Comments on: 100 lines of C that can parse any Protocol Buffer

By: Martin Gadbois

Martin Gadbois — Mon, 27 Jul 2009 15:58:35 +0000

Your 100 lines of code is great; would you care to select a license to publish it?

By: Emmanuel Castro

Emmanuel Castro — Fri, 07 Nov 2008 13:01:36 +0000

You chose the LLVM bitcode format for Gazelle instead of Google Protocol Buffer.

Is there any reason, or was it just because you made the choice before Protocol Buffer was available.

Protocol Buffer seems to have a better language support (C , Java, Python).

Note: I just finished porting the LLVM bitcode reader in Java (translated from the C version). I plan to release a Gazelle in Java with it. Let me a word if you want to see the sources (not packaged yet).

By: Ronald

Ronald — Wed, 16 Jul 2008 16:26:48 +0000

That .c source code was the most boring game I’ve ever played. It only outputted some text with no instructions of how to play or anything. Do you even get to save any highscores if you figure out how to play??

By: josh

josh — Sat, 12 Jul 2008 19:46:33 +0000

@Matthias,

Maybe I wasn’t clear enough: my 100 lines of C does not use any protobuf library. It actually implements the decoding of a protobuf from scratch, and has no external dependencies except for the standard C library.

By: Matthias

Matthias — Sat, 12 Jul 2008 19:42:00 +0000

This should be much shorter. Create a new FileDescriptorProto with one empty DescriptorProto. From that create a FileDescriptor, lookup the “empty message” descriptor, create a DynamicMessage from that and parse the input. The unknown fieldset contains the fields.

Actually, in Java you can directly parse an UnknownFieldSet.