 |
|
| kb_mouse driver status update |
|
|
By looncraz - Posted on April 15, 2008 - 15:25:33 (#22733)
Current version when comment was posted: Final (should be) |
|
|
Should anyone be interested, I am currently testing out a new driver which will repair the kb led/state issue while also providing some support for passing through non-standard interrupts ( i.e. special function keys, such as volume and player controls ).
This will be some time in the coming still, as I have bigger fish to fry in some things which are currently underground ( read: 6 months or so before I release anything at all, except some sporadic testing in the cases I do not have needed hardware ).
Simply post here if you are interested in testing the proper R5 repair or are more interested in getting those extra keys working.
--The loon
|
|
| Cool |
|
|
By BeOS Mr X - Posted on August 27, 2006 - 14:02:28 (#20510)
Current version when comment was posted: Final (should be) |
|
|
Does this mean there will be a 'working' numluck app for the first time ever? I'd like to test that. Good to see you back.
|
|
| BeOS Bug.. |
|
|
By looncraz - Posted on August 27, 2006 - 12:25:13 (#20509)
Current version when comment was posted: Final (should be) |
|
|
I have found the EXACT location and cause of the key state (locks) in BeOS R5. (not present in Dano)
Them issue is with the kb_mouse kernel driver. I have made large strides in rewriting this driver, with mouse support (USB and PS/2.. serial coming) alread in place, but I still have some glitch with modifiers (reporting modifier 239 or something if ANY lock is on, and is somehow not catching certain interrupts).
With any luck, I can get this going with some assistance from the Haiku kb_mouse driver.
When this driver is ready, I will release it for testing.. with inbuilt keyLock memorization. There may also be some limited support for certain keyboads with extended keys. But this will likely work through an extensible architecture in kernel and user-space.
BTW, The Loon is back, and in FULL swing. Expect applications / code from me to be on BeBits within the next few days to the next week.
--The loon
|
|
| useless |
|
|
By BeOS Mr X - Posted on June 23, 2006 - 15:39:59 (#20051)
Current version when comment was posted: Final (should be) |
|
|
All it does is turn on the numlock light, but the numpad buttons don't work like the numbers and everything. So far I've tried all the 'so called' working numlocker utilities which none work properly. An update for this app is encouraged, or if looncraz is still alive he could work with me in email getting it to work with my keyboard.
|
|
| R5 tests... |
|
|
By looncraz - Posted on December 31, 2003 - 11:45:15 (#10247)
Current version when comment was posted: Final (should be) |
|
|
for i in *
do
echo Howdy $i
done
Yeah, good old BASH :-/
Anyway, with the testing with R5, there is still no light at the end of the tunnel.
Even with testing to ensure that the right lights are lit, and then relighting if needed, all 10 max tries fail.
In fact, even use the KeyMaps pref app, the results are the same (even manually).
So, I may need to invert the keymap ( I would prefer to do it live).
So what can be done, is use this and the Inverted keymap.
--The loon
|
|
| Working theory.. |
|
|
By looncraz - Posted on December 22, 2003 - 18:37:24 (#10111)
Current version when comment was posted: Final (should be) |
|
|
Okay,
there seems to be a way it works.. at least sometimes
Try running LoonKeyLocks twice, or more.
See what that does.
I am trying to detemine the best way of detecting when this problem exists, and then how to patch it.
--The loon
|
|
| Re Hmm |
|
|
By InactiveX - Posted on December 21, 2003 - 19:56:43 (#10085)
Current version when comment was posted: Final (should be) |
|
|
I remember the app you're referring to as well.
I am using the commercial version of r5.0.3.
If there's any more info I can supply to you to get this fixed, let me know.
|
|
| Hmm |
|
|
By looncraz - Posted on December 20, 2003 - 13:12:04 (#10062)
Current version when comment was posted: Final (should be) |
|
|
I had the same problem a long time ago with another such app.
I will probably look into the keymaps, and check the locks as reported again, and then making live mods if needed to ensure the proper locks are applied.
For now, this seems to work for most people pretty well.
