HYSTM32_32 LCD stays white for firmwares > 1v48

Posted on
  • Dear Gordon,

    thank you for this great project!

    I am (not yet) an expert like you, therefore I would appreciate help in finding out what happened to the LCD handling in the nightly firmwares for the HYSTM32_32 board?
    Since version 1v48 the LCD stays white and shows no reaction to LCD draw or clear commands.

    Thanks in advance


  • Honestly I'm not sure. I did have a look at it and it seemed as if the LCD initialisation code hung - but I haven't changed anything there in a long while.

    I wonder if someone who had a built environment set up could 'git bisect' and find out where everything went wrong?

  • I have this board and for me anything after 1v49 doesn't work at all. And I mean not only the display but also the board is not responding at all.

    After uploading espruino to the board it doesn't respond on the serial console, I can't write and I don't get anything, just a blank screen.

    Gordon HELP :)

  • So the last working build was 2014-02-03 (1v49). Next build's binary size was reduced and it looks that something got broken.

    Could someone check with his board, please?

  • I've got my work cut out at the moment so I don't have time to figure out what caused this...

    As above, if someone could get the compiler set up, and could git bisect to work out which actual change in Git caused the problem it would be easy to rectify it.

  • I've tracked it down to this exact commit: https://github.com/espruino/Espruino/com­mit/9da681c23cda4be6cc774390c1c084866d7f­1f2f

    ... which has absolutely no relevance to LCDs at all.

    it works before it, and fails with/after it. However if you go on to the next commit, and remove those 2 lines, it still fails. In fact even if you keep the 'jsvIsString' but light an LED instead of doing anything useful it still fails.

    In fact, if you attach a hardware debugger and breakpoint that function, it's never actually reached.

    I have absolutely no idea what's wrong. It almost seems like some code size/compiler issue.

  • That's a bad news I guess... but still thank you for investigating!

  • +1 for fixing this before my boards arrive.

  • Hi Folks,
    I've seen the same problem with 1v60 ... "white screen of death".
    Decided to revert to 1v44. It's a pity, because the HYSTM32_32 is a great board and I really like Espruino ...... a wonderful product.



  • Thanks! - I know it is frustrating... I've spent probably a day on it now, and haven't really got anywhere. I can't really justify spending more time on it.

    As I keep saying though - all the source is there, I've tracked down the exact commit that causes the problem - it looks like a compilation issue. It just needs someone to look around it and try and track down exactly what the problem is. It's quite likely some issue with the linker script and whether it's putting everything in the right place.

  • Hi Gordon,
    OK, thanks for the fast reply, I know you're very busy and you don't have much time to spare. The 1v44 is OK for what I need at the present time. However, another problem I've noticed, this time with the WebIDE, is that it hangs after trying to "send to Espruino" a reasonable chunk of code (about 50 lines). Strange. The simple "toggle a led" uploads and works OK.



  • Sounds similiar to a problem which came up with Chrome M33.
    Sometimers I got totally strange error messages, sometimes it hang up.
    Did you check throttle in settings -> communications ?
    This helped me.

  • As @JumJum said - try that option out.

    The IDE should detect that it's not talking to a board with the newer firmware and should throttle anyway, but it may be there's a bug there.

  • Hey,

    i've got some time on my hands again. My HY 3.2 boards arrived, and I wanted to see if there's any progress on this problem and what the current thread for this board might be.

    I would look at setting up a tool chain and trying to fix the problem, if there's nobody else working on it so far. Any hints/links/recipes for setting up the tool chain would be appreciated, though.


  • There's actually some info here: https://github.com/espruino/Espruino/iss­ues/327

    Marcel has been working on it and says he got it working by getting rid of some warnings... Hopefully he'll be able to contribute a patch soon, but I'm not 100% sure that any of the warnings he fixed could have caused the instability, so it's more likely to be something related to his compiler setup.

  • Just to add, there should be some help on compilation in the readme: https://github.com/espruino/Espruino

    There's also a thread on it here: http://forum.espruino.com/conversations/­151/

  • Hey Gordon,

    Thanks. Always impressed how you handle forum and coding.

    I remember that you mentioned using some special compiler to create smaller binaries. Any info on that one, and is it crucial for anything related to the hy3.2?


  • I don't think it's that crucial. The 3.2 is pretty good (it's the 2.4s that were a bit under-specced IIRC). To be honest if you compile with -Os then the newer the compiler the better - it's just when compiling with -O3 that codesourcery-2013.05-23-arm-none-eabi seemed to do a better job.

  • Post a reply
    • Bold
    • Italics
    • Link
    • Image
    • List
    • Quote
    • code
    • Preview

HYSTM32_32 LCD stays white for firmwares > 1v48

Posted by Avatar for amp @amp