Message 168551 - Python tracker

Message168551

Author	loewis
Recipients	Robin.Schreiber, ezio.melotti, loewis, mark.dickinson, mrabarnett, pitrou, rhettinger, skrah
Date	2012-08-19.08:20:38
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1345364452.42.0.0368206324844.issue15722@psf.upfronthosting.co.za>
In-reply-to

Content
hackcheck fixes the "Carlo Verry hack", which goes like this: py> object.__setattr__(str, 'lower', str.upper) py> 'dammit Carlo!'.lower() 'DAMMIT CARLO!' (from http://bugs.jython.org/issue1058) It shouldn't be possible to monkey-patch a builtin type; I believe this is to prevent crashes when self has the incorrect layout. Other than that, I find that an arbitrary restriction, except that setattr/attribute assignment prevent an assignment from occurring, it shouldn't be possible to bypass this by calling __setattr__. So if the restriction could be lifted, it should be lifted in both cases. What specific decimal test is failing?

Content

hackcheck fixes the "Carlo Verry hack", which goes like this:

py> object.__setattr__(str, 'lower', str.upper)
py> 'dammit Carlo!'.lower()
'DAMMIT CARLO!'
(from http://bugs.jython.org/issue1058)

It shouldn't be possible to monkey-patch a builtin type; I believe this is to prevent crashes when self has the incorrect layout. Other than that, I find that an arbitrary restriction, except that setattr/attribute assignment prevent an assignment from occurring, it shouldn't be possible to bypass this by calling __setattr__. So if the restriction could be lifted, it should be lifted in both cases.

What specific decimal test is failing?

History
Date	User	Action	Args
2012-08-19 08:20:52	loewis	set	recipients: + loewis, rhettinger, mark.dickinson, pitrou, ezio.melotti, mrabarnett, skrah, Robin.Schreiber
2012-08-19 08:20:52	loewis	set	messageid: <1345364452.42.0.0368206324844.issue15722@psf.upfronthosting.co.za>
2012-08-19 08:20:39	loewis	link	issue15722 messages
2012-08-19 08:20:38	loewis	create