Your attached form seems to be missing.
Also filing an issue on our issue list is helpful as we track progress
on problems there.
On Sat, Jul 28, 2012 at 5:23 PM, wrote:
> A form I created with build is causing Collect to crash in a (somewhat) predictable way.
> The form itself is quite simple and fairly short (xml attached). It is basically a series of grouped Date Spinners and MultiSelects (labelled x_DK with 1 option, which is what I generally use for Don't Know/No Response). The form passes validation.
> When I swipe through the form without selecting anything, there are no problems. If I swipe around with a couple of the MultiSelects selected, no problem. However, we I get to >11 of the MultiSelects selected, the next swipe causes a crash. The ADB output at time of crash is 'java.util.Date cannot be cast to java.lang.String.'
> I have tried updating and rolling back to different versions of ODK, with no success. And while I am able to consistently replicate the error, it does not seem to be limited to 1 field - that is, if I hit MultiSelects 1-5 and 7-12, it then only crashes when I hit MultiSelect 6. Alternatively, I can do 1-9, then 11-13, then come back and crash it on 10.
> Any help or guidance on this would be very much appreciated.
> ADB Output:
> W/dalvikvm( 3978): threadid=1: thread exiting with uncaught exception (group=0x40a381f8)
> E/AndroidRuntime( 3978): FATAL EXCEPTION: main
> E/AndroidRuntime( 3978): java.lang.ClassCastException: java.util.Date cannot be cast to java.lang.String
I believe this is an error with a formula in your form.
The /data/IM32 prompt has this relevance condition:
selected(/data/IM6_group/IM6_ND, '99') and
selected(/data/IM7_group/IM7_ND, '99') and
selected(/data/IM8_group/IM8_ND, '99') and
selected(/data/IM12_group/IM12, '99') and
selected(/data/IM13_group/IM13, '99') and
But IM12, IM13, IM14 are date values. I think these should be IM12_ND,
On Mon, Jul 30, 2012 at 8:12 AM, wrote:
Ugg, sorry. Attached here.
I will certainly file this as an issue on the list as soon as someone
confirms it's not some ridiculous human error issue that I've overlooked.
University of Washington