Response file needed for Fortran project on Linux

Inspired by #7212.

A project with lots of Fortran source:

#!/bin/sh
set -eu
work_dir=meson_toolong
build_dir=b
n=12000
mkdir -p ${work_dir}
cd ${work_dir}
rm -rf ${build_dir}
{
  echo "project('toolong', 'fortran')"
  echo "srcs = ["
} > meson.build
for i in $(seq -w ${n})
do
  src_file=f${i}.f90
  {
    echo "Subroutine s${i}"
    echo "End"
  } > "${src_file}"
  echo " '${src_file}'," >> meson.build
done
{
  echo "]"
  echo "library('toolong', srcs)"
} >> meson.build
meson . ${build_dir}
cd ${build_dir}
ninja -v

gives

The Meson build system
Version: 0.59.0
Source dir: /home/mat_cross/documents/Meson/resp/meson_toolong
Build dir: /home/mat_cross/documents/Meson/resp/meson_toolong/b
Build type: native build
Project name: toolong
Project version: undefined
Fortran compiler for the host machine: gfortran (gcc 11.2.1 "GNU Fortran (GCC) 11.2.1 20210728 (Red Hat 11.2.1-1)")
Fortran linker for the host machine: gfortran ld.bfd 2.35.2-4
Host machine cpu family: x86_64
Host machine cpu: x86_64
Build targets in project: 1

Found ninja-1.10.2 at /usr/bin/ninja
ninja: fatal: posix_spawn: Argument list too long

(Python is 3.9.6).

[dcbaker]

For me it falls over trying to do the depscan step, which is a python program we maintain. But it does generate a rule for usin ghte fortran linker with a response file, so this actually an issue with the depscanner

[matcross]

Sorry, my fault. It’s looking good from your submit/depscan-file-limit branch. Thanks again for the fix.