Benjamin Peterson <benja...@python.org> added the comment: 2010/2/11 Brian Harring <rep...@bugs.python.org>: > > Brian Harring <ferri...@gmail.com> added the comment: > > @benjamin: > > Tests can be written; the reason this patch doesn't bundle tests up front is > that I wasn't going to burn the time till I knew they were needed since I > expected the concept to require some debate. > > As for the hacks angle, there isn't anything hackish there- hackish is what > you have to do w/out either of these patches > (http://www.pkgcore.org/trac/pkgcore/browser/snakeoil/snakeoil/caching_2to3.py). > I understand not everyone likes classes mixed on the fly, but it's a > perfectly valid technique w/ many, many valid uses.
While it may be a valid pattern, I still think there are more elegant solutions. > > Converting the innards of 2to3 over to a generalized plugin interface frankly > is overkill from where I'm standing- being able to plugin new fixers is one > thing (very useful in my opinion but orthogonal to what this ticke tis > about), being able to plugin on the fly different refactoring outputers is > another thing. I plan to work on this this spring or summer if I have time. > > Personally what's in place w/in allow_alternate_output_tools.patch is more > than enough for the crazy ideas I've got kicking around. > > If you've got a specific complaint in the patch, an area that is dodgy/hacky, > state it please- I'm more than open to actionable criticisms. I think delegation would be more appropriate than adding mixin classes at runtime. ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue7826> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
