On Mar 27, 2005, at 12:02 AM, Sarwat Khan wrote:
On 26-Mar-05, at 8:43 PM, Bob Ippolito wrote:
Which are inconsistent with PyObjC, because the number of arguments does not equal the number of underscores!
Who wants to write code like that :P You're either using PyObjC or not.
It would be more useful if it had the same API as the actual classes as you would use them from PyObjC. Then you could just drop it in when you need cross-platform code.
So now there are three ways to do it:
1) Use plistlib, which comes with Python anyway, and has a pythonic API (at the expense of having to download it if you're not on a Mac)
2) Use PyObjC, and get the real APIs (at the expense of losing cross-plat support and requiring an extension, but guaranteed compatibility and speed)
3) Use plistservices, and get something that is similar to the real APIs, but mapped in a somewhat ambiguous way, that is different from the PyObjC mapping.
My use of plists are either on a Mac or on a web server. In the second case, #1 and #2 are out (at any rate, it's easier to download plistservices than it is to get the Python's source and extract plistlib). plistlib doesn't make available its ISO 8601 parsing available either. The mapping isn't ambiguous, it's the same as the Java bridge.
It's not that hard to get the plistlib from Python's source.
"doesn't make available its ISO 8601" parsing doesn't really sound like a feature to me. It doesn't make its base64 decoding either :) If the feature was compelling enough to make available, it should be in a "date/time parsing" module, not in a mostly unrelated module that just so happens to need it.
You mean, the same as the Java bridge, where every selector is mapped by a human manually, and is probably *already* unmaintained and dying?
-bob
_______________________________________________ Pythonmac-SIG maillist - Pythonmac-SIG@python.org http://mail.python.org/mailman/listinfo/pythonmac-sig
