Pierre Augier

> Should Cython be blamed? I would say no, I have no evidence yet that Cython is generating broken code, and the fact that lowering optimization levels make things magically...

Unfortunately, I'm not able to reproduce the issue with conda-forge, even with ``` gcc = "==12.2.0" openmpi = "==4.1.4" ``` which are the version I get with Debian 12. It...

Finally, I was able to avoid the problem with gcc 12.2 provided by Debian with ``` export CFLAGS=-Wsign-compare -DNDEBUG -fwrapv -Wall -fstack-protector-strong -Wformat -Werror=format-security -fwrapv -O3 ``` (which is the...

> One way to hack it without passing `--no-build-isolation` is to add an upper bound to `requirements-build-cython.txt`. It would indeed be a simple and quite clean solution: just change `requirements-build-cython.txt`...

> Why everyone else, running other distros and/or GCC versions, should be penalized to use an older Cython? Note that it is very temporary only to avoid a bad issue...

Current outputs with [d41024f](https://github.com/hpyproject/hpy/pull/492/commits/d41024f6446b134e70c4c3c0cf088b1744c79186): ``` $ make print_pypy ==================================== PyPy ====================================== cpy hpy purepy ---------------- ------------------- ------------------- TestModule::test_noargs 330.17 us 250.28 us [0.76] 88.54 us [0.27] TestModule::test_onearg_None 360.27 us 289.94...

Same thing for [fluidsim](https://foss.heptapod.net/fluiddyn/fluidsim) and [fluidimage](https://foss.heptapod.net/fluiddyn/fluidimage). It's not so surprising that the bug is seen in all these packages since the same methods/options are used. In term of code, one...

That would already be much more convenient than the long version!

I come back to this possible enhancement because it would really be useful (anything shorter than `--manifest-path`)... Could such alias be added?

> We didn't get around it yet, but I'd be happy to review a pull request :) I'm not comfortable with Rust but I can try since I guess it...