Path: csiph.com!x330-a1.tempe.blueboxinc.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!feeder.news-service.com!newsfeed.xs4all.nl!newsfeed5.news.xs4all.nl!xs4all!post.news.xs4all.nl!not-for-mail Return-Path: X-Original-To: python-list@python.org Delivered-To: python-list@mail.python.org X-Spam-Status: OK 0.000 X-Spam-Evidence: '*H*': 1.00; '*S*': 0.00; 'guido': 0.04; 'behavior,': 0.07; 'interpreter': 0.07; 'raised': 0.07; 'python': 0.08; 'bubble': 0.09; 'deliberately': 0.09; 'exceptions': 0.09; 'from:addr:ethan': 0.09; 'from:addr:stoneleaf.us': 0.09; 'from:name:ethan furman': 0.09; 'indicates': 0.09; 'message- id:@stoneleaf.us': 0.09; 'not?': 0.09; 'operator,': 0.09; 'received:gator410.hostgator.com': 0.09; '~ethan~': 0.09; 'exception': 0.12; 'wrote:': 0.14; '(!)': 0.16; '(assuming': 0.16; 'bugs,': 0.16; 'called,': 0.16; 'function)': 0.16; 'hazard': 0.16; 'received:67.18.44': 0.16; 'received:72.11': 0.16; 'received:72.11.125': 0.16; 'received:72.11.125.166': 0.16; 'subject:returning': 0.16; 'sure,': 0.16; 'method.': 0.16; 'result,': 0.16; 'call.': 0.19; 'to:name:python-list': 0.19; 'guess': 0.19; 'header:In-Reply-To:1': 0.21; 'code.': 0.22; 'worked': 0.24; "we'd": 0.24; 'function': 0.25; 'later': 0.26; 'object': 0.26; '(even': 0.29; 'instead': 0.29; 'code,': 0.29; 'nested': 0.30; 'done': 0.32; 'does': 0.33; 'to:addr:python-list': 0.33; 'handling': 0.33; 'too': 0.33; 'rather': 0.34; 'expensive': 0.34; 'header:User-Agent:1': 0.35; 'function.': 0.35; 'several': 0.36; 'thread': 0.37; 'another': 0.37; 'could': 0.38; 'but': 0.38; 'earlier': 0.38; 'eric': 0.38; 'returning': 0.38; 'signal': 0.38; 'subject:: ': 0.38; 'some': 0.38; 'should': 0.39; 'called': 0.39; 'returned': 0.39; 'to:addr:python.org': 0.39; 'getting': 0.40; "couldn't": 0.40; 'easily': 0.60; 'more': 0.60; 'below': 0.61; 'matter': 0.63; 'levels': 0.63; 'received:websitewelcome.com': 0.67; 'stated': 0.67; 'imagine': 0.72; 'directly?': 0.84; 'received:gateway04.websitewelcome.com': 0.91; 'snow': 0.91; 'subject:skip:N 10': 0.93 Date: Tue, 31 May 2011 17:30:18 -0700 From: Ethan Furman User-Agent: Thunderbird 1.5.0.10 (Windows/20070221) MIME-Version: 1.0 To: python-list Subject: Re: returning NotImplemented References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-AntiAbuse: This header was added to track abuse, please include it with any abuse report X-AntiAbuse: Primary Hostname - gator410.hostgator.com X-AntiAbuse: Original Domain - python.org X-AntiAbuse: Originator/Caller UID/GID - [47 12] / [47 12] X-AntiAbuse: Sender Address Domain - stoneleaf.us X-BWhitelist: no X-Source: X-Source-Args: X-Source-Dir: X-Source-Sender: mail.admailinc.com ([192.168.10.136]) [72.11.125.166]:1906 X-Source-Auth: ethan+stoneleaf.us X-Email-Count: 1 X-Source-Cap: dG9idWs7dG9idWs7Z2F0b3I0MTAuaG9zdGdhdG9yLmNvbQ== X-BeenThere: python-list@python.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: General discussion list for the Python programming language List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Newsgroups: comp.lang.python Message-ID: Lines: 26 NNTP-Posting-Host: 82.94.164.166 X-Trace: 1306887447 news.xs4all.nl 49042 [::ffff:82.94.164.166]:56503 X-Complaints-To: abuse@xs4all.nl Xref: x330-a1.tempe.blueboxinc.net comp.lang.python:6769 Eric Snow wrote: > Guido indicates earlier in the thread that NotImplemented is used so > that you know that it came from the function that you directly called, > and not from another call inside that function. Why does it matter if > it came directly from the function or not? And couldn't the > NotImplemented still have come from a call inside the operator, rather > than directly? An exception can bubble up from several levels below the original call. Returning NotImplemented (even if returned from another internal function) still has to be deliberately done by the original method. Later in the thread it is also stated that while exception handling performance is not too different in C code, it is very much more expensive in Python code. As to why it would matter if whether the return was from the directly called method vs some nested method... not sure, but I would hazard a guess that by returning NotImplented the interpreter knows everything worked as it should have (assuming no bugs, of course); whereas an exception could easily _not_ have come from the directly called method. Imagine if the methods raised exceptions to signal normal behavior, and some object was buggy -- for __eq__ we'd end up with a False result, instead of getting the exception propagated. (!) ~Ethan~