Kirikaza <kirikaza@rambler.ru> added the comment:
Are you talking about ... I have read no line of code from urllib module. I just try to use urllib.urlopen() and I see it uses not only http_proxy but also HTTP_PROXY and urlopen() prefers the latter variable.
a specific change in a specific place
Let's consider a two-lines sample proxy.py: import urllib urllib.urlopen('http://python.org/ftp/').readlines() Some tests: $ http_proxy= HTTP_PROXY= proxy.py ['<!DOCTYPE HTML ... \n'] $ http_proxy=lower:80 HTTP_PROXY= proxy.py IOError: invalid proxy for http: 'lower:80' $ http_proxy= HTTP_PROXY=UPPER:80 proxy.py IOError: invalid proxy for http: 'UPPER:80' $ http_proxy=lower:80 HTTP_PROXY=UPPER:80 proxy.py IOError: invalid proxy for http: 'UPPER:80' place: 20.5.1, paragraph 7 (the only one where *_proxy are discussed) change: add some words about HTTP_PROXY and maybe about FTP_PROXY
does urllib2.urlopen ... (in 3.x) I don't have Python 3.x so I can't say anything.
---------- _______________________________________ Python tracker <report@bugs.python.org> <http://bugs.python.org/issue9637> _______________________________________