[Carpet] CCTK_TraverseString: invalid group/variable name

[Luca Baiotti]

Thomas,

>> I am aware of the problem of different users having different wishes 
>> and I don't think that the default solution can be good for everyone.
>>
>> Indeed, I would suggest to add parameters for CarpetIO* to specify the 
>> warning level. There are other level 1 warnings which I don't care 
>> about, like
>>
>> WARNING level 1 in thorn IOUtil processor 28 host node0017.admin
>>   (line 599 of 
>> /data1/baiotti/Cactus/configs/belladonna_NoDebug/build/IOUtil/CheckpointRecovery.c): 
>>
>>   -> Recovery directory 
>> '/data20/baiotti/Meudon-45km-5lev-dx.25_noRefBoundInStar_IF/checkpoint' 
>> doesn't exist
>> (because I was setting recovery to autoprobe).
>>
>> Did I convince you?
> 
> Not for this example.
> Autoprobe mode means "Probe for checkpoint files and automatically 
> recover, continue as usual if nothing was found" (this is the 
> description, as taken from IOUtil's param.ccl, of the "autoprobe" range 
> for the parameter IO::recover. If no recovery directory exists, or no 
> suitable checkpoint is found therein then Cactus will issue a warning 
> but continue.

Indeed I wrote what above as a counterexample to setting level 1 
warnings to be lethal ;-) I do want to see the above warning, but - of 
course - I don't want it to be lethal. Did I explain clearly this time?

> Regarding the I/O parameter check warning you mentioned in your first 
> email of this thread:
[...]
> I fully agree that one should be able to configure Cactus to abort a run 
>  on such errors. And this is already there :-) If not specified 
> otherwise in the parfile, you should have seen a follow-up level-0 
> warning like this:
> 
>> WARNING level 0 in thorn IOUtil processor 0 host localhost.localdomain
>>   (line 132 of 
>> /home/tradke/cactus/configs/experimental/build/IOUtil/Utils.c):   -> 
>> error while parsing parameter 'IOASCII::out1D_vars'
> The following I/O parameter controls the runtime behaviour during 
> parameter checking:
> 
>> BOOLEAN strict_io_parameter_check "Stop on errors while parsing I/O 
>> parameters from parameter file ?" STEERABLE = RECOVER
>> {
>> } "yes"
> 
> As you can see, it defaults to yes, which is want you want, I guess. If 
> other people don't like this behaviour they'd explicitely set the 
> parameter to no.

I do not see the level 0 warning and I do not set the parameter you 
mention. This means that probably there is a bug in the output routines. 
Check yourself my data on belladonna in 
/data21/baiotti/prova_recovery-parameters. The parfile is 
check_recovery-parameters.par and the stderr/out is provaREecovParam.o12867

I would really like IOUtil etc. to behave like you said and stop the run.


Let me know,

Luca