Experimental Physics and
Industrial Control System

"Johnson, Andrew N. via Core-talk" <core-talk at aps.anl.gov> · Wed, 7 Feb 2024 21:56:51 +0000

If you follow the links in the PR that Michael posted you’ll see that this fix was to support NaN and Inf values which strict JSON doesn’t handle at all. 

I’d be fine with a PR that adds a flag to choose standard which you want, but adding Inf and NaN support to the JSON output would be a bit more work. They would have to be explicitly checked for and generated as strings. There are a couple of ways to do that, either make yajl_gen do it (in libCom/src/yajl) or have pvDataCPP add the checks. The latter would then work with 3.14.12.x and 3.15 versions of Base.

- Andrew

-- 
Complicity is easy, Simplexity takes real work

> On Feb 7, 2024, at 12:54 PM, Michael Davidsaver via Core-talk <core-talk at aps.anl.gov> wrote:
> 
> On 2/7/24 02:28, NICOLE Remi via Core-talk wrote:
>> Hello everyone,
>> I planned on upgrading EPICS from 7.0.7 to 7.0.8 in our build system,
>> but had some errors in our PV access tests.
>> Apparently, in 7.0.8, the 'pvget my:pv -M json' now returns JSON5:
>>     my:pv:name
>> {value:42.0,alarm:{severity:0,status:0,message:""},display:{description
>> :"My PV description"}}
>> Where before in 7.0.7, it returnd standard JSON:
>>     my:pv:name
>> {"value":42.0,"alarm":{"severity":0,"status":0,"message":""},"display":
>> {"description":"My PV description"}}
>> Was this a conscious choice? I don't see the change in the release
>> notes. I think for interoperability, standard JSON would be better.
> 
> This change was the result of:
> 
> https://github.com/epics-base/pvDataCPP/pull/93
> 
> Further changing to add "-M json5" seems reasonable
> to me.  I will let Andrew say more.
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System