Reference:

• Failed Fedora update of OpenPGM (because there was no side tag): FEDORA-2025-1160d1ba24

The test on badfuncs fails because of the following two lines/functions in the source code:

• https://github.com/zeromq/libzmq/blob/master/src/udp_engine.cpp#L400 - inet_addr

_raw_address.sin_addr.s_addr = inet_addr (addr_str.c_str ());
/* To be replaced by something like: */
inet_pton (AF_INET, addr_str.c_str (), &_raw_address.sin_addr.s_addr);

• https://github.com/zeromq/libzmq/blob/master/src/udp_engine.cpp#L340 - inet_ntoa

const char *const name = inet_ntoa (addr_->sin_addr);
/* To be replaced by something like: */
char name[INET_ADDRSTRLEN];
            inet_ntop(AF_INET, &addr_->sin_addr, name, INET_ADDRSTRLEN); 

Thanks for having caught this misalignment!

I used the fedpkg build --target=f42-build-side-104367 command, but I guess it was too soon after the build of OpenPGM in the side tag. From what I have experienced, we need to wait at least 5-10 minutes in order for the build to find its way in the repo.

Another build is ongoing: https://koji.fedoraproject.org/koji/taskinfo?taskID=128563713 I checked and this time, it is with OpenPGM 5.3.18, as expected.

From what I have experienced, we need to wait at least 5-10 minutes in order for the build to find its way in the repo.

With on-demand tags, just waiting isn't likely to be sufficient, you really need to run the wait-repo --request command to get Koji to regenerate the side tag repo. This was changed because constantly regenerating every single repo every few minutes was a huge suck on koji's resources.

I'll try and get the guide updated today...thanks for fixing this!

Thanks a lot for having caught the issues and for having provided all these explanations! After having improved the documentation, it will help others (and us) in the future; appreciated :)