Fredrik Lundh wrote: > what part of your observations makes you think that the environment isn't > "captured" (i.e. > copied from the process environment) when the os module is > imported ?
Answer: the part that was informed by a fundamental misunderstanding on my part of how the os module obtains information. Based on Mr. Lundh's 'copied from the process environment' remark and his hint, I would infer the following: 1) In the observations I detailed, the 'process' is the pythonwin application 2) When the pythonwin application is invoked, it obtains a copy of the current environment 3) the import os statement obtains information from the copy of the current environment that was obtained when the pythonwin process began 4) a reload(os) or a combination of del os followed by import os also get information from the copy of the current environment that was obtained when the pythonwin process began My questions: a) are the above inferences (albeit oversimplified) correct? b) when the hint refers to a new process, does the term 'process' refer to the same concept that is described in secion 6.1.5 of the Python Library Reference? -- http://mail.python.org/mailman/listinfo/python-list