make test in master fail - json
a.cudbardb at freeradius.org
Tue Sep 15 20:34:49 CEST 2015
> On 14 Sep 2015, at 23:59, Matthew Newton <mcn4 at LEICESTER.AC.UK> wrote:
> On Mon, Sep 14, 2015 at 11:47:25PM +0100, Arran Cudbard-Bell wrote:
>>> On 14 Sep 2015, at 22:58, Matthew Newton <mcn4 at LEICESTER.AC.UK> wrote:
>>> Debian wheezy, make test in master (19d03eeb44c66e2) fails:
>>> There are comments there something about json-c < 0.10. This is json-c 0.10
>>> (debian 0.10-1.2), so I guess it could be that 0.10 is broken as well? Or just
>>> something in flux in master.
>> Weird, it doesn't break in travis with json-c 0.9, it doesn't break on OSX
>> with json-c 0.12. I guess there was some weird transitional thing in json-c
>> 0.10 *sigh*.
> OK, additional reference. Jessie with libjson0 0.11-4 works fine.
> Though can't guarantee the build environments are similar in any
> way (just my two desktop systems, not clean Debian installs).
>> The comments about json-c < 0.10 are because 0.10 was the first get support
>> for 64 bit integers (supposedly).
> Righto. Looks suspiciously like "supposedly" :)
Weird weird things.
Can you test on your system as see if all JSON integer values produce INT32_MAX/INT32_MIN/INT64_MAX/INT64_MIN?
This does not happen with json-c 0.9.
As they seem to on mine, which is why the tests are failing.
I even modified the code to avoid casts from 32->64 bits and it's still producing junk values.
Arran Cudbard-Bell <a.cudbardb at freeradius.org>
FreeRADIUS development team
FD31 3077 42EC 7FCD 32FE 5EE2 56CF 27F9 30A8 CAA2
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 872 bytes
Desc: Message signed with OpenPGP using GPGMail
More information about the Freeradius-Devel