string repr in 2.1

[Michael Hudson]

Robin Becker <robin at jessikat.fsnet.co.uk> writes:


> On related notes if repr no longer does octal escapes does string
> input still accept them?

Yes!  Upgrading isn't meant to be *that* much of an adventure.

> Also my trial of the 2.0 code versus an initial 2.1 coding reveals
> that an re based approach is very slow :( Anyone know simply what's
> involved in doing a codec?

Try this instead:

def _make_dict():
    d = {}
    for c in range(ord(' ')) + range(0x7F,0xFF):
        d[chr(c)] = "\\%03o"%(c,)
    for c in map(chr,range(ord(' '), 0x7F)):
        d[c] = c
    d['('] = '\\('
    d[')'] = '\\)'
    d['\\'] = '\\\\'
    return d

_d = _make_dict()

def _escape21_mwh(self, s):
    return ''.join(map(lambda c,d=_d: d[c], s))

It's faster than your re based version, but not as fast as _escape20;
I get:

$ /usr/local/src/Python-2.0/build-O2/python becker-doit.py 1000
Time for 1000x4 uses of <function _escape20 at 0x81ebb34> = 0.57"
Time for 1000x4 uses of <function _escape21 at 0x81e2174> = 2.30"
Time for 1000x4 uses of <function _escape21_mwh at 0x81e78f4> = 0.94"

It can be rewritten in C if you're really keen, when it'll probably
run somewhat faster than _escape20 (I have a very similar function in
pyrepl).

obligatory-jwz-quote-about-regexps-goes-here-ly y'rs
M.

-- 
  Any form of evilness that can be detected without *too* much effort
  is worth it...  I have no idea what kind of evil we're looking for
  here or how to detect is, so I can't answer yes or no.
                                       -- Guido Van Rossum, python-dev