Did you ever figure it out? I would love to know! I'm running into the same
problem.
On Friday, February 10, 2017 at 7:41:54 AM UTC-8, Sanket Lawangare wrote:
>
> I am using the grpc command line tool to test response for my request.
> This is the official link to the grpc command line tool:
I do not think we support that. Actually I think tokens are intentionally
dropped if the channel is not secure to avoid leaking the tokens over the
wire.
On Wed, Feb 7, 2018 at 3:04 PM, wrote:
> I was able to get this working by simply calling AddMetadata on the client
Hi Vijay,
Thank you very much for the reply!
1. "Finish will always give an ok of 1", I think this is clear and makes
perfect sense. This is corresponding to the last "ok" print in my log:
Tag received: 0x2419440, Completion queue/Next() returned: 1
HandleEachResponse_FINISH_BAD
2. I read that
I was able to get this working by simply calling AddMetadata on the client
context to set the authorization header. This appears to work for both
secure and insecure connections, so I'll go with this approach rather than
using the custom authenticator.
On Wednesday, 7 February 2018 12:20:11
As Mark mentioned, it's not on our roadmap, but if you wanted to put one
together, I can suggest
https://github.com/grpc/grpc/blob/master/doc/core/transport_explainer.md
for an idea of what transports need to do. If the goal is a simple
message-oriented transport without reference to complex
Hi there,
"ok" is documented at
https://github.com/grpc/grpc/blob/master/include/grpc%2B%2B/impl/codegen/completion_queue.h#L124
.
Finish will always give an ok of 1 since it's always ok to be done with
your RPC and check its status.
When starting an RPC, ok will give 0 only if gRPC is sure
Your flow is correct. Is it RPC? I guess that's a more philosophical,
but I don't think there is anything wrong with it. Streaming RPCs are
already kind of weird anyway.
A more pure solution would be something akin to tunneling, where a client
connects to a server, but then runs a gRPC
I started learning grpc very recently, and followed the docs/instructions
and discussions online to build a toy program of an asynchronous streaming
RPC client (.proto and client.cc are attached below). *My grpc version is
1.8.4.*
My main idea for this client is to create a "listener" in a
Got it. So this isn't considered bad form for RPC? Technically it isn't
RPC? In my case:
- Server implements rpc Link (stream UpLink) returns (stream DownLink) {}
- Client initiates connection
- Client waits for messages to come down from the server to be executed
- Clients parses the
I'm looking into how to support client connections to disparate services
using custom oAuth2 tokens. The services may or may not support SSL/TLS.
I have created a custom authenticator plugin following the sample code
at https://grpc.io/docs/guides/auth.html.
I build my custom call credentials
Yes, that is correct. (It's bidirectional because the word "stream" is in
both the request (UpLink) and response (DownLink).)
I mistakenly thought this was another thread (which I replied to yesterday)
which had almost the same issue. Take a look at
I'm unclear on what the last part means: "That's why that issue, (and my
post), suggest treating the message as inverted requests and responses."
Do you mean rpc Link (stream UpLink) returns (stream DownLink) {} ? With a
bidirectional stream?
Thanks
On Wed, Feb 7, 2018 at 11:27 AM, 'Carl
It's unlikely that gRPC will be able to have servers initiate connections /
rpcs to clients. The initial headers need to be sent by the client to the
server, and only the server can send trailers to indicate the RPC is done.
But, after these two, the client and server are peers. That's why
If I understand correctly, your server interceptor wants to know the time
when the RPC is started on the client.
Yes you can write a client interceptor that adds the timestamp to the
headers.
On Tuesday, January 30, 2018 at 7:50:00 PM UTC-8, vicva...@gmail.com wrote:
>
> I am using java
--
It is actually very useful for communicating with microcontrollers. We do
something similar with just sending protobuf request responses using
nanopb. Our microcontroller is resource limited so running the full grpc
stack is not feasible on that side, but might be useful in other situations.
On
As far as I know, no one has requested this before, so it's not on our
roadmap.
What is your use-case for this? I'm not familiar with USB HID, but a quick
web search indicates that it's for connecting things like keyboards and
mice via USB, and it's not clear to me why you'd want to use gRPC for
16 matches
Mail list logo