Jump to content

tdm

Members
  • Content Count

    801
  • Joined

  • Last visited

  • Days Won

    84

Posts posted by tdm

  1. 1 hour ago, acrux said:

    acrux=rehejuhan 😄

    If the aim is to get the keyboard act as printed on the keys (that is my strong preference) then it is a step towards it.

    However - these were the only keys I was able to adjust. About the others I wrote in that issue...

     

    Thank you for noting your github name.  🙂

     

    The aim is to get the keyboard working in the most obvious way possible.  In other words, start by:

     

    1. Select QWERTZ under settings -> system -> keyboard -> layout.

    2. Select German (or similar) under settings -> system -> languages/input -> physical keyboard -> builtin keyboard.

     

    With that done, the keys should generate the symbols as marked, eg:

    1. Pressing a key with no shift should generate the white marking.

    2. Pressing a key with yellow-arrow should generate the yellow marking.

     

    Please note that Android handles the normal shift key.  The kernel does not have any special logic for normal shift.  So I can only hope that it works properly.

     

    I have made the changes you noted on github and created a new boot image.  You can find it here:

     

    http://files.nwwn.com/android/pro1/boot-qwertz-test1.img

     

    This should work with test11 or test12 equally well.  You should be able to flash it with fastboot, eg. "fastboot flash boot boot-qwertz-test1.img".

     

    Please let me know if the keys you noted work properly, and what additional changes should be made.  I will incorporate the changes and we can iterate until it is correct.  🙂

     

     

    Thanks!

     

    • Thanks 4
  2. 26 minutes ago, acrux said:

    Please, if possible, give some "love" to german qwertz layout 🙂

    https://github.com/tdm/android_device_fxtec_pro1/issues/35

    Absolutely.  Reading the issue report, there is a modified keymap in comment by rehejuhan on March 10:

     

    11:0033:8033 (, and ; )
    26:002c:0026 (y and |) <- update, not correct, 0026 is just l, not |
    30:0011:0029 (w and ^)
    38:0013:8029 (r and °)
    39:0028:002b (ä and #)
    42:0056:8056 (< and >)
    47:0026:802b (l and ')

    Shall I use this?

     

    • Thanks 2
  3. Following up on @VaZso and his friend's device... we looked at a USB packet capture and determined that everything was fine until the programmer tool sent a request to write to UFS, and the device never responded.  I have never seen this issue before on any device.  My theory is that UFS is bad and the device needs to be returned for service.

     

    • Thanks 2
  4. 12 hours ago, Jordi said:

    I can reproduce this quite easily at home, in a room where I have poor wifi signal. Unfortunately, my router is not configurable directly, I have to go to the internet site of my provider and there is just a few parameters I can modify.

    That is great (well, great for reproducing the issue).  Can you connect the device to a PC and run "adb shell cat /dev/kmsg" and then cause the issue to happen?  Hopefully the device will show a kernel panic message before it resets.

     

    • Like 3
    • Thanks 1
  5. On 5/15/2020 at 1:51 PM, michael.bosscha said:

    Alright, so I just had my first random reboot. For me, this has never been an issue until today, when I connected to GF's new wireless router. I remarked that I found the signal strength a bit lacking for a device that's not even 10 meters away, but she said she had good connection since starting it up. OK, let's try it out! I connected, entered an URL in an open window and... bam, random reboot. Whut. OK, reboot, reopen Chrome, browse to the FX forum, search for this thr... bam, another reboot. OK, so definitely sure now this is due to the wifi acting up. Reboot, switch off wifi and back to cellular... and smooth sailing again. That was weird... does my phone keep track of system crashes somewhere in a log like Windows does?

    Reboots like this are not tracked anywhere for production devices.  I'm pretty sure you are seeing a bug in the qcacld kernel driver (eg. the Qualcomm wlan module).  Is there anything particular about this wireless router that might cause the issue?  I would like to see if anyone else can reproduce this.

     

  6. On 5/10/2020 at 11:00 AM, VaZso said:

    I agree - it would be good if they would integrate LineageOS fixes into stock firmware or hire somebody who is competent in Android development like @tdm to do this valuable work...

    I have access to the stock code and I could do some builds with bug fixes. But I don't see the point. The biggest (only IMO) reason to run stock is for the locked boot loader and play store certification (aka safety net). IdeaLTE does not share the keys to sign builds, so my builds cannot run locked. And, if you really want safety net to pass in Lineage, there are ways to make that happen.

     

     

    • Like 2
    • Thanks 1
  7. Hey all, I know it's been a while but .. well .. pandemic and all.  I have been doing my "day job" remotely from home with 4 boys running around doing some sort of "school" for a couple hours of the day and being .. well .. being boys the rest of the day.  So it's been a bit of a challenge to manage everything and still run side projects.  But I think I've got most everything under control now and I'm going to try to drive this thing to the finish over the next couple of weeks.

     

    So, as a (re)starting point, I've synced up my Lineage tree and rebuilt from where I left off to make test12.  In other words, test12 is the latest Lineage 16.0 tree with the latest security and bug fixes, but has the same pro1 code as test11.

