• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [DEV] Continuing Triumph ICS Development

Awesome work so far! I just successfully built and 3G is working as good as ever (which is only mediocre).

There is definitely something not quite right about the touch screen. As others have reported, it doesn't respond in browser windows, and it doesn't recognize press & hold in the keyboard. Also, if you swipe off the left side of the screen it registers a hit on the top left corner, causing the page to scroll down really fast, but that's probably a separate bug.

I went into the developer options and turned on the pointer location option. It shows your touches and has 6 pieces of data across the top of the screen:

P: 0/2 - changes to 2/2 when the screen is touched
dX: 0.0 - shows the x position of the touch. Left side is 0, right side is 480. When released it changes to a (seemingly) random value. Sometime the value is negative and the field turns red
dY: 0.0 - same as dX but vertical. top is 0, bottom is 800. When released it changes to a (seemingly) random value. Sometime the value is negative and the field turns red
Xv: 0.0 - shows the horizontal velocity. when the touch is released, it shows the velocity at the time of release
Yv: 0.0 - shows the vertical velocity. when the touch is released, it shows the velocity at the time of release
Prs: 1.0 - Always 1.0 and red, not sure what it means or is supposed to be doing, but red is bad, right?
Size: 1.0 - Changes between 0 and 1 when touched and back to 0 when touch is released. For some reason 1 is red.

Screenshot
vGDGL.png


Not sure if this is of any use, but I figure I'd throw it out there. Keep up the awesome work.
 
yumbrad said:
I couldn't stay away :rolleyes:

Got wifi working. More of a pain than data!

https://github.com/yumbrad/android_...mmit/2fb701633389602b87bdd3915ee3aa7f24d09496

https://github.com/yumbrad/android_...mmit/3ab1e4c95449bb4342de12e56e9a4cbbeec3155c
(sorry the branches have different names. I'll figure out upstream/downstream here soon.. previously only used git for my own stuff)

Now for a weekend break. Really! :p
Click to expand...

Ok, so I pushed up these changes. Do a repo sync and you should be good to go.
 
Congratulations on getting the 3G and WiFi working! What you guys do is amazing, and I enjoyed CM7. CM9 is coming along good as I can see. I'm only 17, wish I knew how to program code and stuff like that. Good luck on getting everything else working, as you are better programmers then Motorola and Virgin Mobile. Motorola = Lazy
 
Note that you can do what Whyzor mentioned in another thread to get rid of the btwlancoex cpu hog.

(that is, remove /system/bin/btwlancoex)
 
gannon5197 said:
Really don't know if this is irrelevant, so mods delete this if needed.
Look up "newboston Java" on YouTube. I'm studying up while I can. It teaches you all about Java and he even has a playlist for developing android apps. I'm studying up while I can.
Click to expand...

You're welcome. :-) that site has been posted on my thread for a while now. now if only someone can help me get Ubuntu to properly install on my laptop, I could contribute to the development. Pm me (to prevent clutter in this thread) if you have any ideas. Virtual box...? Maybe a different Dev/Linux program.

Sorry moderators.
 
Ok I got the touchscreen working a little better w the keyboard, browser and hopefully most other situations.

https://github.com/Whyzor/triumph-kernel-msm7x30/commit/c9eb088b16155515fe868e8f05789f8ea54f79b4

Basically in the bi041p_ts.c file the ABS_PRESSURE should be ABS_MT_PRESSURE. One remaining issue is if the 2nd touch x or y axis is less than 1st touch, the two points are swapped sometimes. I will look into this when I get more time.

For some reason after syncing & building, I still don't have 3g. Can anyone else other than Mantera or Brad get a working 3g build from repo sync?
 
Whyzor said:
Ok I got the touchscreen working a little better w the keyboard, browser and hopefully most other situations.

https://github.com/Whyzor/triumph-kernel-msm7x30/commit/c9eb088b16155515fe868e8f05789f8ea54f79b4

Basically in the bi041p_ts.c file the ABS_PRESSURE should be ABS_MT_PRESSURE. One remaining issue is if the 2nd touch x or y axis is less than 1st touch, the two points are swapped sometimes. I will look into this when I get more time.

