- Add `@Internal` and `@ExperimentalApi`, both are annotated `@Internal`
- Annotate `@Internal` to `package io.grpc.internal`
- AbstractChannelBuilder.ChannelEssentials is annotated `@Internal`
- ChannelImpl.ping() is annotated `@ExperimentalApi`
- Context is annotated `@ExperimentalApi`
- Add `package-info.java` to `io.grpc.inprocess` and `io.grpc.internal`.
Reserve io.grpc for public API only, and all internal stuff in core to
io.grpc.internal, including the non-stable transport API.
Raise the netty/okhttp/inprocess subpackages one level up to io.grpc,
because they are public API and entry points for most users.
Details:
- Rename io.grpc.transport to io.grpc.internal;
- Move SharedResourceHolder and SerializingExecutor to io.grpc.internal
- Rename io.grpc.transport.{netty|okhttp|inprocess} to
io.grpc.{netty|okhttp|inprocess}
Holding the lock while calling the transport can cause a deadlock, as
shown in #696. In previous auditing for deadlock prevention I considered
heavily Call interactions, but failed to consider shutdown() and realize
it was holding a lock while calling transport.shutdown().
We still hold a lock when calling transport.start(). Although it is
conceivable that this could cause a deadlock as the code evolves over
time, I don't believe it can cause a deadlock today or that the risk is
very high. In addition, it would require more effort to solve.
Note that even more importantly, this translates a RST_STREAM error code
to a gRPC status code. This is generally useful, but also necessary for
DEADLINE_EXCEEDED to be more reliable in 0eae0d9.
Fixes#687
This does make use of the fact we are no longer using Multimap. Doing
entries() for ArrayListMultimap must create a new Map.Entry for every
entry. Since we are now using HashMap, we are able to use entries() with
no extra cost.
Merging particular Keys no longer needs to deserialize.
Multimap creates at least one new object for every access, because all
the objects it returns are "live" and when mutated they need to update
the multimap's size. Many common operations thus require at least an
object allocation per key.
Note that previously remove() was non-functional as it removed the wrong
type from the multimap. The type system did not catch this because
remove() is passed an Object for all collection types.
The return type of removeAll() was changed to Iterable to prevent the
need of converting to Key if the caller doesn't consume the return
value.
Although it appears serialize() is now more expensive in terms of
allocations because it first accumulates into an ArrayList, the memory
usage is approximately the same since Multimap.values() makes an
Iterator for each key. The new code would allocate fewer bytes overall
and in fewer allocations, but the older code retained less memory while
processing. If we want to optimize serialize() we can track the number
of entries without needing to do any wrapping like Multimap does. I
didn't bother because the ArrayList is a fraction of the cost compared
to actually serializing the values.
CANCELLED is certainly not the right status code. Communicating the
exception to the client removes the need for logging, which also makes
it more clear which call experienced the problem.
Improved some consistency. writeHeaders was the only non-final
implementation method of ServerStream, even though it is really no
different than the others.
The previous order was unintuitive as the following would execute in the
reverse order:
Channel channel;
channel = ClientInterceptors.intercept(channel, interceptor1,
interceptor2);
// vs
channel = ClientInterceptors.intercept(channel, interceptor1);
channel = ClientInterceptors.intercept(channel, interceptor2);
After this change, they have equivalent behavior. With this change,
there are no more per-invocation allocations and so calling 'next' twice
is no longer prohibited.
Resolves#570
Forgot to add this last file
updated method name
Remove unused function
Remove helper function for threshold edge detection
Remove helper function for threshold edge detection
Re make listener abstract
The CallImpls in ChannelImpl and ServerImpl implement the Call
interfaces; they should be the ones ensuring that inappropriate calling
of methods is handled as the interface describes.
The client can race with the server in cancelling due to deadline. If
server cancels we don't get DEADLINE_EXCEEDED, so double-check on
client-side to reduce the chances of losing the race.
Generally we expect the client to lose the race because of coarse timer
granularity for timer expirary. This change does little to help if the
server's clock runs noticably "fast" relative to the client.
call stack and across thread boundaries.
Strongly modeled after the Go context API https://blog.golang.org/context with support for
- cancellation propagation & cancellation listeners
- typed value binding
- timeout/deadline
The major difference with Go is that ThreadLocal is used for propagation instead of parameter
passing as this is considered more idiomatic for Java.
- Rename flushTo() to drainTo().
- Remove flushTo() from DeferredNanoProtoInputStream (which is renamed
to NanoProtoInputStream), because the optimization is not implemented.
- Rename DeferredProtoInputStream to ProtoInputStream.
#529
Resolves#511.
- In generated code, make CONFIG private and METHOD_* fields public.
