Bill Torpey

It's hard to tell from the code included, but there is a race condition setting linger that is documented here: https://github.com/zeromq/libzmq/issues/3252. Our workaround has been to set the linger option...

Note that Mac OS does not define `uuid_generate_time_safe`, but that's OK -- if the application cares it can do something like this: ``` #ifdef wUuid_generate_time_safe if (wUuid_generate_time_safe(tempUuid) == -1) {...

Not sure what to do about the Windows version -- that looks to be an existing issue (windows/wUuid.c): ``` void wUuid_generate_time (wUuid myUuid) { } ```

If you're suggesting getting rid of all the wUuuid stuff, I'm fine with that. That would also mean getting rid of the bogus implementation for Windows, I assume (windows/wUuuid.c): ```...

But that's not the current situation -- again, take a look at windows/wUuid.c and you'll see what I mean.

Just trying to be a good citizen -- since wUuid... functions are already being #define'd, it seemed like the right thing to do was to stick with that convention. In...

Hi Frank: OK, I see what you mean -- the 'wUuid_' functions that are defined are only the ones used internally by OpenMAMA, which are basically these: ``` ./src/mama/c_cpp/src/c/bridge/qpid/transport.c: wUuid_clear...

Here's my take on mac impl -- note that `wUuid_generate_time_safe` is not #define'd, but simply implemented. Turns out we need a way to know whether `wUuid_generate_time_safe` delegates, so we do...

In fact, yes. We use the following uuid functions in OZ: ``` #ifndef WUUID_H__ #define WUUID_H__ #include typedef uuid_t wUuid; #define wUuid_generate uuid_generate #define wUuid_generate_time uuid_generate_time #define wUuid_generate_time_safe uuid_generate_time_safe #define...