BUG (?): SciPy build failure with latest f2py (bisected) #25211
Description
Using latest NumPy main branch at the time of writing we can see the build failure below locally and in CI for SciPy. Let me know if this is expected, for now I've tentatively treated as a bug.
[388/1606] Compiling Fortran object scipy/sparse/linalg/_propack/_spropack.cpython-311-x86_64-linux-gnu.so.p/meson-generated_..__spropack-f2pywrappers.f.o
FAILED: scipy/sparse/linalg/_propack/_spropack.cpython-311-x86_64-linux-gnu.so.p/meson-generated_..__spropack-f2pywrappers.f.o
gfortran -Iscipy/sparse/linalg/_propack/_spropack.cpython-311-x86_64-linux-gnu.so.p -Iscipy/sparse/linalg/_propack -I../scipy/sparse/linalg/_propack -I../../../python_venvs/py_311_scipy_dev/lib/python3.11/site-packages/numpy/_core/include -I../../../python_venvs/py_311_scipy_dev/lib/python3.11/site-packages/numpy/f2py/src -Iscipy/lib_fortranobject.a.p -Iscipy/sparse/linalg/_propack/liblib__spropack.a.p -I/home/treddy/github_projects/spack/opt/spack/linux-ubuntu22.04-skylake/gcc-11.3.0/openblas-0.3.23-vtzm3if4lq22ni6ftwf6xkqrr2zolk3l/include -I/home/treddy/github_projects/spack/opt/spack/linux-ubuntu22.04-skylake/gcc-11.3.0/python-3.11.2-4qncg2nqev5evxfmamde3e6rnb34b4ls/include/python3.11 -fvisibility=hidden -fdiagnostics-color=always -D_FILE_OFFSET_BITS=64 -Wall -std=legacy -O2 -g -Wno-conversion -fPIC -Wno-maybe-uninitialized -Jscipy/sparse/linalg/_propack/_spropack.cpython-311-x86_64-linux-gnu.so.p -o scipy/sparse/linalg/_propack/_spropack.cpython-311-x86_64-linux-gnu.so.p/meson-generated_..__spropack-f2pywrappers.f.o -c scipy/sparse/linalg/_propack/_spropack-f2pywrappers.f
scipy/sparse/linalg/_propack/_spropack-f2pywrappers.f:6:12:
6 | use __user__routines
| 1
Error: Invalid character in name at (1)
scipy/sparse/linalg/_propack/_spropack-f2pywrappers.f:7:12:
7 | use __user__routines
| 1
Error: Invalid character in name at (1)
I bisected on NumPy main branch, at least the patch is small so maybe not too complex to assess the situation:
3e442fc4fb343a6e7a048f4279e1b916dbe5ec65 is the first bad commit
commit 3e442fc4fb343a6e7a048f4279e1b916dbe5ec65
Author: Rohit Goswami <rog32@hi.is>
Date: Sun Nov 19 10:15:35 2023 +0000
BUG: Add use-block handling for common blocks
Closes gh-19161
numpy/f2py/common_rules.py | 2 ++
numpy/f2py/f90mod_rules.py | 5 +++++
2 files changed, 7 insertions(+)
I wonder if it is the repetition of use string.