From from at alumni.stanford.edu  Fri May  6 00:10:30 2016
From: from at alumni.stanford.edu (from at alumni.stanford.edu)
Date: Thu, 5 May 2016 15:10:30 -0700
Subject: [Rprotobuf-yada] Difficulty getting RProtoBuf to work on Linux
Message-ID: <CANmG5NFB673XZ0=Yd-vnMndMeqd40uEQqa_M8SbAM_fEygQF6w@mail.gmail.com>

Dear RProtoBuf Developers,

I'm teaching a class, and we are trying to use RProtoBuf on Linux.

We are encountering the problem when trying to read in the .proto files.

I've appended a minimal example of the .proto files below. We are trying to
read these in with

readProtoFiles(dir = "proto_dir")

The errors are:

protos_min/P2.proto:7:22:"com.name.common.value.P2Proto.var_2" is already
defined in file "P2.proto".
protos_min/P2.proto:6:9:"com.name.common.value.P2Proto" is already defined
in file "P2.proto".

The .proto files worked with Java protobuf code, but we would like to try
to do all of our work within R.

We are using R 3.2.4 on RHEL 6, and we've tried compiling RProtoBuf against
protobuf versions 2.4.1, 2.5.0, and 2.6.1.

Any advice is most appreciated. None of us has any experience with
protobufs, yet the students need to decode this data to do their work.

Many thanks,

Bill

::::::::::::::
P1.proto
::::::::::::::
package com.name.common.value;

option java_outer_classname="P1ProtoBuffs";
import "P2.proto";

message P1Proto {
    optional P2Proto var_1 = 3;
}

::::::::::::::
P2.proto
::::::::::::::
package com.name.common.value;

option java_outer_classname="P2ProtoBuffs";
import "P3.proto";

message P2Proto {
    optional P3Proto var_2 = 2;
}

::::::::::::::
P3.proto
::::::::::::::
package com.name.common.value;

option java_outer_classname="P3ProtoBuffs";
import "P4.proto";

message P3Proto {
    optional commagg.value.P4Proto var_3 = 1;
}
::::::::::::::
P4.proto
::::::::::::::
package com.name.commagg.value;

option java_outer_classname="P4Protos";

message P4Proto {
    message P4S1Proto {
        optional string var_4 = 3;
    }

    message P4S2Proto {
        optional P4S1Proto var_5 = 1;
    }

    repeated P4S1Proto var_6 = 1;
    optional P4S1Proto var_7 = 2;
    optional P4S2Proto var_8 = 3;
}
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rprotobuf-yada/attachments/20160505/d0f50c5f/attachment.html>

From murray at stokely.org  Fri May  6 01:03:28 2016
From: murray at stokely.org (Murray Stokely)
Date: Thu, 5 May 2016 16:03:28 -0700
Subject: [Rprotobuf-yada] Difficulty getting RProtoBuf to work on Linux
In-Reply-To: <CANmG5NFB673XZ0=Yd-vnMndMeqd40uEQqa_M8SbAM_fEygQF6w@mail.gmail.com>
References: <CANmG5NFB673XZ0=Yd-vnMndMeqd40uEQqa_M8SbAM_fEygQF6w@mail.gmail.com>
Message-ID: <CAECWziJGaviF3UFtMGT=FvyMD4rG=DDLfEQzS4ougQmQKJq4Cg@mail.gmail.com>

Can you try to just :

readProtoFiles("P1.proto")

I think it may be getting confused by trying to read in P4.proto, then
reading it in again through the import statement.

   - Murray


On Thu, May 5, 2016 at 3:10 PM, <from at alumni.stanford.edu> wrote:

