help-smalltalk
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Help-smalltalk] visualgst marathon finished :)


From: Paolo Bonzini
Subject: [Help-smalltalk] visualgst marathon finished :)
Date: Sat, 16 Oct 2010 13:29:54 +0200
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.9) Gecko/20100907 Fedora/3.1.3-1.fc13 Lightning/1.0b3pre Mnenhy/0.8.3 Thunderbird/3.1.3

I pushed a huge refactoring to my repository.  Among the highlights:

- list box selection works properly when pressing Cancel or upon errors

- history pane covers also methods

- single tab strip (to be reevaluated), transcript and workspace on the same strip as browsers. this may reappear later as a bottom sidebar, for now it's gone though.

- multiple tabbed windows

- sometimes faster, sometimes slower.


From the code point of view:

- GtkLauncher is still a singleton, but it is a subclass of GtkShell. i removed all class variable state which is now private to each GtkShell

- BrowserState objects can tell a browser to select that item. This is done with double-dispatch.

- new abstract classes for browser list widgets (which know how to propagate the state to the browser) and browser tabs (to remove a lot of boilerplate)

- new abstract class SourceFormatter for Source/* files

- workspace split into WorkspaceWidget and WorkspaceTab


TODO:

- inspectors, the SUnit browser and the debugger are broken. maybe next week. :) to try you can just cp GtkClassBrowserWidget.st and hack.

- dynamic menus and toolbars for the tabs (kind of a prerequisite for the debugger). gwen, i have no clue here. :) an obvious start is to move _some_ menus and toolbar definitions from GtkVisualGSTTool to GtkLauncherTab, then some design is needed.

- opening windowed browsers should pick a default selection like opening tabbed browsers

- merge finder and state hierarchies, they are dups. the implementor and sender tabs are a bit buggy and there is some duplicated code too. i don't understand the GtkImageModel very well, so i didn't touch this.

- testcases have no hope of passing. i honestly didn't care. maybe it was a bit too early. :/

Paolo



reply via email to

[Prev in Thread] Current Thread [Next in Thread]