Page 12 of 42 FirstFirst ... 2101112131422 ... LastLast
Results 111 to 120 of 411

Thread: DNA Development Blog

  1. #111
    Insider Curt's Avatar
    Join Date
    Jun 2013
    Location
    Atlanta, GA
    Posts
    295
    Supports Inception Designs
    Supports Inception Designs
    Quote Originally Posted by jptheripper View Post
    Anything I can do/donate to the project?
    I wish. Right now I am trying like hell to get all the bugs worked out of the software. I fixed the "doesnt work on some USB buses" at least basically, but there still seems to be a flutter..

    and now my bootloader-loader doesn't fully work. It loads the new code, but then refuses to REBOOT. *grr*

    If you are really good with software and genuinely want to help I suppose I could send you a test rig, but short of that, no.

    Maybe babysit my kids? give me an extra 4 hours in the day? I am running around like a head with my chicken cut off.

  2. #112
    im in SW FL, if you can get the kids here, i'll watch them. Have two of my own (6/5).
    I code all day but its javascript/dojo/jquery and its hacky at best. Need an angel to test with?

  3. #113
    Insider Curt's Avatar
    Join Date
    Jun 2013
    Location
    Atlanta, GA
    Posts
    295
    Supports Inception Designs
    Supports Inception Designs
    hehe no thank you

  4. #114
    Insider Curt's Avatar
    Join Date
    Jun 2013
    Location
    Atlanta, GA
    Posts
    295
    Supports Inception Designs
    Supports Inception Designs
    Only took 10 hours of debugging.

    Code:
    0x08C0, // 08  c0  rjmp +16
    should read

    Code:
    0xC008, // c0 08  rjmp +16
    The insidiousness of this bug is it causes random havok inside the AVR which sorts itself out jumping around all the 'undefined operation"''s contained in an unprogrammed chip, very quickly, then accidentally stumbles into the bootloader code and initializes correctly. This happens so fast that it appears to work.

    UNTIL you install a program in the way, that's why it worked until reboot.

    this was a fun one, I hand-assembled several lines and that is the only one I messed up the high/low bytes.

    I'm getting some sleep now, only halfway there.

  5. #115
    i remember in the early 80's typing out things in basic and not understanding how it works. its insidious that a slight mistype can lead so much troubles.

  6. #116
    those of us that are anxiously awaiting this product appreciate your diligence. Enjoy your well deserved sleep.

  7. #117
    Try coding for IE with JSON. One rouge comma and every browser works except IE explodes. Wont even load.

  8. #118
    Insider Curt's Avatar
    Join Date
    Jun 2013
    Location
    Atlanta, GA
    Posts
    295
    Supports Inception Designs
    Supports Inception Designs
    Ever have one of those days where you realize how stupid you've been? Had one.

    breakthrough on bootloader code size, why the hell was I loading a whole page at a time? that requires logic on the receiving side to parse the data stream. This is USB. it's fast. So now I load the image 4 bytes at a time and fit the entire transaction in a single packet. no stream logic necessary. tossed it.

    Take longer? yeah, instead of <1 second, takes more like 3. pff.

    Then I realized instead of check-summing each packet, it makes more sense to crc the entire image in place, once loaded. that's where it needs to be correct anyway. single for loop. code got smaller.

    And I was doing trivial pointer math on the bootloader side, x2 kinda stuff. and computing offsets. waitaminute I can compute that on the PC side and just feed it "put THIS HERE". code got smaller STILL.

    In the end I compressed the bootloader 128 bytes while troubleshooting the USB on different platforms, it now consumes just shy of 1.8k, or ~20% of program space. Still a bit higher than I'd prefer, but down 5% is nothing to sniff at. For reference the entire feature-complete Morlock image w/RNA/USB/menus is 3.6k, about half of the available space.

    Which has never been the point, I have been trying feverishly to get the platform stable, which I am ready to say it is now. Time to work in earnest on the Morlock code itself. This should not take too long, but the good news is after a pass or two I can ship hardware. Since the platform is stable I am comfortable field-upgrading issues, which IS what this has been about.

  9. #119
    great news curt

  10. #120
    absolutely awesome news. does that include the oled ? or will the first batch be led style

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •