[zeromq-dev] ZeroMQ 4.0.4 stable released

[Frank Hartmann]

Pieter Hintjens <ph at imatix.com> writes:

> On Sun, Mar 16, 2014 at 10:30 AM, Frank Hartmann <soundart at gmx.net> wrote:
>
>> The CMake flow did not make it into stable? Was there a special reason?
>
> Stable is a snapshot of git master from a certain point in time, plus
> fixes over time. I assume the CMake flow wasn't there when we started
> the 4.x stable fork. You could start by making sure it works on libzmq
> master. Then we can see about backporting that to the next stable
> release.

Hm I tried, but did not get master to compile with or without cmake :)
Is the "norm" thing now required or did I hit the middle of something?

[  1%] Building CXX object CMakeFiles/libzmq.dir/src/session_base.cpp.o
In file included from /home/frank/curve/libzmq/src/session_base.cpp:31:0:
/home/frank/curve/libzmq/src/norm_engine.hpp:15:34: fatal error: norm/include/normApi.h: No such file or directory

>> Do you have experienced any reasons why the version should vary
>> occasionally?
>
> No. It's not normal; presumably a race of some kind.
>
>> Currently it feels like some sort of race somewhere. Does the client
>> block while the encryption is established or could it happen, that the
>> first message is sent while other mechanism are still exchanging keys?
>
> The client blocks so it may be an error in libzmq.

Ok, I could try to reproduce it with a test_security_curve.cpp like
setup and see howto go from there.

Are there tools around for catching races mechanically? I used hellgrind
once, but while it showed the problem it showed millions of false
problems too, because of not understanding the mutexes used.

Still it is a great help if you are able to sort through the false
positives, which requires deep code knowledge.

I read clang has now a thread checker too? Latest clang is in Debian/sid
only, so I am a bit reluctant to burn my wheezy system... Can anyone
recommend the threadsanitizer?

regards
  Frank