METHOD_* fields are MethodDescriptors now, users of the CONFIG field
should switch to using the METHOD_* fields.
- Move MethodType into MethodDescriptor (#529).
- Unify the fully qualified method name. It is fully qualified service
name + slash + short method name. It doesn't have the leading slash.
- HandlerRegistry switches the key from short method name to fully
qualified method name.
- Pass CallOptions to Channel.newCall() and
ClientInterceptor.interceptCall().
- Remove timeout from AbstractStub.StubConfigBuilder and add deadline,
which is stored in a CallOptions inside the stub.
- Deadline is in nanoseconds in the clock defined by System.nanoTime().
It is converted to timeout before transmitting on the wire. Fail the
call with DEADLINE_EXCEEDED if it's already expired.
The mapping is poorly suited for gRPC. C and Go don't even do any
mapping. We can improve the mapping in the future, but it is very
important that users don't start depending on the current mapping.
This change is "inspired by" the original code, but is even more
conservative.
Fixes#477
Other classes are already following the convention that ClientFoo for
client-side, and ServerFoo for server-side. Call has been the black
sheep of the family.
- Call -> ClientCall
- Calls -> ClientCalls
- ForwardingCall* -> ForwardingClientCall*
We know the size won't be more than serializedHeaders.length, is
unlikely to be fewer, and very unlikely to be substantially fewer. This
prevents us from needing to resize the array in all cases.
Resolves#321
By default, ArrayDeque will be of size 16, which is an overkill for most
calls. "4" is not a magical number itself, but seems a better guess than
16 since we do have some knowledge of how much it will contain.
Resolves#320
isReady() can provide pushback while the call is in progress, but it
can also provide the pushback necessary when the client creates more
streams than permitted by MAX_CONCURRENT_STREAMS.
As part of this commit, OkHttp is now calling onReady() after call
creation (it previously never called onReady()).
Ideally OKHttp wouldn't do blocking I/O during start(), but it does and
fixing it is non-trivial. OkHttp can either throw an exception when it
encounters an error during start or it can shut itself down. Both
require changes in ChannelImpl, so we just choose to keep OkHttp's
current behavior and deal with it in ChannelImpl.
Operation is a term no longer used in gRPC. StatusException seems clear
and is concise. Moved out of Status class to remove stuttering.
The return types of as*Exception() is now explicitly the
Status.*Exception type.
If we notice something wrong with the headers, then we choose a nice
error message. But we are accidentally overwriting that message with
additional error details.
Set upper and lower bounds for Netty & OkHttp allocators based on transport limitations and benchmark results.
Fix OkHttp OutboundFlowController to chunk payloads larger than frameWriter.maxDataLength
Allow OkHttp to allocate buffers to FrameWriter larger than max DATA length
MessageFramer allows queing of data and explicit flushing. Sinks
generally can benefit from knowing when they are required to flush, so
we now tell them when MessageFramer received a flush so they only have
to flush when required.
ServerImpl.start() now throws IOException to make the error explicit.
This was previously being papered over by wrapping the exception in
RuntimeException.
call super.start(), and makes the subsequent use of other methods on the
call throw IllegalStateException.
Create ClientInterceptors.CheckedForwardingCall that handles exception
in start logic.
Create Forwarding[Server]Call[Listener] for generic decoration use
cases, with an abstract delegate() for flexibility.
Create SimpleForwarding[Server]Call[Listener] to replace now deprecated
forwarding classes.
1. Adds <property name="separateLineBetweenGroups" value="true"/> to CustomImportOrder to enfore blank line between imports groups.
2. Uses checkstyle 6.5, which fixed a bug of "CustomImportOrder checks import sorting according to ASCII order instead of case-insensitive alphabetical order".
Also add OkHttpReadableBuffer unit test, and increase the test string
length to expose the bug.
Also use array equality assertions in the base test, instead of
comparisons whose return value is discarded.
Fixes#231.
When the description is null, the exception message would be in the form
of "INTERNAL: null", which isn't very attractive, and makes it seem like
an error in itself. If the description is null, just use "INTERNAL".
The checkstyle.xml is a slightly modified version of the upstream Google
checkstyle configuration. All changes have comment describing them.
Lots of warnings were corrected. Examples is the only project that has
warnings still, as the necessary changes require some thought.
-Xlint:-options is not available on some earlier JDK 7s, but won't fail
if unsupported. It prevents the warning wanting bootclasspath specified
since target/source is 1.6.
This allows sooner delivery of errors. We never needed to stop delivery
for unexpected EOS, but instead the application would have been required
to request() another message before delivering. Stalling MessageDeframer
sooner removes the need for the application to request another message
before noticing that the buffers are empty.
The Http2ClientStream should not close the buffer in this case since
it's already been given to the deframer and potentially to the user.
