>>>>Hi all,
>>>>
>>>>In our Application (VFP 6 SP5) we receive sometimes an C0000005 with the function SET("CLASSLIB"):
>>>>
>>>>lcClasslib = SET("CLASSLIB")
>>>>or
>>>>IF "MYCLASSLIB" $ SET("CLASSLIB")
>>>>
>>>>It happens on different OS and different machines, but it's not reproducable. Any ideas?
>>>>
>>>>Rolf
>>>
>>>Now i can reproduce the problem on my machine. In my tests i write the result of the function in a logfile. During the program execution there is a noticeable change in the result of the function allthough there are no libs released or new loaded.
>>>
>>>At the beginning the result is ok and looks e.g like this:
>>>
>>>C:\PATH\CLSLIB1.VCX ALIAS CLSLIB1, C:\PATH\CLSLIB2 ALIAS CLSLIB2, C:\PATH\CLSLIB3 ALIAS CLSLIB3...
>>>
>>>After some time the result changes like this:
>>>
>>>C:\PATH\CLSLIB1.VCX ALIAS GODATA_345280178 , C:\PATH\CLSLIB2 ALIAS TABNAME , C:\PATH\CLSLIB3 ALIAS FIELDNAME...
>>>
>>>or
>>>
>>>C:\PATH\CLSLIB1.VCX ALIAS , C:\PATH\CLSLIB2 ALIAS , C:\PATH\CLSLIB3 ALIAS ...
>>>
>>>As you can see, the alias of the loaded librarys changes allthough the libs are not released and new loaded. The correct alias are e.g. replaced by names of open tables or by fieldnames or something other. Sometimes the alias is empty but the keyword "ALIAS" is still there. Also there is sometimes a space between the alias and the following comma.
>>>
>>>Then after some time the crash happens.
>>>
>>>It seems to me, that the VFP-memory-management becomes corrupt. Now i will try to find the reason.
>>>
>>>Rolf
>>
>>
>>Rolf,
>>
>>If you're looking for causes for corrupt memory - I have a candidate for you - are you using a lot of chrtran() ?
>>
>>Though this is supposed to be solved in vfp6sp5, I believe it was still in
>>http://support.microsoft.com/kb/221617/EN-US/
>>
>>
>>I had this situation (message#1251395) - must confess that I used chrtran in sql selects
>>
>>Afterwards I replaced all the chrtran with sys(15) and the problem was gone
>
>
>Hi Gregory,
>
>thanks for the advice. There are some chrtran() in the program, but as you said this should be solved in vfp6sp5. Our problem occurs with vfp6sp5. But i will check it.
>
>Rolf


I did have the problem in vfpsp5 ...