"cannot determine Numba type" when calling AOT-compiled function from AOT-compiled function

If you take the AOT example file from the documentation and add another function to it,

@cc.export('runall', '()')
def runall():
    square(4)

you get this traceback when attempting to compile it. From this I conclude that numba does not support nesting AOT-compiled functions. True? If not I would like to see a working example in your docs.

Traceback (most recent call last):
  File "setup.py", line 26, in <module>
    ext_modules=[cc.distutils_extension()])
  File "tenv/local/lib/python2.7/site-packages/setuptools/__init__.py", line 145, in setup
    return distutils.core.setup(**attrs)
  File "/usr/lib/python2.7/distutils/core.py", line 151, in setup
    dist.run_commands()
  File "/usr/lib/python2.7/distutils/dist.py", line 953, in run_commands
    self.run_command(cmd)
  File "/usr/lib/python2.7/distutils/dist.py", line 972, in run_command
    cmd_obj.run()
  File "tenv/local/lib/python2.7/site-packages/wheel/bdist_wheel.py", line 192, in run
    self.run_command('build')
  File "/usr/lib/python2.7/distutils/cmd.py", line 326, in run_command
    self.distribution.run_command(command)
  File "/usr/lib/python2.7/distutils/dist.py", line 972, in run_command
    cmd_obj.run()
  File "/usr/lib/python2.7/distutils/command/build.py", line 128, in run
    self.run_command(cmd_name)
  File "/usr/lib/python2.7/distutils/cmd.py", line 326, in run_command
    self.distribution.run_command(command)
  File "/usr/lib/python2.7/distutils/dist.py", line 972, in run_command
    cmd_obj.run()
  File "tenv/local/lib/python2.7/site-packages/setuptools/command/build_ext.py", line 78, in run
    _build_ext.run(self)
  File "/usr/lib/python2.7/distutils/command/build_ext.py", line 339, in run
    self.build_extensions()
  File "/usr/lib/python2.7/distutils/command/build_ext.py", line 448, in build_extensions
    self.build_extension(ext)
  File "tenv/local/lib/python2.7/site-packages/setuptools/command/build_ext.py", line 199, in build_extension
    _build_ext.build_extension(self, ext)
  File "tenv/local/lib/python2.7/site-packages/numba/pycc/cc.py", line 293, in build_extension
    ext._prepare_object_files(self)
  File "tenv/local/lib/python2.7/site-packages/numba/pycc/cc.py", line 274, in _prepare_object_files
    objects, _ = cc._compile_object_files(build_ext.build_temp)
  File "tenv/local/lib/python2.7/site-packages/numba/compiler_lock.py", line 32, in _acquire_compile_lock
    return func(*args, **kwargs)
  File "tenv/local/lib/python2.7/site-packages/numba/pycc/cc.py", line 200, in _compile_object_files
    compiler.write_native_object(temp_obj, wrap=True)
  File "tenv/local/lib/python2.7/site-packages/numba/pycc/compiler.py", line 198, in write_native_object
    library = self._cull_exports()
  File "tenv/local/lib/python2.7/site-packages/numba/compiler_lock.py", line 32, in _acquire_compile_lock
    return func(*args, **kwargs)
  File "tenv/local/lib/python2.7/site-packages/numba/pycc/compiler.py", line 157, in _cull_exports
    locals={}, library=library)
  File "tenv/local/lib/python2.7/site-packages/numba/compiler.py", line 926, in compile_extra
    return pipeline.compile_extra(func)
  File "tenv/local/lib/python2.7/site-packages/numba/compiler.py", line 374, in compile_extra
    return self._compile_bytecode()
  File "tenv/local/lib/python2.7/site-packages/numba/compiler.py", line 857, in _compile_bytecode
    return self._compile_core()
  File "tenv/local/lib/python2.7/site-packages/numba/compiler.py", line 844, in _compile_core
    res = pm.run(self.status)
  File "tenv/local/lib/python2.7/site-packages/numba/compiler_lock.py", line 32, in _acquire_compile_lock
    return func(*args, **kwargs)
  File "tenv/local/lib/python2.7/site-packages/numba/compiler.py", line 255, in run
    raise patched_exception
numba.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Untyped global name 'square': cannot determine Numba type of <type 'function'>

File "numbatest/multsquare.py", line 26:
def runall():
    square(4)
    ^

[stuartarchibald]

Thanks for the report. The compilation context needs to be made aware of the other function, this can be done using e.g. @njit:

from numba.pycc import CC
from numba import njit

cc = CC('my_module')
# Uncomment the following line to print out the compilation steps
#cc.verbose = True

@cc.export('multf', 'f8(f8, f8)')
@cc.export('multi', 'i4(i4, i4)')
def mult(a, b):
    return a * b

@njit
@cc.export('square', 'f8(f8)')
def square(a):
    return a ** 2

@cc.export('runall', '()')
def runall():
    square(4)

if __name__ == "__main__":
    cc.compile()

duplicate of #3632.

[stuartarchibald]

This is marked as a good first issue as it requires a change to the documentation for AOT mode to explain how to achieve the above (want to call CC compiled function from other CC compiled function, eed to njit to make compiler aware).

Since this is working toward documentation I wonder: @njit is simply an abbreviation for @jit(nopython=True) , correct?

And if you were calling square only from runall can you skip the @cc.export decorator on square, ie not export it?

[stuartarchibald]

Since this is working toward documentation I wonder: @njit is simply an abbreviation for @jit(nopython=True) , correct?

Yes, http://numba.pydata.org/numba-doc/latest/user/5minguide.html#other-things-of-interest, the 5 minute guide is recommended reading.

And if you were calling square only from runall can you skip the @cc.export decorator on square, ie not export it?

I think this would work, but would need to check that it is working for the right reasons (need to make sure that it's not working solely because the LLVM inliner decided to just inline it into runall!).

[pearu]

What would be needed in numba that it could automatically njit-ing the functions such as square in the given example? Could this be achieved at the rewriting stage or later (as type interference is also required here)?