     

    I'll go over the issue list on github and prioritize later today or tomorrow.  If I recall correctly, the biggest issue remaining was SELinux.  So unless something bigger has been posted, that should be my next issue to tackle.

     

    • Like 3
    • Thanks 10
  8. Hey all, it's been a while since I've been here so I wanted to pop in and give an update.

     

    I live in Seattle, which has been under stay-at-home orders for over two weeks and will continue until at least the end of this month.  I have a wife and 4 boys at home, and we are all healthy.  But it is difficult to get any work done between the constant interruptions.  In addition, I have been given a new project at work that will take a week or two to complete.  So I'm afraid that I won't have much of any time to work on the Pro1 until at least the middle of April, but more likely the start of May.  But rest assured I will continue to work on it when things get closer to normal again.

     

    Stay safe and healthy out there!

     

    • Like 5
    • Thanks 13
  9. 1 hour ago, david said:

    @tdm, I got your tool to work.  I was able to simulate flashing the phone with the No Write option.  I think I have an idea what makes it work and not work, but I will play with it a bit more to be sure and then report back.

    In the meantime, which firehose is being used in the current version of the tool?  Tulip, Crackling, or Bacon?

    Glad you got it to work.

     

    Where did you get the idea that the firehose for any of those devices was used? None are, of course. The pro1 firehose is used.

     

  10. 1 hour ago, david said:

    Thanks for the answers.  I wasn't questioning your decisions.  I was just trying to understand how things were communicating one way or another.  

    I didn't think it was when it was done flashing.  It showed it immediately after I clicked the flash button.  How long should it take to copy the file to the phone when the no write option is enabled?

    Several minutes, same as actual writing.

     

    1 hour ago, david said:

    Is there some way to enable logging so I can give you a log file?

    Yes, I can give you a debug version. But a packet capture with eg. WinPCAP or wireshark would probably be just as useful.

    1 hour ago, david said:

    Also, when I am in EDL mode, my LED light does not flash.  I think I have read that others have had their light flash, but I don't know if that was for the Pro1 or other devices.

    Nope, the edl code knows nothing about led or screen or anything else.

    1 hour ago, david said:

    It seems like the serial over USB concept is a not very good concept.  Lol.  Using an online calculator, to make sure I wasn't making any math errors, here is how long transferring the file would take over the serial api:

    That's crazy. 🙂

    And some of the other EDL tool developers have said that you might need to set the speed to 16384 speed or less, so almost 10 times slower!

    So thanks for not using that method.

    That's the thing, USB does not have a programmatically settable speed. So you can set the virtual serial speed to whatever you like and the actual transfer speed will be the same.

  11. 3 hours ago, VaZso said:

    Also, I have tried to put my own phone to EDL mode using a home-made "deep flash" cable and it seemed to be not possible even though my phone has charging enabled when the phone is switched off so that icon has appeared and could not go to EDL that way...

    If anyone have any ideas, please share it with me.

     

    If xbl is functional this should always work.  And if you can get to the point where you see a boot logo when turning on the device, xbl (and much more!) is functional.

     

    Try putting the cable in the other way (eg. upside down).  @Sean McCreary was unable to get his device into EDL mode using a "deep flash" cable, so he disassembled it to fix it.  A couple months later, he found that if he flipped the cable upside down, it worked.

     

  12. @david you had several posts with several questions so I'll answer them here.

     

    Why did I choose to use WinUSB instead of the Qualcomm driver?  There are a couple of reasons:

    • The Qualcomm driver exposes the device using a serial port.  This is ancient and outdated technology that is difficult to use.  At minimum you need to figure out the virtual "com" port number assigned to the device by the Windows device manager.  This is unnecessary and confusing.
    • The Qualcomm driver is not installed automatically for all versions of Windows.  Finding it and installing it manually is a huge pain.  Again, unnecessary and confusing.

     

    Why are the WinUSB driver and the Qualcomm driver incompatible?  Because programs that are written for the Qualcomm driver use Windows' serial port API and WinUSB does not provide a virtual serial port.  Conversely, my program uses the bulk USB protocol and the Qualcomm driver does not provide access to that.

     

    Could I modify my program to be compatible with the Qualcomm driver?  Sure.  But why move backwards just because Qualcomm is stuck in the 1980s?

     

    Why does the flash tool show "Unknown error" when it is done flashing?  I don't know but if the flashing process completes, it should not affect the operation.  If you are concerned about it, I can work with you to try to figure it out.  Nobody else has reported this issue, except apparently @VaZso who has replied that he also sees it.

     

    WinUSB has no inherent serial speed settings or limitations, nor framing (bits and parity).  These are serial port concepts, not USB concepts.  USB speed is negotiated below the programming API and, as far as I know, cannot be changed programatically.

