Here's an example of how you can extract environment variables from a batch or cmd file without creating a wrapper script. Enjoy.
from __future__ import print_function
import sys
import subprocess
import itertools
def validate_pair(ob):
try:
if not (len(ob) == 2):
print("Unexpected result:", ob, file=sys.stderr)
raise ValueError
except:
return False
return True
def consume(iter):
try:
while True: next(iter)
except StopIteration:
pass
def get_environment_from_batch_command(env_cmd, initial=None):
"""
Take a command (either a single command or list of arguments)
and return the environment created after running that command.
Note that if the command must be a batch file or .cmd file, or the
changes to the environment will not be captured.
If initial is supplied, it is used as the initial environment passed
to the child process.
"""
if not isinstance(env_cmd, (list, tuple)):
env_cmd = [env_cmd]
# construct the command that will alter the environment
env_cmd = subprocess.list2cmdline(env_cmd)
# create a tag so we can tell in the output when the proc is done
tag = 'Done running command'
# construct a cmd.exe command to do accomplish this
cmd = 'cmd.exe /s /c "{env_cmd} && echo "{tag}" && set"'.format(**vars())
# launch the process
proc = subprocess.Popen(cmd, stdout=subprocess.PIPE, env=initial)
# parse the output sent to stdout
lines = proc.stdout
# consume whatever output occurs until the tag is reached
consume(itertools.takewhile(lambda l: tag not in l, lines))
# define a way to handle each KEY=VALUE line
handle_line = lambda l: l.rstrip().split('=',1)
# parse key/values into pairs
pairs = map(handle_line, lines)
# make sure the pairs are valid
valid_pairs = filter(validate_pair, pairs)
# construct a dictionary of the pairs
result = dict(valid_pairs)
# let the process finish
proc.communicate()
return result
So to answer your question, you would create a .py file that does the following:
env = get_environment_from_batch_command('proc1')
subprocess.Popen('proc2', env=env)