoStringShortDirect 52627 22704 -56.86%
BenchmarkGoStringLong 128485 52530 -59.12%
BenchmarkGoStringLongDirect 138377 52079 -62.36%
BenchmarkGoStringShortUnicode 57062 22994 -59.70%
BenchmarkGoStringShortUnicodeDirect 62563 22938 -63.34%
BenchmarkGoStringLongUnicode 139913 55553 -60.29%
BenchmarkGoStringLongUnicodeDirect 150863 57791 -61.69%
BenchmarkGoSliceShort 59279 20215 -65.90%
BenchmarkGoSliceShortDirect 60160 21136 -64.87%
BenchmarkGoSliceLong 411225 301870 -26.59%
BenchmarkGoSliceLongDirect 399029 298915 -25.09%
Fixesgolang/go#12619Fixesgolang/go#12113Fixesgolang/go#13033
Change-Id: I2b45e9e98a1248e3c23a5137f775f7364908bec7
Reviewed-on: https://go-review.googlesource.com/19821
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
Address comment from CL/20095.
Change-Id: I2b3b3230106ad27128440609472003c69bd97825
Reviewed-on: https://go-review.googlesource.com/20173
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
Since the normal Go benchmark machinery cannot readily be used for
Android apps, a new test, TestJavaSeqBench, is added that builds and
runs the new benchmarkpkg package along with its support Java class
SeqBench. Benchmarkpkg mimics Go benchmarking, in particular it
produces benchcmp compatible output.
Excerpts of the output from a Nexus 5:
BenchmarkJavaEmpty 65536000 26 ns/op
BenchmarkJavaNoargs 256000 7685 ns/op
BenchmarkJavaNoargsDirect 64000 17405 ns/op
BenchmarkJavaOnearg 64000 26887 ns/op
BenchmarkJavaOneret 32000 38325 ns/op
BenchmarkJavaManyargs 32000 41720 ns/op
BenchmarkJavaRefjava 32000 38139 ns/op
BenchmarkJavaRefgo 32000 34403 ns/op
BenchmarkJavaStringShort 32000 32366 ns/op
BenchmarkJavaStringLong 8000 127879 ns/op
BenchmarkJavaSliceShort 32000 42462 ns/op
BenchmarkJavaSliceLong 8000 138391 ns/op
BenchmarkGoEmpty 524288000 3 ns/op
BenchmarkGoNoarg 32000 40342 ns/op
BenchmarkGoOnearg 32000 43529 ns/op
BenchmarkGoOneret 32000 45456 ns/op
BenchmarkGoRefjava 32000 55111 ns/op
BenchmarkGoRefgo 32000 57038 ns/op
BenchmarkGoManyargs 16000 67967 ns/op
BenchmarkGoStringShort 32000 57538 ns/op
BenchmarkGoStringLong 8000 128485 ns/op
BenchmarkGoSliceShort 32000 59279 ns/op
BenchmarkGoSliceLong 4000 411225 ns/op
Benchmarks prefixed with "BenchmarkJava" are for calls from Java into
Go. Benchmarks prefixed with "BenchmarksGo" are the other way around.
Note that all Go benchmarks run against a Java interface implementation
while the Java benchmarks calls Go functions directly. In other words,
every Go call serializes an implicit Java reference, explaining the
higher call times. The JavaRefgo and JavaRefjava tests attempt to
quantify the overhead equivalent for Java.
The "Direct" suffix are for variants that runs the benchmarks from a
new thread or goroutine. For Go it makes little difference, but there
is a noticable speedup when calling Go from Java when there is already
a JNI call context earlier in the stack.
The benchmarks are for Android only for now, but the benchmarkpkg
has been added to the common golang.org/x/mobile/bind package in
anticipation of future iOS support.
Change-Id: I3c948dc710b65bc348e7635416324095060a5beb
Reviewed-on: https://go-review.googlesource.com/20095
Reviewed-by: David Crawshaw <crawshaw@golang.org>
RefMap tracks its number of live Ref instances in the 'live' member.
