On Tue, Nov 12, 2002 at 07:20:12PM +0000, Michael Hudson wrote:

"Patrick K. O'Brien" writes:

...

On Tuesday 12 November 2002 11:06 am, Steve Waterbury wrote:

...

Twisted Gurus:

Not yet, but I'd like to be one day. ;-)

...

I'm using the "curry" recipe from the Python Cookbook (pretty much verbatim), and it appears to work fine in my code, but cPickle isn't happy about it:

The short answer is that Python doesn't currently pickle nested class and fucntion objects, i.e. those not defined at the top level of the module.

Unless something's changed since I last looked, you can't pickle functions at all.

...

...

def foo(): ... pass ... import cPickle cPickle.dump(foo, open('testpickle', 'w')) bar = cPickle.load(open('testpickle'))

I'd imagine this would work in most versions of python, given that people have been pickling objects for countless generations.

On Wednesday 13 November 2002 04:50 am, Michael Hudson wrote:

jml@mumak.net (Jonathan M. Lange) writes:

...

...

...

...

def foo():

=2E.. pass =2E..

...

...

...

import cPickle cPickle.dump(foo, open('testpickle', 'w')) bar =3D cPickle.load(open('testpickle'))

I'd imagine this would work in most versions of python, given that people have been pickling objects for countless generations.

Oops, my mistake. But note that whats been pickled is a reference to __main__.foo, not the function object itself:

That's true about all code-like objects. The code itself doesn't get pickled, only the state. The code is retrieved by the unpickler via the references to the location of the code - the package name, module name, class name, function name, etc. -- Patrick K. O'Brien Orbtech http://www.orbtech.com/web/pobrien ----------------------------------------------- "Your source for Python programming expertise." -----------------------------------------------

Steve Waterbury writes:

Michael Hudson wrote:

...

And this is a particularly hopeless case. How did you expect it to work at all? Think like an implementer!

Gee, thanks for the encouragement ... I got it to work, and now I'm not gonna tell you how ... nyah nyah! :^)

Sorry, I wasn't *trying* to be offensive. But if one understands how pickling works, one knows not to try things like this... Cheers, M. -- If you don't use emacs, you're a pathetic, mewling, masochistic weakling and I can't be bothered to convert you. -- Ron Echeverri

I have a stupid device that's talking to a Twisted web server that demands HTTP headers to be capitalized Like-This, however Twisted only capitalizes the first character.. On line 567 in twisted/protocol/http.py (Twisted 1.0.0) what we have right now is: l.append("%s: %s\r\n" % (name.capitalize(), value)) # content-length turns into Content-length But what my stupid device wants to see is: l.append("%s: %s\r\n" % ('-'.join(map(lambda n:n.capitalize(),name.split('-'))), value)) # content-length turns into Content-Length Any chance someone could patch that up? -bob

On Tue, Nov 12, 2002 at 04:03:24PM -0500, Bob Ippolito wrote:

I have a stupid device that's talking to a Twisted web server that demands HTTP headers to be capitalized Like-This, however Twisted only capitalizes the first character.. [...] Any chance someone could patch that up?

Doubtful. http://www.w3.org/Protocols/rfc2616/rfc2616-sec4.html#sec4.2 "Field names are case-insensitive" Jeff

On Tue, 12 Nov 2002 16:40:00 -0500, Bob Ippolito wrote:

Well, it's already doing capitalize(), might as well capitalize it the way that everything else does. It certainly wouldn't break anything to send it with string.capwords(name, '-') instead of name.capitalize().

If I recall correctly, it's doing capitalize() because it was *originally* just sending the headers lowercase, and somebody else had a broken device and/or script that was expecting headers to be formatted exactly as they are now (in particular, Cache-control). If people are really encountering broken devices like this on a regular basis (and they seem to be) perhaps it would be best to make the header-formatting code by default do all-lowercase, but be easily pluggable? This would have the advantage of eliminating a bunch of unnecessary string allocations during HTTP request processing. Another option would be to have the headers decided on a case-by-case basis, by looking up in a hashtable. This would let us have a little popularity contest to determine how existing servers send their "case-insensitive" headers and record each decision... -- | <`'> | Glyph Lefkowitz: Traveling Sorcerer | | < _/ > | Lead Developer, the Twisted project | | < ___/ > | http://www.twistedmatrix.com |