Startup time is regressing

[Matt Mackall]

On Mon, 2010-11-08 at 17:16 +0000, Antoine Pitrou wrote:
> Matt Mackall <mpm <at> selenic.com> writes:
> > 
> > Some other numbers for comparison:
> > 
> > python 2.6.6 --version: 0.00501s
> > python 2.6.6 -c 'print "hello world"': 0.02431s
> > 
> > python 3.1.2 --version: 0.00536s
> > python 3.1.2 -c 'print("hello world")': 0.04105s
> 
> You should use the -E version to Python, so that site-local settings (e.g.
> PYTHONPATH) don't interfere in the measurement.

$ time (for i in `seq 100`; do python2.4 -E -c 'print("hello world")'
> /dev/null; done)

real	0m1.484s
user	0m0.463s
sys	0m0.753s

$ time (for i in `seq 100`; do python2.5 -E -c 'print("hello world")'
> /dev/null; done)

real	0m1.712s
user	0m0.640s
sys	0m0.870s

$ time (for i in `seq 100`; do python2.6 -E -c 'print("hello world")'
> /dev/null; done)

real	0m2.379s
user	0m0.957s
sys	0m1.210s
12:08calx~$ time (for i in `seq 100`; do python3.1 -E -c 'print("hello
world")' > /dev/null; done)

real	0m4.045s
user	0m1.983s
sys	0m1.813s

Spot the trend. No Python 2.7 or 3.2 in Debian unstable.

-- 
Mathematics is the supreme nostalgia of our time.