Open GoogleCodeExporter opened 9 years ago
Can you find a minimum .proto file that can reproduce the problem and then
attach the source?
Original comment by xiaof...@google.com
on 1 Jul 2013 at 5:23
Hi,
sorry, I can't isolate the problem.
It seems to be some 'race' bug, because when I change 'protoc' argument
order ( but I still use the same bundle of .proto files ), compilation
succeed.
Please see attached ( a little bit complex ) example.
You can simulate the problem easily - see two shell scripts - one will
fail, one will succeed. It seems there is no sense in order of arguments
in 'success' script, because all higher level protocols (
authentication, system information ... ) are based/encapsulated into
lower level protocols (nano, header, ...).
Original comment by dusan...@gmail.com
on 2 Jul 2013 at 8:56
I didn't see the attachment?
Original comment by xiaof...@google.com
on 2 Jul 2013 at 5:25
Was re-sent to your private mail.
Original comment by dusan...@gmail.com
on 3 Jul 2013 at 6:47
The actual problem is that there are two .protos both defining an extension to
MessageOptions using the same extension field number. This conflict causes
protocol compiler to fail when parsing custom options.
The solution is to remove one of the duplicated extension definitions. For
protobuf, the protocol compiler should generate a more meaningfully error
message in this case.
Original comment by xiaof...@google.com
on 10 Jul 2013 at 6:31
Any update on this issue? It seems to be still a problem (I just ran into it
using 2.6.1).
Jason
Original comment by j...@squareup.com
on 3 Feb 2015 at 4:09
Original issue reported on code.google.com by
dusan...@gmail.com
on 1 Jul 2013 at 1:03