[protobuf] .proto file version constant

2010-09-08 Thread Jesper
I would like to be able to have a version constant in the .proto file, which can be encoded into messages, so that the communicating parties can verify that they are using a sufficiently new .proto file. I tried using a custom FileOption as described here> http://code.google.com/apis/protocolbuffer

[protobuf] Issue 217 in protobuf: build failed on fedora 9 - i386

2010-09-08 Thread protobuf
Status: New Owner: ken...@google.com Labels: Type-Defect Priority-Medium New issue 217 by qinzuoyan: build failed on fedora 9 - i386 http://code.google.com/p/protobuf/issues/detail?id=217 What steps will reproduce the problem? 1. ./configure 2. make 3. What is the expected output? What do you

Re: [protobuf] doubt

2010-09-08 Thread Daniel Wright
They're effectively the same -- it's just a style question. If B only makes sense in the context of A, I'd go with the second version, otherwise I'd go with the first. But the generated code should be the same except for the name of B in the generated code. 2010/9/8 alf > what is different o w

[protobuf] doubt

2010-09-08 Thread alf
what is different o what is recommended is the same ? message A { optional string stuff = 1 repeated message A = 2 } message B { optional string stuff=1 } or message A { optional string stuff = 1 repeated message A = 2 message B { optional string stuff=1 } } -- You recei

[protobuf] Could Java initializers be elided for primitive fields in many cases?

2010-09-08 Thread David Dabbs
IIRC Java initializes these primitive types as private boolean foo_boolean;// false private int foo_int;// 0 private longfoo_long; // 0L private float foo_float; // 0.0f private double foo_double; // 0.0d But java_primitive_field.cc unconditi

Re: [protobuf] Java reflection package resolution

2010-09-08 Thread Jon Robison
Ah, okay - that would make sense. I was using the builders after the user input the data to build a new message, so it would make sense to use them before as well. I'll try it soon. On Tue, Sep 7, 2010 at 8:09 PM, Kenton Varda wrote: > Descriptors are just descriptors, not implementations. Sev