Certainly works here perfectly.
Are you in R5 by any chance?
--The loon
|
|
| Didn't quite work for me |
|
|
By InactiveX - Posted on December 20, 2003 - 10:15:10 (#10061)
Current version when comment was posted: Final (should be) |
|
|
I installed as described, and when I rebooted, sure enough the NumLock light was on.
However, in StyledEdit the keypad behaved as if the NumLock was off. Pressing the NumLock key and the light remained lit, but now worked as it should.
I am using a UK keyboard and have it set right in preferences.
Any ideas? I would really like to have this working. Thanks.
|
|
| Handy... thanks, Loon |
|
|
By Jace - Posted on December 19, 2003 - 23:26:32 (#10055)
Current version when comment was posted: Final (should be) |
|
|
As this is one of my long standing annoyances with BeOS, this is welcome. Thanks :-)
(I hope OBOS puts this into the OS in a non-hack way, though)
|
|
| OH.. and about Zeta.. |
|
|
By looncraz - Posted on December 19, 2003 - 22:43:54 (#10054)
Current version when comment was posted: Final (should be) |
|
|
I have nothing to do with that Zeta junk any more. I just don't blend well with such efforts apparently. I am better on my own for a while.
--The loon
|
|
| Settings.. |
|
|
By looncraz - Posted on December 19, 2003 - 22:41:43 (#10053)
Current version when comment was posted: Final (should be) |
|
|
I know, I know...
It is just a junk method I've been using for odds and ends..
I have been toying with how I like to store settings, just programming wise...
I'll be modifying it eventually so that some goes into an attribute handling class, and some into class for handling special data (user, owner, permissions, icon, full path, etc...), another class for reading and writing to files, and handling startpoints and such automatically. (So you say prefs->WriteLine("ojhejhe"); or ReadNextLine(); or FindInFile("setting"); or GetValueForTag or something for plain text setting handling, and the of course ReadMessage('_msg'); or however.
As for now, some junk to get me by, so I can get my work done.
Works perfectly well, and that is all that matters!
--The loon
|
|
| settings class |
|
|
By AnonymousCoward - Posted on December 19, 2003 - 14:46:00 (#10049)
Current version when comment was posted: Final (should be) |
|
|
Man, that settings-class is horrible! I sure hope they're not using that in Zeta...
|
|
| memson.. |
|
|
By looncraz - Posted on December 19, 2003 - 12:24:51 (#10048)
Current version when comment was posted: Final (should be) |
|
|
I don't like cross-compiling for different hardware, as I can't reliably test it.
If you would like, you may send me your changes and your binary, and I'll host it here.
Thanks!
--The loon
|
|
| Yes.. |
|
|
By looncraz - Posted on December 19, 2003 - 12:23:40 (#10047)
Current version when comment was posted: Final (should be) |
|
|
This is a true fix.
It does not invert the keymap, or anything of that nature.
It tells the keyboard to set its locks, properly.
I discovered that just doing it that way made it work perfectly.
--The loon
|
|
| Is this any better than using |
|
|
By xfilecsm - Posted on December 19, 2003 - 08:47:10 (#10046)
Current version when comment was posted: Final (should be) |
|
|
the american numlock inverted keymap? The only problem I have with that keymap is that it doesn't automatically work in a terminal. Will this let me bootup with numlock on and working? I tried it early this morning (maybe too early) and it was just like the american numlock inverted keymap.
|
|
| PPC version |
|
|
By memson - Posted on December 19, 2003 - 05:41:28 (#10044)
Current version when comment was posted: Final (should be) |
|
|
Recompiled it for a PPC target - it seems to work fine (baring in mind my apple extended keyboard 2 has a caps lock that stays depressed whist caps lock is 'on' and therefore I can't really test that bit of it... Numlock came on as expected.
To recompile for PPC - remove the default parameters from the method bodies (should only appear in the class declaration, mwcc is strict about this point) and create a PPC 'BeApp' project, remembering to add the metrowerks C++ lib in addition to the source.
Matt
|
|
|
|
|
