I don't think it's necessary to depend on future, although we can certainly
borrow code from it. But how many places in the code slice a range (and of
them, how many of them have to be xrange in Python 2)?

Aaron Meurer

On Sat, Jan 31, 2015 at 10:39 PM, Chris Smith <[email protected]> wrote:

> I was talking with @asmeurer and @jayesh92 this afternoon about removing
> xrange from the SymPy codebase. There's an interesting dilemma that arises:
> there is nothing in PY2 that will emulate the new range in PY3. The most
> crippling example is that while range in PY3 can accept a slice like
> range(10)[::2], xrange cannot. So in a file where one needs to use an
> xrange-like range, it can be imported from compatibility as range (and
> compatibility supplies the xrange function which will no longer be
> necessary once PY2 is dropped). *But if, in the same file, one slices a
> range an error will be raised because the imported xrange (under the alias
> range) cannot be sliced that way:
>
> from sympy.core.compatibility import range
> range(10**8)  # ok. We imported range which is really xrange so this works
> ...
> range(10)[::2] --> error since xrange cannot be sliced this way
>
> We can write list(range(10))[::2] to work around the issue but then we are
> are forcing people to use PY2 idioms to work around PY2 functions (xrange)
> disguised as PY3 functions (range). That's about as bad as allowing users
> to use PY2 keywords (like xrange) in a codebase where we are trying to keep
> things compatible with PY3.
>
> There's some really good news to help in this (and other issues): the
> "future" project at
> I was talking with @asmeurer and @jayesh92 this afternoon about removing
> xrange from the SymPy codebase. There's an interesting dilemma that arises:
> there is nothing in PY2 that will emulate the new range in PY3. The most
> crippling example is that while range in PY3 can accept a slice like
> range(10)[::2], xrange cannot. So in a file where one needs to use an
> xrange-like range, it can be imported from compatibility as range (and
> compatibility supplies the xrange function which will no longer be
> necessary once PY2 is dropped). *But if, in the same file, one slices a
> range an error will be raised because the imported xrange (under the alias
> range) cannot be sliced that way:
>
> EXAMPLE
> ========
> from sympy.core.compatibility import range
> range(10**8)  # ok. We imported range which is really xrange so this works
> ...
> range(10)[::2] --> error since xrange cannot be sliced this way
>
> We can write list(range(10))[::2] to work around the issue but then we are
> are forcing people to use PY2 idioms to work around PY2 functions (xrange)
> disguised as PY3 functions (range). That's about as bad as allowing users
> to use PY2 keywords (like xrange) in a codebase where we are trying to keep
> things compatible with PY3.
>
> There's some really good news to help in this (and other issues): the
> "future" project at http://python-future.org/ . They have already workd
> through the issues related to this (and other PY2/3 incompatibilities) so
> that in our compatibility file all we would have to do is
>
> if PY3:
>     range = range
> else:
>     from future.builtins import range
>
> And then in a file where we want to use xrange we use the compatibility
> range instead and now the code in our EXAMPLE above will work -- no
> workaround is necessary. That's the way it should be. I would recommend
> that as long as we support PY2 we should include `future` like we included
> `mpmath` until we drop PY2 support.
>
>  --
> You received this message because you are subscribed to the Google Groups
> "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/sympy.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sympy/0e4afab2-6e47-4544-b985-3e670469ff11%40googlegroups.com
> <https://groups.google.com/d/msgid/sympy/0e4afab2-6e47-4544-b985-3e670469ff11%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sympy.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAKgW%3D6J4f7nKVdXjnv25btBaMgO%3Do0_xGm46AMW2XSuxKttbfA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to