Added cleanup code to MessageDeframer and AbstractClientStream to make
sure that we free the Buffer when appropriate.
Some early grpc users on Android are using a very old Guava. They are
working on upgrading, but it will take time. This alone is not enough
for "old guava" compilibility; expect more to come.
OkHttp no longer cancels all calls on shutdown, as we want to allow
graceful shutdown. Such cancelling behavior will likely be provided by
Channel in the future.
OkHttp should now have a thread-safe implementation of newStream.
Previously 'lock' was not held when checking goAway and the checking in
AbstractClientTransport was redundant.
Netty was thread-safe, but it was very hard to tell what guarantees were
necessary and what guarantees each piece was providing.
Previously streams were being partially orphaned if there was an
interruption during stream creation. To handle cancellation,
AbstractClientStream's cancel() had to be changed remove the
"optimization" otherwise, again, the stream would be orphaned.
This change loses asynchronous notification of channel state-change and
a way to wait until the channel is actually connected. Both of these are
expected to be added back as part of a health API. The important
distinction from Service is that ChannelImpl never permanently fails and
can revert from being started to connecting again.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83875407
There may already be a transportError, in which case we want to augment
it and not overwrite it.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83807747
Summary of changes:
1) Merged the interfaces MessageDeframer2.Sink and DeframerListener into
MessageDeframer2.Listener. This simplifies the interface of
MessageDeframer2 quite a bit.
2) Added a deliveryPaused() handler to MessageDeframer2.Listener, which
is called by the deframer when there is not enough data to read/deliver
the next message.
3) Modified AbstractStream and AbstractClientStream to manage the timing
of when the closed() event is delivered to the listener. The
transportReportStatus ultimately controls this by creating a task to
close the listener. It either runs this task immediately or when the
next deliveryPaused() event occurs.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83620903
No major refactorings/simplifications were done. Only gRPC v1 support
infrastructure was removed.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82737436
Any place that force-sets the protocol to 2 or assumes the old value
is now removed. Unfortunately, it seems InProcessTransportTest has
some non-obvious dependency on gRPC v1.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=81597524
have "-bin" suffix in their names.
Split Metadata.Marshaller into BinaryMarshaller and AsciiMarshaller.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=81306135
We don't use the decompressor at all in gRPC v2, but it may still have
resources that should be close()d.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=81271990
check for it, so that we can detect intermediate proxies that do not support
trailers.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=81084983
Remove synchronization on stateLock as we are not required to be thread safe
Add better toString for stream impls
Internal cleanup of various 'status' fields in AbstractClientStream
Remove 'stashTrailers' as we've already extracted status in layer above correctly
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=80678356
The client now checks certificates and performs hostname verification.
Tests check certificates, so the server uses a cert that the client
trusts.
Only the client portion of SslContextFactory was previously used.
Applications that want to ignore certificates (i.e., for testing) can
use io.netty.handler.ssl.util.InsecureTrustManagerFactory instead.
The MOE configuration was already failing to work, and so required the
simple mapping for examples in addition to what was needed for the new
certs.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=80434148
contention between client and server in the same test process.
Name the threads of default thread pools.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=80326782
The GFE should not be sending DATA frames with endOfStream=true.
However to make things work in the short-term, if we receive one we
still need to process the DATA frame.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=80243944
A recent refactoring moved code so our previous workaround stopped
producing any effect.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=79965647
This will replace the channel-wrapping technique as the official
solution of intercepting calls on the client side. It is done in the
same manner as ServerInterceptor.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=79280263
We know we don't like the API, but don't yet have the new one in place.
Remove the old one now instead of waiting further, because we are
getting more users who are more likely to try to start using the API.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=79090165
The TestService proto is temporarily supplied as a generated jar (until the open source protoc compiler supports grpc).
Copies of messages.proto, empty.proto, and message_set.proto are scrubbed and included in the source under integration-testing.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=78711468
pools etc) that are shared among channels and servers as default values when
the application doesn't provide its own.
It uses reference counting to shut down resources with a delay.
Changed the channel and server builders to use it for default values of
executors and event loop groups.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=78629204
overview:
1) Lots of @SuppressWarnings :)
2) Remove dependencies on StandardCharsets.XXX (which is Java 7)
3) Moved testing/utils/ssl/* to .../stubby/util so that the netty transport doesn't depend directly on the testing module.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=78460727
The semantics in gRPC changed where server sending OK before client
half-closes is now permitted.
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=78404390
nmittler
Kun Zhang
Carl Mastrangelo
Xudong Ma
Eric Anderson
Louis Ryan
Jack Coughlin
David Connelly
Josh Humphries
Thomas ten Cate
Jakob Buchgraber
Jorge Canizales
nabeelmian
yangg