-
Notifications
You must be signed in to change notification settings - Fork 205
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
Giving default value to arg causes segfault #265
Comments
FWIW, here's what lldb's backtrace says when running
|
I'm using conda-forge to supply my build of both python and boost, and there seems to be a difference between their old builds and their more recent ones. Until I can pin down the differences, I'm closing this for now. |
I just ran into this exact issue building our software with conda. I was accidentally linking to the system libpython (was weak linking on mac, but for some reason the linker was hard-linking with an rpath also) instead of the conda one. the problem seemed to go away if I just changed to leave symbols unresolved. |
Yeah, at some point they (Anaconda and also conda-forge) started statically linking the python executable, which means that all symbols from As a result, it's a mistake to link your module against any version I don't specifically remember how I fixed my issue described above, but I suspect it was by leaving the symbols unresolved, as you say. |
On my Mac, the following program segfaults as soon as the module is imported:
(I tested with boost-1.68 and 1.69. I haven't checked 1.70 yet.)
The issue seems related to, but not quite the same as #38. Inspired by the comments in that issue, I wrote an even more minimal program that produces a segfault. Merely using
arg("x")=0
induces a segfault.compiler and OS details
Edit: I can also reproduce the segfault on Linux, using Anaconda's gcc-7 compiler:
gcc details and example invocation
The text was updated successfully, but these errors were encountered: