================================================================================ Change log NOTE: -- FIRST number changed when I rewrite kernel of program -- SECOND, when I change user interface, adding new "greate" method -- THIRD, little bugfix, or adding/rewriting self-documentation, or just put some comment in script. Not of all of this little changes need to describe in this section. 1.0.4 -- Some spell check of documentation. Thanks to all helpers. new raise instractions with some help for developers. Bugfix in re.subn() method. 1.0.3 -- A lot of bugfix. If user use unknown for python 'encoding', script try to search substitution in some dictionary, and made some warnings and exceptions. 1.0.2 -- Compatibility with python 2.1 tested and bugfixed, instarction how to turn it on you can see in installation instructions. Rename _AdvancedMO object into NoUniMatchObj, because it may be imported into other prgrams. Bugfix in .group() method of NoUniMatchObj object. More Documentation. 1.0.1 -- Compatibility with python 2.1 restored, but after some action described in installation guide. A little bugfix in NoUniMatchObj() class. Bugfix in compatibility with python 2.2 block. I was tested some important methods in Linux RedHat 8.0 python 2.2 vim 6.1 1.0.0 -- 1) Change internal representation of text. Now internal object not in unicode. This is not comfortable for vim. 2) Killing back compatibility with python 2.1: python 2.2 required; pyhton 2.3 recommended. 3) Now setitem supported step argument (text_obj=[start:stop:step]=sequence) 4) three new methods adding: insert(), remove() and reverse(). There are no more FIXME mark in program, but have some TODO: no sort() and pop() methods. New class NoUniMatchObj() need documentation. 0.2.1 -- Back compatibility block. Warning! This module now already compilable under python 2.1 and I was test some functions (like Text.center()). But more powerfull test I don't do. 0.2.0 -- I found greatest bug in my program when they try to work with multibyte encoding. This bug force me to rewrite many methods which use expressly or by implication +byte_offset feature by vim. Now working of this script don't depend off availability of +byte_offset feature in vim. There are two new methods provide for access to real byte-offset in vim: Text.offset2LC() and Text.LC2offset() NOTE: later, in version 1.0.0, I rewrite this again. :( 0.1.4 -- new interaction method: replace_i(); New containers with exceptions and buttons 0.1.3 -- now all of _RegExp using unicode flag for compile the pattern adding decode() method (just return self.text). __str__() and encode() methods now automatically use self.encoding to encode self.text. Writing documentation 0.1.2 -- improvement interactive() method 0.1.0 -- adding interactive() method 0.0.0 -- initial version: support mutable string and file methods, version info, _RegExp subclass with addition count() method. All working good (I'm hope). A lot of known bugs marked in the script by FIXME comment