changelog

Python Development Environment (Python IDE plugin for Eclipse). Features editor, code completion, re

	* bike/refactor/renameMethod.py: Added prompt
	functionality. 

2002-10-02  Phil Dawes  <[e-mail]>

	* bike/query/getTypeOf.py: Added functionality to handle scanning
	for types of references (needed by getReferencesToMethod)

	* bike/query/getReferencesToMethod.py: Implemented to use
	common.py stuff

2002-09-30  Phil Dawes  <[e-mail]>

	* bike/query/common.py: Split out common query code into common.py

2002-09-27  Phil Dawes  <[e-mail]>

	* bike/refactor/renameFunction.py: Now uses the getReferences stuff
	
2002-09-26  Phil Dawes  <[e-mail]>

	* bike/transformer/WordRewriter.py: Split this class out of the
	renameClass module.	

	* bike/transformer: Added new transformer package to contain code
	which rewrites the internal representation of the sourcecode

	* bike/bikefacade.py: Implemented locateNode to use the new
	fastparserast stuff.

	* bike/parsing/extended_ast.py: Added code to set fqn (fully
	qualified name) attributes on the fastparser ast

2002-09-25  Phil Dawes  <[e-mail]>

	* bike/refactor/renameClass.py: Finished modification to use
	getReferencesToClass. 

2002-09-06  Phil Dawes  <[e-mail]>

	* bike/query/getTypeOf.py: Refactored getType() functionality out
	of ast classes and into a query module.

	* bike/parsing/fastparser.py: Added parsing of import and from
	lines to fastparser, and optimised a little.

2002-09-04  Phil Dawes  <[e-mail]>

	* bike/refactor/renameClass.py: Started modification to use
	getReferencesToClass. 

2002-08-30  Phil Dawes  <[e-mail]>

	* bike/query/getReferencesToClass.py: New module which returns a
	sequence of references to a class

	* bike/parsing/fastparser.py: Wrote new parser which builds a tree
	of functions and classes, very quickly.

2002-08-14  Phil Dawes  <[e-mail]>

	* bike/parsing/tokenutils.py: Fixed bug where attempting to
	tokenize small strings would result in attributeerror

2002-08-05  Phil Dawes  <[e-mail]>

	* bike/logging.py: Logging module. Eventually will be replaced by
	import logging, but for now is a copy of the code by Vinay Sajip
	(to save people from having to download the module)

	* bike/parsing/brmtransformer.py: Added simple fix - tuple being
	concatinated to list was causing problems


---------------------- 0.6.7 ------------------------------------
	
2002-07-31  Phil Dawes  <[e-mail]>


	* ide-integration/BicycleRepairMan_Idle.py: Added Canis fix for
	the bug which stopped brm from displaying the menu when starting
	idle via right-click file.

	* bike/bikefacade.py: Added Canis windows compatability fix for
	dealing with .pyw files

	
2002-07-25  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Fixed bug where extracting an
	expression from a line just below a block opener (e.g. if foo:)
	would cause a parse error.

2002-07-21  Phil Dawes  <[e-mail]>

	* README.idle: Doc about problems on windows

---------------------- 0.6.6 ------------------------------------
	
2002-07-19  Phil Dawes  <[e-mail]>

	* bike/parsing/doublelinkedlist.py: Optimised to use a list rather
	than a class instance for each link element.

2002-07-17  Phil Dawes  <[e-mail]>

	* bike/parsing/typeinfo.py: Fixed bug - package wasn't checking
	parent packages (and Root) for types. This meant that modules
	imported from higher up the tree weren't being found.

	* bike/parsing/extended_ast.py: Renamed 'Top' to 'Root'

---------------------- 0.6.5 ------------------------------------

2002-07-15  Phil Dawes  <[e-mail]>

	* ide-integration/bike.el: Minor bug fix

2002-07-12  Phil Dawes  <[e-mail]>

	* ide-integration/bikeemacs.py: Added code so that it waits for
	buffer to be reloaded in emacs before sending the request to
	update the next one.

	* ide-integration/bike.el: Added bug fix - if prompt for rename
	was on file not loaded, emacs would send a message back to brm
	when it ran find-file, which buggered up the 'rename-this-method?' 
	interaction.

2002-07-11  Phil Dawes  <[e-mail]>

	* ide-integration/bike.el: Added support for GNU emacs

	* README.emacs: renamed from README.xemacs

	* 

2002-07-09  Phil Dawes  <[e-mail]>

	* bike/bikefacade.py: Added code to handle more windows filename
	brokenness

2002-07-05  Phil Dawes  <[e-mail]>

	* bike/refactor/renameMethod.py: Added same optimisation to
	renameMethod.

	* bike/refactor/renameClass.py: Added optimisation: source string
	is scanned for classname before visiting ast. (with the lazy ast
	creation this means that files aren't parsed unless they contain
	the keyword (or are linked to the source via function call).

	* bike/parsing/extended_ast.py: Refactored .source member into
	.getSource() method and implemented so that tokens are
	synchronised with source if they're modified. 

2002-07-03  Phil Dawes  <[e-mail]>

	* bike/parsing/extended_ast.py: Added lazy parsing of ast when
	source.getModule() is called.

	* bike/parsing/typeinfo.py: Added PackageTypeInfo which deduces
	child types by interrogating the package rather than relying on
	children to register themselves in the typeinfo. This was required
	to facilitate lazy parsing of ast. (since parse would have had to
	be done to add module node to package). A lookup of a module now
	causes getModule() to be called on the sourcefile, thus triggering
	parse of source into ast.

---------------------- 0.6.4 ------------------------------------

2002-06-27  Phil Dawes  <[e-mail]>

	* bike/parsing/tokenutils.py: Added bug fix for superflous
	brackets. Basically, if brackets are unecessary (e.g. '(a).foo()')
	then they aren't in the parse tree but are in the token
	stream. This buggers up the token matching.
	This fix just ignores unmatching ')'s. It was the simplist thing
	that could possibly work, but may cause problems later - needs
	lots of testing.

---------------------- 0.6.3 ------------------------------------
	
2002-06-26  Phil Dawes  <[e-mail]>

	* bike/parsing/tokenhandler.py: Fixed bug where getattr renamed
	all occurences of the methodattr in the rest of the tokens

2002-06-21  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Fixed bug where commend before
	the extracted block would make brm think it should be extracting
	an expression rather than a block.

	* bike/parsing/addtypeinfo.py: Fixed the same bug in _getImportedType

	* bike/refactor/renameClass.py: Fixed bug where the module search
	routine in the 'From foo import *' code was using the parent scope
	as the search position. This doesnt work if 'from' is nested in a
	class since parent scope is a module, not a package. Added method
	_getParentPackage() to handle this correctly.

2002-06-13  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Fixed the 'loop' bug. extract
	method will now recognise variables assigned to by the extracted
	function and used in a loop by the calling function.

2002-06-12  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Lots of refactoring prior to
	fixing a bug.

2002-06-10  Phil Dawes  <[e-mail]>

	* bike/parsing/doublelinkedlist.py: Rewrote 'insert' so that it
	doesnt move the iterator. Renamed old method to
	insertAndPositionIteratorOnNewElement.

2002-06-07  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Fixed the parse error when
	extracting code from an if...else... block. The problem was that
	the code below the extraction contained the 'else', but not the
	'if'. This is hacked around by inserting an 'if 1: <indent>
	pass'. This seems to work, but I welcome a better way of handling
	this.

	* bike/parsing/output.py: Added code to ensure that indent never
	dips below 0. This fixes a bug in parsing during method extracts
	out of nested blocks.

	* bike/refactor/extractMethod.py: Fixed the 'don't know what type
	this isAssAttr(Name('item'), 'decl', 'OP_ASSIGN')' bug

---------------------- 0.6 --------------------------------------
	
2002-05-28  Phil Dawes  <[e-mail]>

	* ide-integration/test/*: Added manual test script for testing ide
	integration before a release
	
	* ide-integration/BicycleRepairMan_Idle.py: Now imports code into
	brm when loaded, or saved for the first time.

	* ide-integration/bike.el: Now imports all loaded python files
	into brm. Also checks to see if file was just updated to avoid
	superflous reloads.