     

  13. 48 minutes ago, silversolver said:

    The Nextel iDEN used a custom TDMA system (time division multiple access) and worked well for what it was. However, from Wikipedia, "Sprint Nextel provided iDEN service across the United States until its iDEN network was decommissioned for additional LTE network capacity on 30 June 2013." https://en.wikipedia.org/wiki/IDEN

    I guess all those cool NexTel handsets are museum pieces now.....I had wondered, but not been curious enough to look LOL.

    Did Sprint ever use TDMA?  I thought AT&T was the only one that used TDMA (before they moved to GSM)?

     

  14. 3 minutes ago, Craig said:

     

     

    No, Verizon users cannot get CDMA data, because Verizon won't activate it, but thats nothing to do with the device.  The only way people are using Verizon on Pro¹ now is LTE only after having it activated on another device.  So Verizon doesnt even know the Pro¹ MEID as verizon is tieing cdma service to the meid of the "mule" device.

    I assume the answer to all three questions is yes, but if you're right that they didn't include CDMA voice bands, then Fxtec has been lieing to us a very long time, and I find that unlikely.  But now I'm going to look into it further.   Why would they include only cdma data bands but not voice? that makes no sense.

     

    I'm not going to say anyone is lying.  Certainly it appears that some CDMA data capabilities exist.  But it is not fundamentally a CDMA device.  I would say that it is a GSM device that is compatible with modern CDMA networks that implement data-only voice services.

     

  15. 4 minutes ago, Craig said:

    Have you looked at the specs for this phone?  We have all the CDMA radios and bands.  Verizon not activating us doesn't change the fundamental fact that this phone is indeed a CDMA device, and can operate without LTE or GSM.

    From site homepage:

    image.png.b03d466c266352d1e34d14ff3013a6ce.png

     

    That is all data.  And I presume Verizon users can get data.

     

    Can you make a CDMA voice call?  Can you show me your CDMA MEID?  Can you operate the device on the network without a SIM card?

     

    The answer is no to all three.  Any actual CDMA device would answer yes to all three.

     

  16. 2 minutes ago, Craig said:

    Just wanna take issue with that.  This phone supports GSM & CDMA.   I'm still hoping to activate 2nd SIM on Sprint MVNO...

    Just cuz Verizon hasn't approved Pro¹ (yet?) for use on their CDMA network, doesn't mean the phone doesn't support CDMA.

    I do understand that people are using Verizon on stock devices with just LTE (without GSM or CDMA) with a SIM card first activated on another device, but for that workaround to be effective, need voice over lte working. 

    Sorry, no.  It is not a CDMA device.  GSM and CDMA are fundamentally and technically different protocols.  Just because Verizon (and I assume Sprint) has seen the future and is ditching traditional CDMA for VoLTE and they can support random GSM devices does not magically make them CDMA devices.

     

    On the other hand, if you can take your SIM card out of the Pro1 and still make a voice call on Verizon, like you can with any real CDMA device, then I will retract my statement.

     

  17. On 3/16/2020 at 7:47 AM, NukaNegan said:

    Can we get a website similar to what Google gives for its hardware binaries and firmware images, like so?

    https://developers.google.com/android/drivers

    It would make it way easier to support the community and the development of alternate ROMs, or to roll your own without having to:

     

    This isn't a Pixel device and FxTec is no Google.  You'll be lucky to just get support equal to other manufacturers.

     

    On 3/16/2020 at 7:47 AM, NukaNegan said:

    Is there a walkthrough for pulling binaries off a device?

     

    Read the Lineage wiki.  I'm sure this is covered in detail.  Hint: extract-files.sh.

     

    On 3/16/2020 at 7:47 AM, NukaNegan said:

    What is the procedure to factory restore a device if you brick it when making a ROM?

     

    If you need to do a full factory restore (you won't if you are careful), just grab the latest fastboot package and flash it, then "fastboot -w" to erase userdata.

     

    • Thanks 5
  18. 53 minutes ago, Hook said:

    Well, I went ahead and tried. I'm now thinking this was reported before, but can't remember if it was for Verizon on LOS.  I can confirm that Verizon will nor work on LOS. LOS fails to recognize that the SIM is provisioned for VoLTE and all the switched that provision it for the phone on the phone menu (*#*#4636#*#*) sta grayed out no matter whatother settings you play around with.  I sure hope there is a solution down the road.

    Guess it's back to stock for me. 😉

    I do not have Verizon, so I cannot really fix it.  It would be great if we had a developer that did use Verizon.  But if none appear within the next few weeks, I'll probably need to have a Verizon SIM card sent to me.  However, as this is a GSM phone and Verizon is a CDMA carrier, I see this as a sort of "extra" functionality -- an enhancement request of sorts.  It will take a back seat to getting all the other bugs fixed and official builds running.

     

    • Thanks 1
    • Confused 1
  19. 3 hours ago, dreamflasher said:

    After how smoothly the whole production process and communication was, really surprising that the software update process is so rough. Fxtec -.-

    Production and software updates are the fault of IdeaLTE. Lack of communication is clearly the fault of FxTec.

     

    • Like 5
×
×
  • Create New...

Important Information

Terms