Re: [Qemu-devel] qobject/qjson.c:69: failed assertion `obj != NULL'

[G 3 <programmingkidx@gmail.com>]

On Nov 21, 2016, at 3:36 PM, Eric Blake wrote:

> On 11/21/2016 02:12 PM, G 3 wrote:
>>
>> On Nov 21, 2016, at 5:02 AM, Paolo Bonzini wrote:
>>
>>>
>>>
>>> On 21/11/2016 07:30, G 3 wrote:
>>>> When I try to use qemu-system-i386, I see this error message:
>>>>
>>>> qobject/qjson.c:69: failed assertion `obj != NULL'.
>
>>>>
>>>> The string argument is this: { 'seconds': %qd, 'microseconds': % 
>>>> qd }
>>>>
>>>> Would anyone know how to fix this problem?
>>>
>>> Please include the full backtrace.
>
>> #3  0x003ec69c in qobject_from_jsonf (string=0x443f44  
>> "{ 'seconds': %qd,
>> 'microseconds': %qd }") at qobject/qjson.c:69
>> #4  0x0033f6e4 in timestamp_put (qdict=0x2052e00) at
>> /Users/john/desktop/qemu-2.0.0/monitor.c:476
>
> Why on earth are you trying to compile qemu 2.0.0?  The latest version
> we support on this list is currently 2.6.x for patch backports, with
> 2.7.0 as the latest stable release; and very soon now it will be 2.7.x
> for backports and 2.8.0 as the latest stable release.
>
> The source of your problem is that your platform defines PRId64 as  
> 'qd',
> but the qemu JSON parser only recognizes lld (POSIX) or I64d  
> (mingw) for
> parsing 64-bit numbers.  We could enhance the JSON parser to recognize
> the non-standard qd in addition to the hack we already have for mingw,
> but I'd argue that using qobject_from_jsonf() is already less-than- 
> useful.
>
> On the other hand, timestamp_put() (which now lives in qapi/qmp- 
> event.c,
> rather than monitor.c) is still using qobject_from_jsonf() with PRId64
> in current git master.  If you don't want to hack the JSON parser  
> (and I
> recommend that you don't), the alternative is to get rid of that
> function call and replace it with an open-coded construction of a  
> QDict
> with manual population of its two members.  It's hard to argue that  
> the
> complexity of maintaining our pseudo-printf JSON parser for  
> constructing
> a QObject with one line is worth the effort compared to the three or
> four lines to construct the same QObject by hand.

I would just like to note for anyone else who might face this problem  
in the
future that replacing %PRId64 with %lld fixed the problem. I made the
replacement in timestamp_put().