Read and write binary .plist files.
This module is a drop-in enhancement of the plistlib module in the standard libary. There are two different APIs which are roughly equivalent. The legacy API provides an interface very similar to that of plistlib, and allows for the use of this module in code already written to use plistlib. The standard API provides an interface like the json, pickle, and marshal modules in the standard library. Two classes are additionally made available.
writePlist(obj, path_or_file[, binary])
Write obj to path_or_file. If path_or_file is a string, assume it's a path and open that path for writing to.
If binary is True (default: False), write a binary plist file. Otherwise an XML one is written.
writePlistToString(obj[, binary])
Serialize obj to a plist formatted string.
If binary is True (default: False), format as a binary plist file, Otherwise format as an XML one.
readPlist(path_or_file[, binary])
Read an object from a plist formatted file. If path_or_file is a string, assume it's a valid path and open up the file at that location for reading first.
If binary is True, assume a binary formatted plist. If it is False, assume an XML formatted plist. Otherwise, automatically detect the type. The default behavior is to detect the type automatically.
readPlistFromString(s[, binary])
Read an object from the plist formatted string, s.
If binary is True, assume a binary formatted plist. If it is False, assume an XML formatted plist. Otherwise, automatically detect the type. The default behavior is to detect the type automatically.
dump(obj, fp[, binary])
Serialize obj as a property list formatted stream to fp (a .write()-supporting file-like object).
If binary is True (default: False), serialize as a binary formatted plist, otherwise as an XML one.
dumps(obj[, binary])
SSerialize obj to a property list formatted str. The arguments have the same meaning as in dump().
load(fp[, binary])
Deserialize fp (a .read() and .seek()-supporting file-like object containing a property list document) to a Python object.
If binary is True, assume a binary formatted plist. If binary is False assume an XML formatted one. Otherwise, automatically detect the formatting. The default behavior is to detect the formatting.
loads(s[, binary])
Deserialize s (a str instance containing a property list document) to a Python object. The arguments have the same meaning as in load().
Fill()
This allows for the conversion of a Fill type object from binary property lists into a Python object and vice versa. I don't know what Fill objects are for. There are no options or attributes.
UID(value)
This allows for the conversion of UID typed objects from binary property lists into Python objects and vice versa. The value can be any positive integer.
The actual plist format is more restrictive than is enforced here. In particular, key values in a plist dictionary must be strings, which is not enforced. There may be other unenforced restrictions, be reasonable.
plutil notes mention "Output of float precision in binary format". Not sure what that means.
Deduplication isn't as aggressive as it could be, OS X's plutil gets smaller files. Ordering is different too, which might be nice to match.
I figured out details of the binary plist format from examining the perl program pluitl.pl, version 1.6 which is available under the following terms:
Any use is fine with attribution.
Author: Pete M. Wilson Website: http://scw.us/iPhone/plutil/ Email: scw@gamewood.net Copyright: 2007-2008 Starlight Computer Wizardry
Wording on some documentation modeled very heavily on wording in documentation in the Python standard library.