Mail To: Vassilis Papathanassiou <papval@otenet.gr>
Subject: Re: BNET with various MagiC versions
----------------------------------------------------------------------------

On Sat, 9 May 1998, Vassilis Papathanassiou wrote:
>Hi Ronald
>
>  Sorry for the delayed answer, after north Greece I went to Dubai and then
>for five days at the exotic Maldives (Indian ocean), so I've lost almost 10
>more days. It was nice, but returning home I found my 1GB Quantum Empire
>drive (for the falcon) dead. Dead means the motor doesn't spin to speed, so
>although the surface might be ok, I can't use it.
>  I consider that a disaster, since all my Falcon specific software is gone.
>No sources were lost, but I'm sure you know what it means, ie all
>configuration files, images, 200MB archives, demos, games art programs etc,
>etc. This was mostly a backup disk for zips etc and I had no indication that
>something is going to happen ie bad sectors etc.
>  If you have the slightest idea of ways to make it spin again, I'd be
>greatly appreciated. For the moment I've installed a SyQuest EZ 135 "disk"
>to be able to work. In this disk I've installed MagiC 5.11 (this is the
>version System Solutions sent me, with a ver 4.xx manual! )
>
>So far I have no problems with BNET, and this time it works heavily (!) to
>transfer stuff from the TT. Needless to say we have the same version, I had
>no time to add even a line after all this mess.
>
>>I don't know when this will reach you, but I thought I had better write it
>>down while it is still fresh in my memory.  I have had opportunity to test
>>BNET under some more MagiC versions, and it now does appear as if the bugs
>>I had on the Mega ST were specific to that MagiC version, or more probably
>>to some group of versions above your 4.xx and below the 5.13 for which it
>>worked all along.  I have now also tested under MagiC 5.20, which worked
>>fine despite the fact that the MagiC bug I discovered in version 5.13 is
>>still present...
>>
>Hmm, I must be careful with this bug, although MagiC sends the paths to the
>XFS without any leading '\' and without drive of course, since it knows the
>drive the XFS belongs to.
>
>>My friend with the two TTs has also tested BNET both before and after he
>>upgraded to MagiC 5.20.  All of that worked fine too.   :-)
>>
>Great news, but still I think I'm missing something concerning the memory
>management. Of course I do NOT use any dirty tricks, it's just that BNET
>drive needs local memory to keep folders/files that the kernel 'thinks' are
>accessible at any time.
>Anyway, now with version 5.xx (kernel version 4) I'll do some more tests
>and we'll see...
>
>>Except for the unfortunate case of some buggy MagiC versions it now seems
>>as if the actual bugs of BNET are beaten.  There are still areas where it
>>needs to be improved, such as speed, recursive directory operations, etc,
>>but the basic functionality can be regarded as complete.
>>
>Sure, it's very difficult to debug it completely but if it works ok to the
>point where we can make it public, the remaining bugs will be ironed out.
>I'll fix the DTA problem (I hope this is what you mean by 'recursive directory
>operations' and not Dreaddir etc since with the later there is no problem on
>my setup).
>
>  Another thing I want to do is to 'secure' file operations when more than 2
>machines are connected. This needs a different file handle structure in
>BNET.APP. Since you consider the 'basic functionality' complete, then I'll
>send you the sources for debuging/converting to BetaDos.
>
>Regards     Vassilis
>
>
>