However, when a reference was removed and later added, 'live' wasn't
updated accordingly. Fix and add a test.
Change-Id: I806e17ea0319d76db4d07b5f8d9107b146ee80db
Reviewed-on: https://go-review.googlesource.com/19975
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
On some Android devices (my HTC One S running Android 4.1.1),
SeqTest failed the testAssets test because the LoadJNI hack to
locate a valid context fails.
Instead, make testAssets set up a valid context acquired from
InstrumentTestCase.
Change-Id: If6e11173dbacff45eb6cb0f409f56cbd88186e30
Reviewed-on: https://go-review.googlesource.com/19896
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Avoid taking a good name (seq) away from users.
Fixesgolang/go#14168
Change-Id: I88e90cb74b479e348c642a1caa27096ed4a6d68e
Reviewed-on: https://go-review.googlesource.com/19601
Reviewed-by: David Crawshaw <crawshaw@golang.org>
The existing implementation has memory leaks on two local variables
that are not deleted after use. Android app will crash after this issue.
add UnitTest for this issue.
Fixesgolang/go#14346
Change-Id: Ic233d15556ac97b35e00e13279a572c48a03049f
Reviewed-on: https://go-review.googlesource.com/19532
Reviewed-by: Elias Naur <elias.naur@gmail.com>
The Seq Java class has a special case for null references. Expand
the special case to Go so that null references from Java are properly
translated to nil.
Fixesgolang/go#14228
Change-Id: I915d1f843c9db299d6910480f6d10dae0121a3b4
Reviewed-on: https://go-review.googlesource.com/19460
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Java methods from Go are run on a thread pool managed on the Java side,
to avoid the complexities of getting Go threads to play along with the
Android JVM. However, for call stacks that contain a Java->Go->Java
chain, this behaviour confuses Java code sensitive to specific threads
if the Go->Java call is executed on an arbitrary thread from the pool.
For example, most Android UI changes must happen on the single UI
thread.
Replace the thread pool with direct calls to mimic ObjC<->Go and
Java->Go calls. Threads not already attached to the JVM are attached.
Introduce a thread local variable to detach such threads at thread exit.
Change-Id: I8cb65803c9278666ae77a0c7a65dc2d9c7e739e1
Reviewed-on: https://go-review.googlesource.com/19334
Reviewed-by: David Crawshaw <crawshaw@golang.org>
To implement an interface in Java the interface type must be listed
in the class declaration. Emulate the implicit Go interface
implementations in Java by listing all possible (non-empty) interfaces.
For example, given
type (
S struct{}
I interface {
M()
}
)
func (s *S) M() {
}
in Go, the Java class S will be declared to implement the Java
interface I.
Fixes a TODO.
Change-Id: I5b0d2dd65938004ab29029f481cace4b8fb4b26f
Reviewed-on: https://go-review.googlesource.com/19417
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
An inspection of the RefTracker inc and dec methods suggests
that a particular Java Stub instance is included in the javaObjs
map if and only if its reference count, refcnt, is larger than zero.
A newly created reference created by RefTracker.createRef has zero
refcnt but was also inserted in the javaObjs map, violating the invariant.
Fix that by not inserting new references in javaObjs. Without the fix
a Stub instance that were never passed to Go would leak, along with
any other instances it referenced, transitively.
This fixes a Java reference tracking problem, I have not verified if
the same problem applies to the Go and ObjC sides of the Seq machinery.
Change-Id: I3ede90d5258630bc837fe61bba850df222d09a26
Reviewed-on: https://go-review.googlesource.com/19261
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
This allows reuse of the code in custom gomobile bind tools.
Change-Id: I4e013ca871d0fa64983e7efb5e1e9dad8ac723c0
Reviewed-on: https://go-review.googlesource.com/18581
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Fixes the bug in 64bit platforms.
