[IronPython] bug: list.index broken in transition from ip 0.9.5 to ip 0.9.6

Dino Viehland dinov at exchange.microsoft.com
Mon Dec 12 18:11:00 CET 2005

Thanks for the bug reports - these are related to one of our changes to separate Python equality from CLR equality.  The "in" bug is the same issue as this bug just in a different spot.

Both of these changes are related to a fix to address how to treat python equality & hashing versus CLR equality & hashing.  Because Python frequently uses value equality (rather than reference equality) and allows types to throw from __hash__ and CLR types can't throw from GetHashCode() we had to separate the two.  Obviously I missed a few spots in the implementation where we are still calling the CLR's equals instead of python's (in this case we're actually calling Array.IndexOf).

The destabilization you're seeing now is our push to get all of these risky changes out of the way now before we switch to the beta releases.  Sorry for the inconvenience but hopefully this will result in a better 1.0 all around.  

-----Original Message-----
From: users-bounces at lists.ironpython.com [mailto:users-bounces at lists.ironpython.com] On Behalf Of J. de Hooge
Sent: Monday, December 12, 2005 4:32 AM
To: 'Discussion of IronPython'
Subject: [IronPython] bug: list.index broken in transition from ip 0.9.5 to ip 0.9.6


Yet another disguise of probably the same underlying bug:

aList = [['a']]
anItem = ['a']

print aList.index (anItem)

Traceback (most recent call last):
   at __main__.Initialize() in C:\activ_dell\prog\fun\src\funTry.py:line 4
ValueError: list.index(x): x not in list

The number of workarounds in my code is getting a bit too big now, since I
am using this everywhere.

Unfortunately I'll have to revert to 0.9.5 awaiting these things to be

Guys, you've been making a great product upto now.
My sincere respect for that.
As far as I can see the IP team has been impressively productive.
Don't know what the project pressures are, but please keep taking the time
you need!

I'd personally rather wait some more months for 1.0 than having things done
in a rush.

I propose basic stuff like this is added to the regression test suite.

Kind regards
Jacques de Hooge
info at geatec.com

users mailing list
users at lists.ironpython.com

More information about the Ironpython-users mailing list