2002-05-27  Phil Dawes  <[e-mail]>

	* ide-integration/bike.el: Added menu for brm

	* bike/parsing/undo.py: Added an internal undo buffer size

	* ide-integration/BicycleRepairMan_Idle.py: Added generic
	exception reporting to idle

	* ide-integration/bikeemacs.py: Added generic exception
	reporting. All exceptions are handled and reported back to the
	user.

2002-05-24  Phil Dawes  <[e-mail]>

	* ide-integration/BicycleRepairMan_Idle.py: Exposed undo to idle

	* ide-integration/bikeemacs.py: Exposed undo to emacs

	* bike/bikefacade.py: Exposed undo functionality to facade

	* bike/parsing/undo.py: Added undo functionality

2002-05-15  Phil Dawes  <[e-mail]>

	* bike/refactor/test_extractMethod.py: Fixed bug in expression
	extracting code (which caused a parse exception with correct code)

	* bike/bikefacade.py: Added public interface

---------------------- 0.5 --------------------------------------

	* ide-integration: Created new directory for integration modules 

2002-05-14  Phil Dawes  <[e-mail]>

	* bike/ui/BicycleRepairMan_Idle.py: Exposed extract function to idle

	* bike/ui/bikeemacs.py: Exposed extract function to emacs

	* bike/refactor/extractMethod.py: Added support for extracting
	expressions into methods (e.g.  a = 1+2   ->  a = self.oneplustwo())
	Added support for extracting functions.

2002-05-13  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Extract Method pretty much
	finished. Now on with the testing...

2002-05-03  Phil Dawes  <[e-mail]>

	* bike/ui/BicycleRepairMan_Idle.py: Added extract method support
	for idle

2002-05-02  Phil Dawes  <[e-mail]>

	* bike/ui/bike.el: implemented extract-method for emacs

	* bike/bikefacade.py: Exposed extract-method to the outside world

	
2002-05-01  Phil Dawes  <[e-mail]>

	* bike/refactor/extractMethod.py: Implemented simple extract
	method

2002-04-28  Phil Dawes  <[e-mail]>

	* bike/parsing/matchTokensToAST.py: The 'new' addtokens. Instead
	of each AST node getting its own slice of tokens, there is only
	one token list and each node has 2 linked list iterators pointing
	to the start and end of its tokens in within it.
	Thus the process of calculating the position of these iterators is
	now called 'token matching' rather than 'adding tokens to ast'.

	Having only 1 token stream vastly simplifies things. There are no
	consistency problems when renaming or moving tokens, and no need
	to 'divide' the comments amongst the AST nodes. 
	This enables one matching algorithm to be able to cater for all
	ast nodes without specialized logic.
	
	* bike/parsing/vector.py, bike/parsing/doublelinkedlist.py:
	containers which support the same bi-directional iterator
	interface. Used in refactoring the parsing module to use a linked
	list of tokens

2002-04-12  Phil Dawes  <[e-mail]>

	* bike/bikefacade.py: Removed use of tokens

---------------------- 0.4.3 ------------------------------------

2002-04-11  Phil Dawes  <[e-mail]>

	* bike/ui/BicycleRepairMan_Idle.py: fixed import problem in
	windows

	* bike/parsing/load.py: Added code so that if you select a
	directory it will load all the modules and packages in it even if
	it isnt a package (i.e. doesnt have a __init__.py), but would
	recurse into non-package sub directories.

	* bike/ui/*: This is no longer a package, since the only 2 files
	are installed elsewhere.

	* setup.py: Made BicycleRepairMan_Idle.py a root level
	module. This means it can be used with idle by just adding
	[BicycleRepairMan_Idle] to your .idle file.
	

2002-04-06  Phil Dawes  <[e-mail]>

	* bike/parsing/addtokens.py: Addtokens class now takes tokens in
	constructor. The addTokens function is renamed to
	addTokensToASTFromSrc. This will pave the way for Source nodes to
	hold tokens rather than the source as a string

	* bike/parsing/tokenhandler.py: Fixed bug which caused name ast
	node to not be tokenized

	* bike/parsing/load.py: Added check for __init__.py before
	recursing into sub directories. This means that non-package sub
	directories will not be recursed into. (i.e. the user must load
	those seperately)