• Replies 4 replies
  • Subscribers 18 subscribers
  • Views 675 views
  • Users 0 members are here
Options
Related

Simulator crashes when viewing memory

dbrobert
Environment:
- MacOS 10.12.3
- Eclipse Neon.2, plug-in 2.2.4
- SDK 2.2.4

Detailed description of the issue:

When attempting to view memory for the attached project, the simulator crashes. This is an inconsistent bug I have seen in several of my projects but have been unable to isolate as to the cause.


Steps to reproduce the issue:

1. Open attached project and run.
2. View memory using the menu option.


Additional information:

I was attempting to see if I could reduce memory usage in another project by changing my use of a string which I need to manipulate to a character array. So I modified my simple TestWatch project to have two variables so I could compare memory usage. When I tried to view the memory the simulator crashed. Since I knew the memory viewing in the project worked before I knew it could only be the addition of one of the two variables. As it turns out, I can eliminate the crash by commenting out the array of characters. However, I don't know if this is the true reason for the crash nor do I know if that is the issue with my other projects that crash when viewing memory. However, hopefully this will bring us closer to a solution.

I tried all the usual tricks, like restarting Eclipse and removing the Garmin directory from the temp dir, etc but nothing prevented the crash except commenting out the array of characters.


Code sample to reproduce the issue:
  • Former Member
    Former Member
    Hello,

    I was able to reproduce this and figure out a little more. It looks like the sim is crashing with arrays that have typ long, double, or char that are defined at the global level. I have created a ticket to investigate.

    Thanks,
    -Coleman
  • Hello,

    I was able to reproduce this and figure out a little more. It looks like the sim is crashing with arrays that have typ long, double, or char that are defined at the global level. I have created a ticket to investigate.

    Thanks,
    -Coleman


    This is great news! I looked at my other project that crashes (METAR/TAF) and confirmed that the view memory functionality is fine until I assign an array of Unix timestamps to a global array, so that is likely my longtime issue that I've previously complained about. ;)

    Thanks for looking further into this.

    Cheers,
    Douglas
  • We've gotten this fixed, and it should be available in the next SDK release.
  • Since this is indeed fixed (as of SDK 2.2.5), might I suggest that the thread be closed? And perhaps, to increase the usefulness of this sub-forum, before a thread is closed that the last message be the SDK release in which the bug was fixed and the title be changed to be prefixed with [FIXED]?

    Cheers,
    Douglas