For some reason after syncing & building, I still don't have 3g. Can anyone else other than Mantera or Brad get a working 3g build from repo sync?
Click to expand...


Yes I just repo synced and built. 3g is working for fine me. WiFi however, is not.

When I turn it on, it definitely thinks longer than it used to but it never fully gets turned on. I get errors like:
E/WifiHW ( 227): Supplicant not running, cannot connect

Did you try a make clean and make clobber before building?

EDIT: After rebooting WiFi is working. So if it doesn't work the first time it boots, you should just have to reboot and it'll work.
 
Whyzor said:
Ok I got the touchscreen working a little better w the keyboard, browser and hopefully most other situations.

https://github.com/Whyzor/triumph-kernel-msm7x30/commit/c9eb088b16155515fe868e8f05789f8ea54f79b4

Basically in the bi041p_ts.c file the ABS_PRESSURE should be ABS_MT_PRESSURE. One remaining issue is if the 2nd touch x or y axis is less than 1st touch, the two points are swapped sometimes. I will look into this when I get more time.

For some reason after syncing & building, I still don't have 3g. Can anyone else other than Mantera or Brad get a working 3g build from repo sync?
Click to expand...

Thanks, I'll sync that up and update the kernel.

And make sure you do a reboot after the first boot to get the rest of the stuff working properly.
 
Just a suggestion (don't mean this is a pushy way) but as you are way better than the rest of us yumbrad (Isaac 2.0) is there anyway you could work on tethering next :D. If you can't/don't want to, that's fine, but it is something I use quite often. Thanks for time in developing and getting most things working on ics :).
 
louis321 said:
I don't know if he is back or willing to help but I just saw on rom miui isaacj87 is back online.....
Click to expand...

If you read what he said I kinda doubt it, but you never know. Also I think yumbrad is more of a Tickerguy than an Isaac (no offense to Isaac)... Just saying. Anyways sorry for talking the thread more off topic.
 
PerkinsII said:
Just a suggestion (don't mean this is a pushy way) but as you are way better than the rest of us yumbrad (Isaac 2.0) is there anyway you could work on tethering next :D. If you can't/don't want to, that's fine, but it is something I use quite often. Thanks for time in developing and getting most things working on ics :).
Click to expand...

I second that! :rolleyes:I use it all the time to connect my tablet and laptop. This is what would hold me back from getting the alpha release
 
I just commited a new version of the touchscreen driver that actually uses my re-write, tested to work with ICS. Prior to this, it was based on Isaac's re-write, which added some bloat to what should be a quick interrupt service. His version & mine are functionally the same, he re-wrote it in preparation for some future improvements that never materialized.

https://github.com/Whyzor/triumph-kernel-msm7x30/commit/5e7882d8a9fa32feeb451193ed805db5691cd353

BTW I highly recommend switching to the SIO scheduler for the kernel as default, so it doesn't block other threads when there's a lot of I/O as in the NOOP scheduler.
 
Whyzor said:
I just commited a new version of the touchscreen driver that actually uses my re-write, tested to work with ICS. Prior to this, it was based on Isaac's re-write, which added some bloat to what should be a quick interrupt service. His version & mine are functionally the same, he re-wrote it in preparation for some future improvements that never materialized.

https://github.com/Whyzor/triumph-kernel-msm7x30/commit/5e7882d8a9fa32feeb451193ed805db5691cd353

BTW I highly recommend switching to the SIO scheduler for the kernel as default, so it doesn't block other threads when there's a lot of I/O as in the NOOP scheduler.
Click to expand...

Awesome. I'll have to add it up after alpha 0.5 is posted. Yeah, your original one didn't work with ICS originally, that's why I didn't have it in there to begin with. But this is great news.
 
Wow guys, I'm impressed! I stopped getting emails from the forums about a month ago when it looked like development was stagnant and the ICS kernel still wasn't booting. I hopped on here planning to start poking at the kernel again since I got a new machine, and the work is much farther than I expected! Great work guys!