Change-Id: I499995d69db03d136dcb2ca7f930ba1e309c8b6d
Reviewed-on: https://go-review.googlesource.com/19070
Reviewed-by: David Crawshaw <crawshaw@golang.org>
not harmful, but this consistency may simplify cgo handling
in some custom build systems
Change-Id: Id4692986725b3737c23cdeb9ce1a1fa2bc9f7dad
Reviewed-on: https://go-review.googlesource.com/18615
Reviewed-by: David Crawshaw <crawshaw@golang.org>
This removes the last dependency on an Android class in the running
gobind code. (Tests still need some work before they will run on the
desktop.)
Change-Id: I49bfb545d4587c6f430c0938fa1ca4d513b56d77
Reviewed-on: https://go-review.googlesource.com/17252
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
The aim of the RefTracker is to "... pin Java objects so
they don't get GCed while the only reference to them is
held by Go code." But in the case of null objects there is
nothing to GC. Therefore we do return a single Ref which
contains a null object, but we ignore the logic that is used
to pin for GC purposes.
Change-Id: If3771ec0180d09485963c3297abccb39a1a8d9ab
Reviewed-on: https://go-review.googlesource.com/13647
Reviewed-by: David Crawshaw <crawshaw@golang.org>
error message example:
gomobile: unsupported named type github.com/sridharv/bugreports/typealiasmissing.Alias
Objective-C and Java generation code currently panics, which need to be
fixed separately. (See TODO in bind/seq.go)
Fixesgolang/go#13190
Change-Id: Ie46dc58ea800522b8ab7cb8ac662ad561e0ca82e
Reviewed-on: https://go-review.googlesource.com/16780
Reviewed-by: Burcu Dogan <jbd@google.com>
for example,
package testpkg
var AnInt int64
will be mapped to
@interface GoTestpkg: NSObject
+ (int64_t) AnInt;
+ (void) setAnInt:(int64_t)v;
@end
Followup of cl/15340
Update golang/go#12475
Change-Id: Ie26c92af977fc3dd62dcad2b10c6a5c1c1b8941b
Reviewed-on: https://go-review.googlesource.com/15770
Reviewed-by: David Crawshaw <crawshaw@golang.org>
A framework generated with gomobile bind -target=ios has two global
constructors: one initializing a data structure and another using it.
These constructors are defined in different translation units, which
(I believe, reasoning from C++ global constructors) means their order
of initialization is undefined.
A capturing block is stack allocated. Its memory is invalid after the
function returns. Make a copy of the interface initializer blocks so
they can be saved to the heap.
Block implementation background:
http://www.cocoawithlove.com/2009/10/how-blocks-are-implemented-and.html
Updates golang/go#12590
Change-Id: Ia7ae9f4bbd8df6e6e79949de54b3e6c48148c700
Reviewed-on: https://go-review.googlesource.com/14549
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
To improve readability for default numeric types
represented as strings, use 0.0 for floating point types.
(No tests make use of this change so golden update
not necessary; not significant enough for new test/test mod.)
Fix spelling on error statement; this was the only
instance of that spelling error in the repository.
Change-Id: I373890725b33da11c6780ba93674d89541bf758c
Reviewed-on: https://go-review.googlesource.com/13645
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
Introduce options -javapkg and -prefix for gobind command.
The following generates java class Testpkg with package name com.example.
gobind -lang=java -javapkg=com.example testpkg
The following generates objective-c files where function and type names
are prefixed with ExampleTestpkg.
gobind -lang=objc -prefix=Example testpkg
As discussed in golang/go#9660 and golang/go#12245.
Gomobile support is not yet implemented.
Change-Id: Ib9e39997ce915580a5a2e25643c0c28373f27ee1
Reviewed-on: https://go-review.googlesource.com/13969
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Tests now pass on OS X.
Change-Id: Id2e6c5bd9ddd4070962ff9bc121d7902e8c9a647
Reviewed-on: https://go-review.googlesource.com/13694
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
For interface binding, we need only protocols.
Change-Id: I377ec9c8de0a3414d48272892b169fb94823e8ff
Reviewed-on: https://go-review.googlesource.com/13541
Reviewed-by: David Crawshaw <crawshaw@golang.org>
