Re: YAJL 1002 parse error: trailing garbage

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

Javier,

Data-into parsers report errors back to the calling RPG program using 
the IBM-provided QrnDiReportError() subprocedure.   This will cause the 
RPG program that runs DATA-INTO to receive an error that must be caught 
with techniques like (E) or MONITOR.   The QrnDiReportError() routine 
does not provide a way to specify a message, it can only specify an 
error number, so you should be getting error number 1002.

For your convenience, YAJLINTO *also* puts a human-readable *DIAG 
message into the job log, as well as the DATA-INTO trace log. Again, 
this is /in addition to/ reporting it as an error via 
QrnDiReportError().   The QrnDiReportError() is the primary means of 
sending back the error message.

It sounds to me like you (or whomever wrote the code to call DATA-INTO) 
is ignoring the error sent by RPG, and then complaining that the extra 
diagnostics are not reported as an error?    If you want an error, don't 
ignore the error code sent by RPG!


On 6/23/2021 12:21 PM, Javier Sanchez wrote:

> Jon, Charles:
>
> Well the thing is that I expected YAJLINTO to let me know who called it.  I
> will have to put up my own logs.  I'm actually using another dev's work so
> he's not using the (E) extender.
>
> The issue here is that if that error is true then at the point where it's
> happening, the rest of the call stack should be broken too but that is not
> the case.  So that's why I thought maybe it could parse successfully the
> json but found more invalid data just after the last valid construct, and
> is reporting it with no further harm.  Then it becomes a warning.  Maybe
> Scott could give me his comments on the above.
>
> Thanks folks!