David Boddie wrote: > On Thu Apr 27 17:47:15 CEST 2006, Ian Bicking wrote: > > >>For something a bit more sophisticated, this is a good argument for >>importing from URLs. Not inlined into the code, but if you added say >>'http://svn.classsite.com/repos/student-name' to sys.path, and an import >>hook that knew how to read such sites (I think both importing from svn >>and plain HTTP are possible, since you don't have to do a listdir to >>attempt an import). > > > HTTP isn't too hard. I wrote an import hook to do this a few years ago: > > http://www.boddie.org.uk/python/xhtmlhook/ > > Of course, it does other things besides import .py files over HTTP, but > it shows that it can be done. > > These days it could be implemented more cleanly using the improved import > hook mechanism.
You are right, that wasn't hard at all: http://svn.colorstudy.com/home/ianb/httpimport.py I haven't tested this seriously, and I'm not sure about the performance. Mmm... data loading also fits in. -- Ian Bicking / [EMAIL PROTECTED] / http://blog.ianbicking.org _______________________________________________ Edu-sig mailing list [email protected] http://mail.python.org/mailman/listinfo/edu-sig
