Re: [Python-Dev] Deprecating undocumented, unused functions in difflib.

2010-12-01 Thread Terry Reedy
On 12/1/2010 8:22 PM, Nick Coghlan wrote: On Thu, Dec 2, 2010 at 6:23 AM, Terry Reedy wrote: It would be easiest to just remove the two lines above. Or should I define functions _xxx names that issue a deprecation warning and attach them as attributes to each object? (Defining instance methods

Re: [Python-Dev] Deprecating undocumented, unused functions in difflib.

2010-12-01 Thread Terry Reedy
On 12/1/2010 8:22 PM, Michael Foord wrote: I would still be tempted to go through a single release of deprecation. You can add a test that the names are gone if the version of Python is 3.3. When the tests start failing the code and the tests can be ripped out. I was wondering how people remem

Re: [Python-Dev] Deprecating undocumented, unused functions in difflib.

2010-12-01 Thread Michael Foord
On 01/12/2010 20:23, Terry Reedy wrote: Difflib.SequenceMatcher object currently get two feature attributes: self.isbjunk = junk.__contains__ self.isbpopular = popular.__contains__ Tim Peters agrees that the junk and popular sets should be directly exposed and documented as part of the api, the

Re: [Python-Dev] Deprecating undocumented, unused functions in difflib.

2010-12-01 Thread Nick Coghlan
On Thu, Dec 2, 2010 at 6:23 AM, Terry Reedy wrote: > It would be easiest to just remove the two lines above. > Or should I define functions _xxx names that issue a deprecation warning and > attach them as attributes to each object? (Defining instance methods would > not be the same). Given that f

[Python-Dev] Deprecating undocumented, unused functions in difflib.

2010-12-01 Thread Terry Reedy
Difflib.SequenceMatcher object currently get two feature attributes: self.isbjunk = junk.__contains__ self.isbpopular = popular.__contains__ Tim Peters agrees that the junk and popular sets should be directly exposed and documented as part of the api, thereby making the functions redund