On Fri, Jun 18, 2010 at 7:21 PM, Floris Bruynooghe <f...@devork.be> wrote: > On Fri, Jun 18, 2010 at 05:56:07PM +0000, Luke Kenneth Casson Leighton wrote: >> rexec, regardless of the "breakage", actually had other benefits, >> most notably demonstrated by overloading import to allow scripts to be >> loaded out of sql databases rather than off the filesystem, on-demand. >> whether sandboxing was broken or not was completely irrelevant, as >> far as the overriding / sub-classing of rexec was concerned. > > Shouldn't this be possible in a clean way with a PEP 302 import hook?
ahh, that would mean writing new code :) rexec.py, _aside_ from that two-line breakage, was just perfect for the task, assuming a level of technical ignorance by the users, which was perfectly reasonable in the circumstances where the product was deployed. l. _______________________________________________ EuroPython 2010 - Birmingham, 17-24 July 2010 - http://www.europython2010.eu EuroPython mailing list EuroPython@python.org http://mail.python.org/mailman/listinfo/europython
