On Wed, 06 May 2009 23:33:20 +0100, Luis Alberto Zarrabeitia Gomez
<ky...@uh.cu> wrote:
Quoting Rhodri James <rho...@wildebst.demon.co.uk>:
So the answer to my question would be "nothing"?
Indeed, there is nothing broken with the search and replace feature of
his editor. When he is copying a non-recursive function, it is
_useless_ to do a search and replace. When he is copying a recursive
function, it is _required_ to do a search and replace.
On the contrary, it's merely not very useful for a non-recursive
function. It still changes the function's name, which was one of the
objectives. Seriously, the lack of effort in (say) doing a quick
incremental search for the function's name is a no-brainer.
So, the very same task requires you to either perform a task that will
almost always be useless (or even dangerous), or to read the source code
to find out if the function is recursive, so that you can use the search
and replace only then.
If you're copying code without reading it, you're going to get bitten
anyway some time. My sympathy is still limited.
(I know that the "problem" is present in almost all programming
languages... but that's not what is being discussed. Bearophile's
concerns seem legitimate, and you should not dismiss them so lightly
just because there are ways to do more work and hopefully avoid the
problems. I'd say that the "problem" is even aggravated in python,
where the dynamic nature of the language makes it near to
impossible to build good refactoring tools)
As with that other great "problem" of python, the lack of private
class attributes, a tiny amount of self-discipline solves 90% of the
issues.
--
Rhodri James *-* Wildebeeste Herder to the Masses
--
http://mail.python.org/mailman/listinfo/python-list
