Raymond Hettinger added the comment:

Oscar, the solution proposed by Serhiy looks like a better choice.

I'm wary of increasing the API complexity of the itertools.  Right now, their 
learnability is aided by having simple signatures and no side-values.

The itertools are modeled on functional tools in other languages with mature 
APIs.  I look to those languages to provide an indication of whether proposed 
features are needed in practice.  AFAICT, there is no precedent for a 
takewhile-with-failed-value combo.

I appreciate your request (especially because it was accompanied by a use case) 
but am going to decline.  IMO, the module as a whole is better served by 
keeping the tools simple and clean. 

If an individual itertool doesn't have an exact fit to a particular use case, 
it may indicate that the programmer would be better served by a simple 
generator which can express the logic more cleanly than a tricked-out itertool 
with side-values.

----------
resolution:  -> rejected
status: open -> closed

_______________________________________
Python tracker <rep...@bugs.python.org>
<http://bugs.python.org/issue18821>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

Reply via email to