-
Notifications
You must be signed in to change notification settings - Fork 62
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Build fails with PHASTA code #392
Comments
Writing down what (little) I've figured out, focusing on the first error message, it's complaining specifically about the destruction of Line 101 in 97523f8
Things I still don't understand:
I'm guessing the answer to most of these questions lies somewhere in C++ lore I've yet to become familiar with... |
Hi @jrwrigh. What repo/branch/commit is this? |
This is on master |
Thanks. To confirm, you are using SCOREC/core master @ 97523f8 (the current HEAD)? |
Actually, in double checking, it's the branch in #391 so that I can actually build it with CGNS. If I run with CGNS enabled on On the #391 branch, CGNS disabled compiles fine, but running with CGNS causes the above error messages. |
Ooooh, I think I know what's happening. It's the change in the C++ standard. Currently CGNS "requires" C++14, whereas the rest is C++11. I'm guessing that change in standard is what's causing the issues. Specifically looking at this part of the CMakeLists code: Lines 40 to 45 in 97523f8
Trying to compile with C++11 flags and CGNS enabled runs into different compiler errors, specifically with the CGNS source files ( |
@cwsmith who is the maintainer of phLinks? I see @seegyoung as possibly the original author. While I was not familiar at first with phLinks, it looks like it is what builds all the part boundary stuff so it is rather critical. |
@jrwrigh @KennethEJansen I'm probably not going to have time to look at this in depth until later next week or the following week. I can provide brief inputs and answers until then. On that front, I'll first look into the C++ standards mentioned. |
Understood @cwsmith . Actually, I was hoping that we could get this delegated to someone else given your scarce time OR determine that C++14 compliance for SCOREC/core was not in the near term achievable horizon in which case we should try to get the CGNS back compatible to C++11. To be clear, we have no expectation of you picking up support of CGNS issues and won't be bugging you about that. Further, we might ask @a-jp for help (though I expect his support time is also scarce since this was a very valuable volunteer project that should not come with expectations of ongoing support). We as the consumers of this development will have to shoulder the load to get it compatible with phLinks and other C++11 blocked components or find a way to bypass them (since PETSc will likely be building its own version of the part boundary interfaces and does not really need us to do that). |
Looking into it, there's nothing about the CGNS library that's C++, so I'm guessing that it's simply that the source code of I'll spend a bit of time just taking whatever the compiler's suggestion to fix the non-C++11 compliant issues are and see if that resolves the issues. |
Thanks. This seems the best approach and hopefully won't offend @a-jp too much. If it does, we can keep it on a branch that we use until SCOREC/core is ready to be C++14 compliant. |
@jrwrigh Would you please share the cmake configuration command (i.e., |
|
HI I am using the version tag/2.2.7 and run into the same error as discussed here. I am using GCC/11.3.0. I am looking forward to updates:-) |
PR #420 should fix the |
#420 was just merged into develop. If the nightly tests pass I'll mark this as closed. |
Trying to build core, I get:
This is with GCC 13.1.1, which might have something to do with it. I'm not super comfortable with C++, so I'm not quite sure what the issue is (other than something is being accessed outside array bounds). I'll take a look at it, but I'm not sure why I'd suddenly be running into issues.
The text was updated successfully, but these errors were encountered: