libui/TODO.md

- change all private names to uipXxxx
- make it so Windows API calls that do logLastError(), etc. abort whatever they're doing and not try to continue, just like wintable
- figure out what to cleanup in darwin terminate:
	- delegate
	- deleted objects view
- assign control IDs on windows
	- GWL(P)_ID
	- related? [12:25] <ZeroOne> And the blue outline on those buttons [ALL clicked buttons on Windows 7] won't go away
		- I get this too
- make sure all terminology is consistent
- 32-bit Mac OS X support (requires lots of code changes)
- add a test for hidden controls when a window is shown
- SWP_NOCOPYBITS (or was it WS_CLIPCHILDREN?)
	- buttons not in tab get drawover issues
	- buttons in tab without transparent drawing code get copied into the label when stack shown and rehidden
- see if we can clean up the Darwin backend
- make the Unix and Windows backend use uiParentDestroy() instead of relying on the autodestroy behavior of the backends
	- this requires explicitly handling delete events
- verify retainment for uiParents in GTK+

ultimately:
- make everything vtable-based
	- figure out where updateParent() plays into this
		- figure out what to do about custom containers
		- make the code flow of all platforms fully symmetrical
- add some sort of runtime type checking
More TODOs. 2015-04-10 13:29:45 -05:00			`- change all private names to uipXxxx`
			`- make it so Windows API calls that do logLastError(), etc. abort whatever they're doing and not try to continue, just like wintable`
			`- figure out what to cleanup in darwin terminate:`
			`- delegate`
			`- deleted objects view`
Moved the control IDs TODO to TODO.md as that's a long-term TODO; it depends on how container controls play out. 2015-04-10 17:31:40 -05:00			`- assign control IDs on windows`
			`- GWL(P)_ID`
More TODOs. 2015-04-12 11:53:16 -05:00			`- related? [12:25] <ZeroOne> And the blue outline on those buttons [ALL clicked buttons on Windows 7] won't go away`
			`- I get this too`
More of the previous commit. More TODOs. 2015-04-10 22:24:21 -05:00			`- make sure all terminology is consistent`
More TODOs. 2015-04-11 17:24:53 -05:00			`- 32-bit Mac OS X support (requires lots of code changes)`
Fixed a major flaw in the GTK+ implementation I only realized now: by calling gtk_widget_show_all() in uiWindowShow(), we override the user's explicit hide settings! Fix it by calling gtk_widget_show() there (to only show the window), and gtk_widget_show_all() in both the new control constructor (to show the new control initially) and in the container constructor (to show that initially too). 2015-04-11 23:59:32 -05:00			`- add a test for hidden controls when a window is shown`
Fixed build warnings. More TODOs. 2015-04-15 22:14:36 -05:00			`- SWP_NOCOPYBITS (or was it WS_CLIPCHILDREN?)`
More TODOs. 2015-04-12 11:59:39 -05:00			`- buttons not in tab get drawover issues`
			`- buttons in tab without transparent drawing code get copied into the label when stack shown and rehidden`
TODO updates. 2015-04-17 12:19:21 -05:00			`- see if we can clean up the Darwin backend`
Implemented uiParentDestroy() on the Windows and Unix backends. These don't actually use it yet; that'll come later. (Ultimately, there should only be one path into window destruction, which leads to one path in uiParent destruction and one path in child destruction.) More TODOs. 2015-04-17 14:52:28 -05:00			`- make the Unix and Windows backend use uiParentDestroy() instead of relying on the autodestroy behavior of the backends`
			`- this requires explicitly handling delete events`
More TODOs. 2015-04-17 15:02:38 -05:00			`- verify retainment for uiParents in GTK+`
More TODOs. 2015-04-10 13:31:16 -05:00
			`ultimately:`
			`- make everything vtable-based`
			`- figure out where updateParent() plays into this`
			`- figure out what to do about custom containers`
			`- make the code flow of all platforms fully symmetrical`
Fixed build warnings. More TODOs. 2015-04-15 22:14:36 -05:00			`- add some sort of runtime type checking`