I just pulled all of the current source and have an image which booted right up on my phone. Note to any Arch Linux users out there though--I got almost all the way to the end of the build and OpenJDK6 segfaulted. I had to go pull the Oracle jre6/jdk6 from AUR.

I saw someone saying their Wifi still wasn't working after the Wifi fixes got merged. I just wanted to check in that Wifi is working on my phone just fine. I'm only seeing ~7Mbps on the Speedtest.net app (I have a 25Mbps connection), but my cable has been a little wonky today so it might not be the phone causing the slowdown there.

I do kernel/driver development at work (not Android though) so I'd like to chip in and get m y hands dirty with something low priority (I'm also married and working towards a Masters so I don't have a ton of free time). I see the cameras and MMS aren't working, but those are probably too critical for me.. is anyone looking at HDMI yet? I might order a cable and start looking there. I had a logcat going during my first boot of the ROM and saw the HDMI service attempted to start, at least.
 
Also, wanted to chime in that on the current build the swipe detection seems a little off. I have to swipe from literally the very edge of the screen to the other edge of the screen to do things like switching between home screens.

On the other hand, the keyboard acting funky almost seems to me like the touchscreen is being TOO responsive. If I try to type like I normally would, the letter I want pops up followed by 2-3 more letters from around it. If I'm extremely careful and make sure to touch only the letter I want, typing on the keyboard is working fine for me. The toggle between 123/ABC not working seems to be like it's detecting two pushes on the toggle button. If I keep jamming on the thing I can get it stuck in 123 instead of ABC.

I wonderif these issues could be related and the touchscreen driver just needs to have some filtering added to it?

I see what you guys mean with the Browser touch behavior as well. That's.. odd.

EDIT: More touch screen testing results -

I downloaded "Multitouch Visible Test" from the Market. There's two things I noticed playing with it -

1. If you slide all the way to the bottom of the touch screen, the coordinates at times will jump to (0,0), which is the upper left of the screen.
2. If you're doing multi-touch and the touches get within 50 pix of each other on either axis, they will snap to the same value. This may be a limitation of the touchscreen itself.

Other than that, the touchscreen is extremely responsive (much more than stock and the CM7 builds on my phone) and I didn't see any missed touches or other weird behavior in my quick testing
 
aaronfitz said:
Also, wanted to chime in that on the current build the swipe detection seems a little off. I have to swipe from literally the very edge of the screen to the other edge of the screen to do things like switching between home screens.

On the other hand, the keyboard acting funky almost seems to me like the touchscreen is being TOO responsive. If I try to type like I normally would, the letter I want pops up followed by 2-3 more letters from around it. If I'm extremely careful and make sure to touch only the letter I want, typing on the keyboard is working fine for me. The toggle between 123/ABC not working seems to be like it's detecting two pushes on the toggle button. If I keep jamming on the thing I can get it stuck in 123 instead of ABC.

I wonderif these issues could be related and the touchscreen driver just needs to have some filtering added to it?

I see what you guys mean with the Browser touch behavior as well. That's.. odd.
Click to expand...

Assuming you built from Mantera's repo, then the problems you're seeing have already been fixed. Whyzor fixed a bug in the touch screen driver in the kernel which was causing most of the problems, but the new kernel hasn't been merged into Mentera's repo yet.

You should go download the alpha 0.5 zip from the other thread and transplant the boot.img into your build. That's what I did and all the touch screen issues were resolved.
 
NateMckn said:
Assuming you built from Mantera's repo, then the problems you're seeing have already been fixed. Whyzor fixed a bug in the touch screen driver in the kernel which was causing most of the problems, but the new kernel hasn't been merged into Mentera's repo yet.

You should go download the alpha 0.5 zip from the other thread and transplant the boot.img into your build. That's what I did and all the touch screen issues were resolved.
Click to expand...
Ah, gotcha. I'll go give that a go. The first post in this thread said "Now using Whyzor's latest touchscreen driver", which made me think his fixes were merged.

EDIT: I gave that a go and it's still not happy. I see mantera's repo didn't get updated with Whyzor's new touchscreen driver.. I'm going to pull it over and build my own kernel to see if it plays better.
 
You must log in or register to reply here.
Back
Top Bottom