Hi Fabio,

do you have any rational reason why you post these bug reports to us, and not to MSFT? And why in he** don't you wait two weeks until VFP9 is released and check if the bugs still remains before you post these messages? It's simply boring to read your complains which may have been fixed without your knowing it, and it makes you look not so serious which I know you are!

>Previous issue Thread #966108
>
>Bug/Issue : 23 of 36
>
>TITLE: PRB: VFP can put bad keys into a index
>
>VERSION: 09.00.0000.1720 and previous
>
>EXPECTED: VFP check a memo/blob index key correctly
>
>OBSERVED: if you define index with a memo/blob expression
>before that a value of key exists too much along,
>VFP ignores the problem and renders the index inconsistent. If subsequently the index comes recreated, the creation ruotine controls the length and generates an error.
>
>REPRO:
>

>ON ERROR MESSAGEBOX(MESSAGE())
>CLEAR
>CREATE CURSOR MEMOINDEX  (F1 M)
>INSERT INTO MEMOINDEX  VALUES (REPLICATE('1',200))
>? VARTYPE(F1)
>? VARTYPE(F1+'0')	&& M+C => C THIS IS INCORRECT FOR ME, and it generates problems
>
>INDEX ON F1 	TAG Ctag	&& VFP check memo and fire a error
>
>INDEX ON '0'+F1 TAG Ctag1	&& VFP IGNORE THE FUTURE PROBLEM
>
>* NEXT INSERT A KEY INTO CTAG1, BUT WHAT IS IT THE KEY VALUE ?
>
>INSERT INTO MEMOINDEX  VALUES (REPLICATE('1',270))
>
>REINDEX && THIS FIRE A ERROR NOW
>
>IF VERSION(5)>800
>	CREATE CURSOR BLOBINDEX (F1 W)
>	INSERT INTO BLOBINDEX  VALUES (REPLICATE(0h01,200))
>	>	? VARTYPE(F1)
>	? VARTYPE(F1+0h00)	&& W+Q => Q THIS IS INCORRECT FOR ME , and it generates problems
>
>	INDEX ON F1 	TAG btag
>
>	INDEX ON 0h00+F1 TAG btag1
>
>	INSERT INTO BLOBINDEX  VALUES (REPLICATE(0h01,270))
>	>	REINDEX && THIS FIRE A ERROR NOW
>ENDIF
>ON ERROR
>