Menu

#3 App refuses to load

open
nobody
None
5
2012-12-02
2012-12-02
gh403
No

The app I'm developing won't load onto an emulated calculator. It will load in other emulators such as TI's Flash Debugger and Wabbitemu.

I have attached the failure error.

I have no idea what could be causing the problem; although the emulator is definitely at fault, since 2 other emulators plus a physical calculator does it correctly. I'd be happy to provide more information if requested, although right now I don't know what to provide!

Discussion

  • gh403

    gh403 - 2012-12-02

    Error when trying to load app

     
  • Benjamin Moody

    Benjamin Moody - 2012-12-02

    Could you please post (or email privately) a copy of the .8xk file that is not working?

    Also, you say it works on a physical calculator; what program did you use to install it (TiLP, TI-Connect, TIGL)?

     
  • gh403

    gh403 - 2012-12-02

    I'm emailing you a "traumatized" version of my app. I've gotten rid of almost all the code (I don't want to publish it yet), but left the basic structure, so if you do get it to load, there is no guarantee that it will do anything other than crash. (Getting it to load would be a success in my book!)

    If I remember correctly, I used TiLP to load it to a physical calculator... although it's been a while since I've had access to actual hardware, so my memory could be wrong there.

     
  • gh403

    gh403 - 2012-12-02

    Ok, I can't figure out how to attach something to a Sourceforge email. I attached it to the bug report instead.

     
  • gh403

    gh403 - 2012-12-02

    Example app which refuses to load

     
  • Benjamin Moody

    Benjamin Moody - 2012-12-02

    Thanks. The problem is that that file has the wrong page numbers (or it has the pages in the wrong order - I can't tell what your intention was.) It has one page marked as "page 0", then one marked as "page 2", then one marked as "page 1" with a signature at the end. If the linking program honors the stated page numbers (like TilEm does, and TiLP does when using a traditional link cable), the signature will end up in the middle of the app instead of at the end, and the OS will reject the app. If the linking program ignores the page numbers and sends pages according to their position in the file (which, I believe, TiLP does when sending an app via direct USB), it will be accepted, but the pages won't be in the order you (presumably) intended.

    You need to ensure that the app pages are in the correct order when you create the app. I don't know what tools you're using, but I have just noticed that rabbitsign will allow you to create an app with pages out of order. It never occurred to me to check for that situation.

     
  • gh403

    gh403 - 2012-12-02

    I see. I'm using a somewhat strange toolchain; I'm running Linux (as you probably gathered), tracking the source with git, assembling and linking with ZDS, and signing with native rabbitsign.

    I've looked through the linker options, and I can't tell that there is an option to specify what order to write the pages to the file to. There is a segments dialog, but it appears to be in order (http://i.imgur.com/oYEzq.png). I guess I'm trying to say, is this something that I need to fix?

     
  • Benjamin Moody

    Benjamin Moody - 2012-12-03

    Well, I would say that it is something that needs to be fixed at either the ZDS or Rabbitsign level, because even if the app as-is works for you, it'll be broken for about half your users. ;)

    As to ZDS, I don't use it, but I'm thinking that perhaps you want to put the sections in the opposite order (PAGE0->PAGE1->PAGE2.) But I don't understand why your current settings would result in page2 being placed in the middle. Could you send me copies of your .zws file and the unsigned .hex file? (BTW, you can email me at floppusmaximus at users.sf.net if you prefer.)

    I'm thinking I will change Rabbitsign so it will detect this situation, either refusing to sign the app or rearranging the pages into the specified order.

     

Log in to post a comment.