>>>>>>>>>lo=eval("m.toThis." + m.lcObjName)
>>>>>>>>
>>>>>>>>That's I believe what I tried. But may be I did put mismatched quotes... I'll re-try.
>>>>>>>>
>>>>>>>>UPDATE. Yes, it worked. I had a very bad headache the whole day yesterday at work, may be I just could not concentrate... My colleague told me he also had it and my son did, so I think it's weather related...
>>>>>>>
>>>>>>>use
>>>>>>>

>>>>>>>lo=GETPEM(m.toThis,m.lcObjName)
>>>>>>>

>>>>>>
>>>>>>I thought about it too. But anyway, evaluate works and it seems good enough for me.
>>>>>
>>>>>It work ( slower ),
>>>>>but it is formally wrong when the object reference is static
>>>>>as in :
>>>>>"m.toThis."
>>>>
>>>>Could you please elaborate more? Also GetPem doesn't have object reference as a return type, that's why I thought it would not work.
>>>
>>>??? why ???
>>>If GETPEM didn't return an object reference, what it could return?
>>>
>>>>
>>>>Thanks.
>>>
>>>I mean you this
>>>

>>>* this it is VFP static code and then it is faster
>>>r1=GETPEM(m.toThis,m.lcObjName)
>>>* this return the same value,
>>>* but when someone uses unnecessary dynamic code, it do a mistake
>>>r2=eval("m.toThis." + m.lcObjName)
>>>? "this print true",r1=r2
>>>

>>
>>So everything done before VFP 6 was a mistake? GETPEM was added in VFP6.
>
>I think he meant a mistake in terms of speed

No. It is mistake.
Same concept is applied to m. prefix, read a variable without m. is mistake.

Speed up is a side effect.