>
> Dear RProtoBuf Developers,
>
> I'm teaching a class, and we are trying to use RProtoBuf on Linux.
>
> We are encountering the problem when trying to read in the .proto files.
>
> I've appended a minimal example of the .proto files below. We are trying
> to read these in with
>
> readProtoFiles(dir = "proto_dir")
>
> The errors are:
>
> protos_min/P2.proto:7:22:"com.name.common.value.P2Proto.var_2" is already
> defined in file "P2.proto".
> protos_min/P2.proto:6:9:"com.name.common.value.P2Proto" is already defined
> in file "P2.proto".
>
> The .proto files worked with Java protobuf code, but we would like to try
> to do all of our work within R.
>
> We are using R 3.2.4 on RHEL 6, and we've tried compiling RProtoBuf
> against protobuf versions 2.4.1, 2.5.0, and 2.6.1.
>
> Any advice is most appreciated. None of us has any experience with
> protobufs, yet the students need to decode this data to do their work.
>
> Many thanks,
>
> Bill
>
> ::::::::::::::
> P1.proto
> ::::::::::::::
> package com.name.common.value;
>
> option java_outer_classname="P1ProtoBuffs";
> import "P2.proto";
>
> message P1Proto {
>     optional P2Proto var_1 = 3;
> }
>
> ::::::::::::::
> P2.proto
> ::::::::::::::
> package com.name.common.value;
>
> option java_outer_classname="P2ProtoBuffs";
> import "P3.proto";
>
> message P2Proto {
>     optional P3Proto var_2 = 2;
> }
>
> ::::::::::::::
> P3.proto
> ::::::::::::::
> package com.name.common.value;
>
> option java_outer_classname="P3ProtoBuffs";
> import "P4.proto";
>
> message P3Proto {
>     optional commagg.value.P4Proto var_3 = 1;
> }
> ::::::::::::::
> P4.proto
> ::::::::::::::
> package com.name.commagg.value;
>
> option java_outer_classname="P4Protos";
>
> message P4Proto {
>     message P4S1Proto {
>         optional string var_4 = 3;
>     }
>
>     message P4S2Proto {
>         optional P4S1Proto var_5 = 1;
>     }
>
>     repeated P4S1Proto var_6 = 1;
>     optional P4S1Proto var_7 = 2;
>     optional P4S2Proto var_8 = 3;
> }
>
> _______________________________________________
> Rprotobuf-yada mailing list
> Rprotobuf-yada at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rprotobuf-yada
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rprotobuf-yada/attachments/20160505/5f3668c0/attachment.html>

From jon at thon.cc  Tue May 17 03:45:35 2016
From: jon at thon.cc (Jonathon Love)
Date: Tue, 17 May 2016 11:45:35 +1000
Subject: [Rprotobuf-yada] RProtoBuf - OSX version on CRAN built with old
	version
Message-ID: <79fd70d5-eac0-60c0-d60f-93d81e36b5d2@thon.love>

hi folks,

the new 0.4.3 version of RProtoBuf for OSX on CRAN is built using an 
older version of Protocol Buffers (possibly 2.5) than the 0.4.2 which 
was built with 2.6. this results in a body of my code no longer working.

i suspect it has something to do with the new R build platform for R 3.3

does anyone here have any connection with the build process of RProtoBuf 
for OSX on CRAN? i've tried a few other avenues, but they've all come up 
empty. i also thought posting here would clue people into potential 
problems they might have in the future.

here is a version of RProtoBuf with a test which demonstrates the issue:

https://github.com/jonathon-love/rprotobuf

with thanks

jonathon


From edd at debian.org  Tue May 17 03:52:46 2016
From: edd at debian.org (Dirk Eddelbuettel)
Date: Mon, 16 May 2016 20:52:46 -0500
Subject: [Rprotobuf-yada] RProtoBuf - OSX version on CRAN built with
	old	version
In-Reply-To: <79fd70d5-eac0-60c0-d60f-93d81e36b5d2@thon.love>
References: <79fd70d5-eac0-60c0-d60f-93d81e36b5d2@thon.love>
Message-ID: <22330.31086.143094.739400@max.nulle.part>


On 17 May 2016 at 11:45, Jonathon Love wrote:
| hi folks,
| 
| the new 0.4.3 version of RProtoBuf for OSX on CRAN is built using an 
| older version of Protocol Buffers (possibly 2.5) than the 0.4.2 which 
| was built with 2.6. this results in a body of my code no longer working.
| 
| i suspect it has something to do with the new R build platform for R 3.3
| 
| does anyone here have any connection with the build process of RProtoBuf 
| for OSX on CRAN? i've tried a few other avenues, but they've all come up 

As I mentioned to you a couple of days ago (when the list was still hanging),
OS X builds are administered by Simon Urbanek who is famously busy.

If you have alternate build tools you could just set up a repo yourself. The
drat package on CRAN makes that easy -- see my site or the CRAN pages for it.

Dirk

| empty. i also thought posting here would clue people into potential 
| problems they might have in the future.
| 
| here is a version of RProtoBuf with a test which demonstrates the issue:
| 
| https://github.com/jonathon-love/rprotobuf
| 
| with thanks
| 
| jonathon
| 
| _______________________________________________
| Rprotobuf-yada mailing list
| Rprotobuf-yada at lists.r-forge.r-project.org
| https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rprotobuf-yada

-- 
http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org