possible problems with empty map in map --- how to submit unit test about qcbor HOT 8 CLOSED

Hi Michael,

I'll look at this in the next day or two. It is likely a bug as I don't recall testing this and there are a lot of corner cases related to empty maps and arrays.

I don't have any file I/O in all of QCBOR for a reason. Some embedded systems don't have files. I'll debug this by making a test case and setting a break point at it (in Xcode).

Thanks for reporting.

from qcbor.

There is no makefile in the tests directory, and no "make check" in the main makefile.
It would be great if it had that, then we could run the whole thing under travis-ci.

from qcbor.

I changed my code to just put a 1 into the target, rather than an empty map: 0xa1 0x01 0x01
SAME THING.

from qcbor.

#81 shows this problem against the top of your tree.
I'm perplexed. I am decoding multiple items without a problem.
I have also tried adding a second entry to the map.

from qcbor.

Well, I screwed up my unit test case. Standby... :-)

from qcbor.

okay, fixed: I compared to TYPE_ARRAY, and that was wrong.
But, a printf():

   printf("type: %d nest: %d labeltype: %d label: %d\n",
          Item.uDataType,
          Item.uNestingLevel,
          Item.uLabelType,
          Item.label.int64);
```

shows me:
type: 5 nest: 0 labeltype: 0 label: 0

the labeltype being 0 is weird.

from qcbor.

My test case was inaccurate: I simplified too much. But my underlying issue remains in my code.

from qcbor.

Further: -DCONFIG_ variables not set consistently across users of QCBOR includes.

from qcbor.