Hello,

This is the first reminder (step 3 from https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fails_to_install/#_package_removal_for_long_standing_ftbfs_and_fti_bugs).

If you know about this problem and are planning on fixing it, please acknowledge so by setting the bug status to ASSIGNED. If you don't have time to maintain this package, consider orphaning it, so maintainers of dependent packages realize the problem.

This bug appears to have been reported against 'rawhide' during the Fedora 36 development cycle.
Changing version to 36.

Any good new for this? This currently blocks Fedora-Scientific_KDE-Live livespin:
https://koji.fedoraproject.org/koji/taskinfo?taskID=83062887

Unfortunately, after fixing several issues, I'm blocked on a tricky 16-bit float bug due to GCC ABI changes. This can only be fixed by Julia developers, and we're currently looking into it with them, but it's hard to make any guarantees. The only simple solution would be to be able to use an old GCC but I have no idea if that's possible.

Hello,

This is the second reminder (step 4 from https://docs.fedoraproject.org/en-US/fesco/Fails_to_build_from_source_Fails_to_install/#_package_removal_for_long_standing_ftbfs_and_fti_bugs).

If you know about this problem and are planning on fixing it, please acknowledge so by setting the bug status to ASSIGNED. If you don't have time to maintain this package, consider orphaning it, so maintainers of dependent packages realize the problem.

Currently upgrade to Fedora 36 Beta fails like this:
```
Error: 
 Problem: package julia-1.7.0.0-1.fc36.x86_64 requires libmbedcrypto.so.3()(64bit), but none of the providers can be installed
  - package julia-1.7.0.0-1.fc36.x86_64 requires libmbedtls.so.12()(64bit), but none of the providers can be installed
  - package julia-1.7.0.0-1.fc36.x86_64 requires libmbedx509.so.0()(64bit), but none of the providers can be installed
  - problem with installed package julia-1.7.2-1.fc35.x86_64
  - mbedtls-2.16.12-1.fc35.x86_64 does not belong to a distupgrade repository
  - julia-1.7.2-1.fc35.x86_64 does not belong to a distupgrade repository
```

It looks like the version of julia in Fedora 36 is older than the one in Fedora 35.

Do you have a link to the upstream bug related to GCC ABI changes that is waiting on Julia developers?

It'd be unfortunate to ship Fedora 36 without Julia, should this be considered a release blocker for Fedora 36 itself?

We had investigated this on a chat, but I've just filed an issue to keep track of this: https://github.com/JuliaLang/julia/issues/44829

Julia no longer builds on F36 so I'm unable to even build the versions that were included in previous Fedora releases. We should probably remove the existing package for now, which was built before GCC12 was included in F36, given that it's broken. But I don't know how to do that.

(In reply to Milan Bouchet-Valat from comment #7)
> We had investigated this on a chat, but I've just filed an issue to keep
> track of this: https://github.com/JuliaLang/julia/issues/44829
> 
> Julia no longer builds on F36 so I'm unable to even build the versions that
> were included in previous Fedora releases. We should probably remove the
> existing package for now, which was built before GCC12 was included in F36,
> given that it's broken. But I don't know how to do that.

Why do not you raise a bug into gcc?
This seems to be the usual ping pong game before each gcc release. :-)

The game goes like this: the gcc developers blame the language developers while the language developers blame the gcc developers.

So after the round above the julia developers say that it is not their fault since the same tests work flawlessly in gcc 11.
Now in the round the issue would be to find what do the gcc developers think about this.

According to Julia developers, this is probably not a bug in GCC, it's just that Julia relies on a particular GCC ABI and will need to adapt to the new version.

https://github.com/JuliaLang/julia/issues/44829#issuecomment-1091277273

So forcely avoiding symbol collision between libgcc and julia makes testsuite pass:

https://koji.fedoraproject.org/koji/taskinfo?taskID=85284148
https://koji.fedoraproject.org/koji/taskinfo?taskID=85284173

Milan, would you have a look at the above scratch build? If they are okay, I can push commit to Fedora julia scm and build it.

Yes, thanks, as long as tests pass it's probably OK. Can you make a pull request?

Okay, I will recheck spec file and will make a PR.

https://src.fedoraproject.org/rpms/julia/pull-request/2

Proposed as a Freeze Exception for 36-final by Fedora user mtasaka using the blocker tracking app because:

 julia on F-36 FTI and FTBFS, this affects cantor FTI,  Fedora-Scientific_KDE-Live-36 fails to compose.

FEDORA-2022-559a87e02f has been submitted as an update to Fedora 37. https://bodhi.fedoraproject.org/updates/FEDORA-2022-559a87e02f

FEDORA-2022-559a87e02f has been pushed to the Fedora 37 stable repository.
If problem still persists, please make note of it in this bug report.

FEDORA-2022-e4c44f5a21 has been submitted as an update to Fedora 36. https://bodhi.fedoraproject.org/updates/FEDORA-2022-e4c44f5a21

Discussed in ticket: https://pagure.io/fedora-qa/blocker-review/issue/727

The decision to classify this bug as an AcceptedFreezeException was made:

"We usually allow FTBFS/FTI fixes in as FEs, this should also fix some deliverables on top of that."

FEDORA-2022-e4c44f5a21 has been pushed to the Fedora 36 stable repository.
If problem still persists, please make note of it in this bug report.