Closed DimitryAndric closed 1 year ago
Thanks for reporting. It is fixed on 4.11RC in 87d47fc3812743f49438b7771d02c2a5589bdf3f.
You can see a diff here...
https://github.com/LLNL/Silo/commit/be29ddf0352bc8e5a7eecc8772a3acb64dfde18c
You can download a patch by adding .patch to the above url...
https://github.com/LLNL/Silo/commit/be29ddf0352bc8e5a7eecc8772a3acb64dfde18c.patch
So, do this...
wget https://github.com/LLNL/Silo/commit/be29ddf0352bc8e5a7eecc8772a3acb64dfde18c.patch
patch src/hdf5_drv/silo_hdf5.c be29ddf0352bc8e5a7eecc8772a3acb64dfde18c.patch
Fix will be in upcoming 4.11.1 release aiming for end of Jan, 2023.
Thank you, I will submit the fix for inclusion in the FreeBSD port, until it gets updated to 4.11.1.
During a recent run of the FreeBSD ports tree with clang 15.0, I was notified that Silo failed to build due to a new warning emitted by this recent version of clang:
I have kept staring at that piece of code, but I am unsure what is supposed to do:
It is complaining about the last
DB_OBJ_CASE()
line, which effectively expands to:Most of the other
DB_OBJ_CASE()
macros have a multidimensional last argument, but the problem is thatDBcurve_mt::yvarname
is declared aschar yvarname[256];
so indeed, attempting tostrdup(m.yvarname[i])
does not really make sense: you should pass a char pointer tostrdup()
, not a plain char.Of course I could pass some compiler option to simply silence the warning, but it would seem better to fix the code to do what is right. However, I do not have a clue as to what is correct, in this case?