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

Re: [E-devel] E17 + Accessx + Alt-Tab Bug

On Sat, 11 Mar 2006 22:44:10 -0500 David Carlo <dcarlo@swillers.com> babbled:

> Hello.  I'm running the current CVS version of E17 as of 2006-03-06.  I 
> am a C5 quadriplegic so I type with one "finger".  I use the sticky keys 
> feature in the accessx extension of xorg's X server.  If I hit the Alt 
> key (to stick it) then the Tab key, the next window gets focus and 
> dialog showing a list of windows shows up but does not disappear unless 
> I hit another key.  Is this a bug or a feature?

hmm - well if the keybinding that brings it up has a modifier key - e listens and waits for the modifier to be released before popping down the dialog. You can also it Escape to close it - but it is intended to "stay up" while browsing. as the sticky key feature likely doesn't produce a key down for alt - merely sets the modifier and unsets it afterwards (with the key up event happening before the alt) he has no idea this happened until the tab is pressed (as x then traps the key and sends e the event as opposed to the focused window). so e never gets a chance to see the key be released. you can re-bind window walking to any key combination you like or any key (f1, print screen or any single key on the keyboard) so you don't need to even use stickykeys.). if you do it that way the dialog will come up and stay up until u hit escape (pressing the "tab" key (whatever you bind it to) will advance to the next window in the list.

i hope this helps - i think this is more of a problem with e trying to be VERY generic with the way it can be configured and automagically trying to figure out how best to behave given a configuration - and the stickykeys simply not working the way a normal keyboard would in that situation and thus e's automagic guesses not working perfectly. still i don't think its much of a bug as it is kind of behaving as i would expect it to in your situation.

normally a user first presses the alt key then the tab key, holding alt down - this will activate the winlist popup and list the windows to focus - as long as alt is held down this stays. they just hold alt and keep pressing tab to go to the next window in the list (or alt+shift_tab to go backwards). then on release of alt - it goes away. since you are not able to press 2 keys at once - it to me seems to be doing just the right thing as you may want to advance through more than 1 window - and thus you need to explicitly press escape to abort the winlist dialog as you are not able to provide the implicit hint of "please pop down the winlist dialog now" by releasing alt.

if you want something different you CAN reconfigure bindings to work best for you and your situation. part of being heavily customisable is that not everyone is the same and everyone has different needs. we can't be perfect "one size fits all" and we may have missed things that you may need or want - and over time and versions i hope we can add them. we cant do it all overnight, but you can change things to work better. as i see it it sounds like it is actually working perfectly well as it is given how it works when you can press 2 keys at once. :) what you may be asking for is an action which simply and immediately focuses the next or previous window without dealing with the winlist. it's a little inconvenient for u at the moment as u have to press an extra "escape" at the end - as opposed to release alt, but if u wanted to go thru a list of 5 windows and move 5 along - the winlist would work better as-is. i wont be adding the implied action u are after for e17 - as we have a freeze on features until release, but when we do e18 i am sure we will be filling in all such gaps - so i ask you to hold on as you are and configure things to work better or adapt (and realise that the code is working the way it does to make certain things faster/easier - but may just not be exactly as u expect for you), and then when e17 is released and we start going "send us your ideas for how to improve e17 for e18 and why we should do it - give examples" so we can get a list of what people are missing/wanting - definitely speak up then and we will try to address it.

> Thanks in advance for any help.
>     --Dave

------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster@rasterman.com
Tokyo, Japan (東京 日本)