[protobuf] Re: Issue 248 in protobuf: protobuf will not compile without thread library

2010-12-07 Thread protobuf
Updates: Status: Accepted Owner: ken...@google.com Comment #2 on issue 248 by ken...@google.com: protobuf will not compile without thread library http://code.google.com/p/protobuf/issues/detail?id=248 Hmm. I don't think we should automatically fall back to thread-hostile co

Re: [protobuf] 2.4.0 and lazy UTF-8 conversions in Java

2010-12-07 Thread Kenton Varda
On Wed, Dec 1, 2010 at 3:33 AM, Evan Jones wrote: > The instanceof approach to switch between the two is a good idea. When I > wrote my implementation, I was concerned about the thread-safeness issues, > although I don't think I ever considered this particular version. However, I > think this can

Re: [protobuf] Any Python wrappers for the C++ implementation?

2010-12-07 Thread Kenton Varda
Cool. Serialization and parsing themselves should actually be improved even more than that, but having other Python code around it waters down the numbers. :) Also, note that if you explicitly compile C++ versions of your messages and link them into the process, they'll be even faster. (If you

Re: [protobuf] protobuf json codec

2010-12-07 Thread Kenton Varda
Hi Siju, This is pretty cool. I added it to the third-party add-ons wiki. http://code.google.com/p/protobuf/wiki/ThirdPartyAddOns Just a minor complaint: You are not Google, so your code should not live under the com.google package. Can you please move it to a different package? Perhaps com

Re: [protobuf] Preferred Citation?

2010-12-07 Thread Kenton Varda
There's a (slightly outdated) list of contributors here: http://code.google.com/p/protobuf/source/browse/trunk/CONTRIBUTORS.txt But I think it's probably best to list Google as the author. I did not invent protocol buffers; I just wrote version 2 and open sourced it, and I had help. I don't re

[protobuf] Re: Issue 196 in protobuf: Python: Ascii output is not assured to be in utf-8

2010-12-07 Thread protobuf
Updates: Status: Accepted Labels: -FixedIn-2.4.0 Comment #3 on issue 196 by ken...@google.com: Python: Ascii output is not assured to be in utf-8 http://code.google.com/p/protobuf/issues/detail?id=196 Jisi, I'm not convinced that this is fixed. The as_utf parameter simply p

Re: [protobuf] Re: java parse with class known at runtime (and compiled proto)

2010-12-07 Thread Kenton Varda
Evan is correct. The best way to write code which deals with a generic protobuf type is to have it take a default instance as a parameter. From that you can do everything else. This is actually better than passing around Class objects, because it allows users to use DynamicMessages with your cod

Re: [protobuf] Nullpointer Exception when .parseFrom(byte [])

2010-12-07 Thread Kenton Varda
The file you attached is not useful without source code. But what would be even better is if you could provide the stack trace for the NullPointerException. The parser should never throw NPE and I'm not aware of any bugs which cause it to throw NPE. BTW, it's very easy for bytes which are not ac

Re: [protobuf] "File already exists in database" on Mac OSX

2010-12-07 Thread Kenton Varda
I assume this is happening at startup? Is it possible that you are somehow linking two copies of the same .pb.cc into your binary? There are cases where the linker doesn't catch this, especially when dynamic linking is involved, but sometimes even with static linking. Speaking of which, are you

[protobuf] Re: Issue 247 in protobuf: Ability to redirect file output to stdout

2010-12-07 Thread protobuf
Comment #2 on issue 247 by ken...@google.com: Ability to redirect file output to stdout http://code.google.com/p/protobuf/issues/detail?id=247 What happens if multiple files are generated? For example, if you have "option java_multiple_files = true;" in your .proto file. This seems like

[protobuf] 3 issues changed in protobuf

2010-12-07 Thread protobuf
Updates: Status: Fixed Labels: FixedIn-2.4.0 Comment by liuj...@google.com: Fixed in r358 Affected issues: issue 223: Python Package Doesn't Contain Compiler Plugin Module http://code.google.com/p/protobuf/issues/detail?id=223 issue 224: Protobuf installation error in Cy

[protobuf] Re: Issue 179 in protobuf: Visual C++ error C1091 when compiling protoc generated code with over 64k descriptor

2010-12-07 Thread protobuf
Updates: Status: New Comment #2 on issue 179 by ken...@google.com: Visual C++ error C1091 when compiling protoc generated code with over 64k descriptor http://code.google.com/p/protobuf/issues/detail?id=179 This proves complicated to fix, because we have lots of code that passes aro

Re: [protobuf] Any Python wrappers for the C++ implementation?

2010-12-07 Thread Yang Zhang
On Tue, Dec 7, 2010 at 7:08 PM, Kenton Varda wrote: > Cool.  Serialization and parsing themselves should actually be improved even > more than that, but having other Python code around it waters down the > numbers.  :) The times are from a minimal microbenchmark using Python's timeit module: nru

[protobuf] Re: Issue 210 in protobuf: Java code should detect incompatible runtime library version

2010-12-07 Thread protobuf
Updates: Cc: liuj...@google.com Comment #13 on issue 210 by liuj...@google.com: Java code should detect incompatible runtime library version http://code.google.com/p/protobuf/issues/detail?id=210 Hmm, not very clear about the maven version compatibility... So suppose you have a pom.

Re: [protobuf] Any Python wrappers for the C++ implementation?

2010-12-07 Thread Kenton Varda
On Tue, Dec 7, 2010 at 9:19 PM, Yang Zhang wrote: > > Also, note that if you explicitly compile C++ versions of your > > messages and link them into the process, they'll be even faster. (If you > > don't, the library falls back to DynamicMessage which is not as fast as > > generated code.) > > I

[protobuf] Re: Issue 210 in protobuf: Java code should detect incompatible runtime library version

2010-12-07 Thread protobuf
Comment #14 on issue 210 by aantono: Java code should detect incompatible runtime library version http://code.google.com/p/protobuf/issues/detail?id=210 It would if you specify, what Maven calls, a range -> [1.0,2.0), which would basically read that anything inclusive between 1.0 and 2.0, b

Re: [protobuf] Any Python wrappers for the C++ implementation?

2010-12-07 Thread Yang Zhang
On Tue, Dec 7, 2010 at 9:40 PM, Kenton Varda wrote: > On Tue, Dec 7, 2010 at 9:19 PM, Yang Zhang wrote: >> >> > Also, note that if you explicitly compile C++ versions of your >> > messages and link them into the process, they'll be even faster.  (If >> > you >> > don't, the library falls back to