[PATCH evemu 02/11] py: Replace print statement with function call

[Peter Hutterer]

On Sat, Jan 04, 2014 at 01:41:57PM +0100, Daniel Martin wrote:
> On Sat, Jan 04, 2014 at 02:14:38PM +1000, Peter Hutterer wrote:
> > On 4/01/2014 01:27 , Daniel Martin wrote:
> > >On Fri, Jan 03, 2014 at 10:24:37AM +1000, Peter Hutterer wrote:
> > >>On Thu, Jan 02, 2014 at 10:47:52PM +0100, Daniel Martin wrote:
> > >>>As of Python 3 the print statement has been replaced by a function and
> > >>>Python 3 has no backward compatibility for it (like Python 2.7), causing
> > >>>syntax errors on such statements.
> > >>>
> > >>>Replace all print statements with the function call and bump the Python
> > >>>package requirement to version 2.7. Because, these changes make the
> > >>>scripts incompatible with Python 2.6.
> > >>>
> > >>>Signed-off-by: Daniel Martin <consume.noise at gmail.com>
> > >>>---
> > >>>  configure.ac                    |  2 +-
> > >>>  src/convert-old-dumps-to-1.1.py | 18 ++++----
> > >>>  src/make-event-names.py         | 94 ++++++++++++++++++++---------------------
> > >>>  3 files changed, 57 insertions(+), 57 deletions(-)
> > >>>
> > >>>diff --git a/configure.ac b/configure.ac
> > >>>index d9f434f..0bd51a9 100644
> > >>>--- a/configure.ac
> > >>>+++ b/configure.ac
> > >>>@@ -31,7 +31,7 @@ AC_PROG_LIBTOOL
> > >>>  AC_PROG_CC
> > >>>  AC_PROG_CXX
> > >>>  AC_PROG_INSTALL
> > >>>-AM_PATH_PYTHON([2.6])
> > >>>+AM_PATH_PYTHON([2.7])
> > >>
> > >>I'd rather keep the min requirement on 2.6 for RHEL6 compatibility and use
> > >>from __future__ import print_function.
> > >
> > >Just installed 2.6(.8) to test this and it works even without the
> > >__future__ import. print is already a function:
> > 
> > the from __future__ import statement removes print as a language statement.
> > so it simply prevents us from accidentally introducing a python3-invalid
> > print statement.
> > 
> > >
> > >     >>> help("__builtin__.print")
> > >     Help on built-in function print in __builtin__:
> > >
> > >     __builtin__.print = print(...)
> > >         print(value, ..., sep=' ', end='\n', file=sys.stdout)
> > >     ---8<---
> > >
> > >I've digged through the NEWS file, but couldn't find any hint since when
> > >this is the case. Anyone else knows some details?
> > 
> > found it: http://docs.python.org/2/library/functions.html#print says
> > available since 2.6.
> > 
> > btw, not sure how you did the conversion but the 2to3 script may uncover a
> > few more things.
> 
> I've made it the classic trial and error way. But, I've tested 2to3 now.
> It would remove the "from __future__ ..." and add a list() around

the removal is because it's not _needed_ but I want it in there anyway :)

> dictionaries, which looks odd:
>     list(some_dict.items())
> No other changes pop up.

that's documented here:
http://docs.python.org/2/library/2to3.html#to3-fixers
though unfortunately it doesn't say _why_ it does it. this one here does
explain it:
http://stackoverflow.com/questions/17695456/why-python-3-needs-wrap-dict-items-with-list
I suspect the code we have in evemu is probably safe (without having checked it
explicitly)

Cheers,
   Peter