Author gvanrossum
Recipients andymaier, benjamin.peterson, chris.jerdonek, cvrebert, docs@python, ethan.furman, ezio.melotti, gvanrossum, mark.dickinson, martin.panter, mikehoy, rhettinger, steven.daprano, terry.reedy
Date 2014-09-06.17:17:57
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1410023878.27.0.0816209916129.issue12067@psf.upfronthosting.co.za>
In-reply-to
Content
- This bug should discuss doc updates, not question the rules.

- The rules have evolved over time and the docs stayed behind.

- We should definitely update the 2.7 docs as well as the 3.4 and 3.5 (in development) docs.  The 2.7 docs need to be different than the 3.x docs.

- The language reference manual should clearly state the rules so that implementers can use them as guidelines for implementation.

- There are several sets of relevant rules:

(a) How is each operator translated into a series of lookups and method calls, etc.  It's similar to other binary operators except that the reverse for __lt__ is __gt__ instead of __rlt__, and there's an extra rule that if __ne__ doesn't exist we compute __eq__ and take the opposite.

(b) The default implementation (e.g. default == falls back to 'is', < raises TypeError).

(c) The rules for built-in types, especially numbers (if there are still special cases that aren't explained by the __xx__ methods on the various numeric types).
History
Date User Action Args
2014-09-06 17:17:58gvanrossumsetrecipients: + gvanrossum, rhettinger, terry.reedy, mark.dickinson, benjamin.peterson, ezio.melotti, steven.daprano, cvrebert, chris.jerdonek, docs@python, ethan.furman, martin.panter, mikehoy, andymaier
2014-09-06 17:17:58gvanrossumsetmessageid: <1410023878.27.0.0816209916129.issue12067@psf.upfronthosting.co.za>
2014-09-06 17:17:58gvanrossumlinkissue12067 messages
2014-09-06 17:17:57gvanrossumcreate