urlimport 0.86c-dev

Allow importing files through the web, by adding urls to pythonpath

[ urlimport.py ]
Enables remote module importing through urllib2 and perforce.

author: Jure Vrscaj <[e-mail]>

last contributed versions:
version: 0.72c (myevan <[e-mail]> fixes and improvements)
version: 0.85b (alex <[e-mail]> fixes and improvements)
version: 0.72d (jure <[e-mail]> found a bug and alex fixed it).
homepage: http://urlimport.codeshift.net
license: MIT

download: all the project's files may be downloaded from bitbucket with hg:
hg clone http://bitbucket.org/alexbodn/urlimport

Usage:

import sys, urlimport
urlimport.config(**{
'ssl_key.https://your.url':'key for this url'
'ssl_cert.https://your.url':'cert for this url',
'opener.https://your.url':'opener for this url',
'cache_time.http://your.url':'timedelta to keep files in cache for this url',
'no_cache.url':True, # don't cache modules from this url
'user_agent.url':'user agent header for this url',
'ssl_key.host':'key for this host'
'ssl_cert.host':'cert for this host',
'opener.host':'opener for this host',
'cache_time.host':'timedelta to keep files in cache for this host',
'no_cache.host':True, # don't cache modules from this host
'user_agent.host':'user agent header for this host',
'ssl_key':'key for unspecified urls',
'ssl_cert':'cert for unspecified urls',
'opener':'opener for unspecified urls',
'no_cache':'True, #don't cache modules for unspecified urls
'user_agent':'user agent header for unspecified urls',
'cache_time':'timedelta to keep files in cache for unspecified urls',
#if cache_time is missing, keep forever, but check last_modified and etag
'cache_dir':'root of the cache dir', # if missing, a temp dir will be created,
# if empty, no cache will be used
'py_version_string': '$PYTHON_VERSION', # string in path to be replaced by
# $major.$minor
'debug': int(level of output text. see debug() function)
})
# note: if you supply your own opener, i'd recommand inheriting/implementing the
# status handling features from class DefaultErrorHandler below.
sys.path.insert(0, "http://your.url") #this may already be there, like setting
#PYTHONPATH
#but there is a danger default import
#would try it before
expect to have access to your modules at http://your.url

TODO:
- check what's mpath all about.

TODO later:
- python 3.1 support. preliminary work done
- url import of zipped modules
- support of other urls, like ssh, svn, hg, bzr, cvs etc.
- reload c extensions. well, they don't allways reload in local import either,
but my module does reload locally, so why not remotely?
introspection will help.
- swithch the usage of debug with some logging, and do some cleanup of the
calls.