Shared library loading in ROOT6

[klendathu2k]

Exclusively use #pragma statements to load shared libraries for ROOT6, to avoid problems when we mix TSystem::Load with #prama cling load.

[genevb]

Looks like you're doing the same thing that you proposed in PR #751 , but that was for a Cling-compiled macro, while this will be for cons-compiled code. I'm not aware of a reason to oppose, other than Dmitri's comments in PR #751 a couple weeks ago.

[genevb]

Looks like you removed the gSharedLibs line that was forcing these includes, so perhaps remove these includes too?

[klendathu2k]

They are related, and I think this PR may be needed to resolve the CI issues with #751.

Looks like you're doing the same thing that you proposed in PR #751 , but that was for a Cling-compiled macro, while this will be for cons-compiled code. I'm not aware of a reason to oppose, other than Dmitri's comments in PR #751 a couple weeks ago.

[plexoos]

@klendathu2k Do you use a specific test command to verify the change? If so, could you share it?

[klendathu2k]

setup 64b setenv GROUP_DIR /star/u/jeromel/work/STAR/group starver SL24x config/v0.3.0-rhel7-root6.24.06 root4star ... throws no errors.

…

On 2025-06-18 11:08, Dmitri Smirnov wrote: plexoos left a comment (star-bnl/star-sw#753) [1] @klendathu2k [2] Do you use a specific test command to verify the change? If so, could you share it? -- Reply to this email directly, view it on GitHub [1], or unsubscribe [3]. You are receiving this because you were mentioned.Message ID: ***@***.***> Links: ------ [1] #753 (comment) [2] https://github.com/klendathu2k [3] https://github.com/notifications/unsubscribe-auth/ANL4LVB3X6YWPKWZVMU7T7L3EF6G7AVCNFSM6AAAAAB7NRUTV2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDSOBUGYYTSMJQG4

[plexoos]

Thanks! Which commands should I run to reproduce the issue?

[plexoos]

... to avoid problems when we mix TSystem::Load with #prama cling load

I cannot reproduce the issue. I don't see any problem running the following commands:

[cvmfswrite01] ~/> setup 64b
[cvmfswrite01] ~/> setenv GROUP_DIR /star/u/jeromel/work/STAR/group
[cvmfswrite01] ~/> starver SL24x config/v0.3.0-rhel7-root6.24.06
[cvmfswrite01] ~/> which root4star
/star/nfs4/AFS/star/packages/SL24x/.sl73_x8664_gcc485/bin/root4star
[cvmfswrite01] ~/> root4star
   ------------------------------------------------------------------
  | Welcome to ROOT 6.24/06                        https://root.cern |
  | (c) 1995-2021, The ROOT Team; conception: R. Brun, F. Rademakers |
  | Built for linuxx8664gcc on Sep 02 2021, 14:20:23                 |
  | From tags/v6-24-06@v6-24-06                                      |
  | With g++ (GCC) 4.8.5 20150623 (Red Hat 4.8.5-11)                 |
  | Try '.help', '.demo', '.license', '.credits', '.quit'/'.q'       |
   ------------------------------------------------------------------

 *** Start at Date : Fri Jun 20 10:11:44 2025
root [0] 

[genevb]

@ plexoos , to be clear, you should use the standard NFSv4 GROUP_DIR to experience the issue: /star/nfs4/AFS/star/group Jerome made a fix for his GROUP_DIR which he does not consider a final solution.

…

On Jun 20, 2025, at 10:13 AM, Dmitri Smirnov ***@***.***> wrote: I cannot reproduce the issue. I don't see any problem running the following commands: ... [cvmfswrite01] ~/> setenv GROUP_DIR /star/u/jeromel/work/STAR/group

[plexoos]

@ plexoos , to be clear, you should use the standard NFSv4 GROUP_DIR to experience the issue: /star/nfs4/AFS/star/group

Thanks for pointing that out, Gene. It wasn’t clear to me from the previous messages that GROUP_DIR specifically needed to be set to that path.

That said, since I’ve already tested it using the fixed setup scripts and couldn’t reproduce the issue, it makes me wonder: Is there really a need to modify the code, if adjusting the environment resolves it?