Re: Coming soon to your mailbox (part 2)

Hi Imre,

> Please don't close this mail, i have realy something you might be interested
> in.

Yes very much ;-)

> In the last couple of weeks i have been busy writing my own version of
> chkdsk, it is now completely written but i still have to give a much better
> testing then what i have already done.

It is in my plans to (strting 2003 1st q) to write "disk dammaging sw" to test
all this

> I does some more checks on the boot of floppies, but doesn't correct them. A
> more general check should be implemented for hard disks and other media (it
> should work on anything that has a FAT12 or FAT16 file system, FAT32
> checking will not directly be guaranteed to work because of lack of memory
> (speed up data structures that take a few Kbytes on FAT16 take a few Mbytes
> on FAT32, because a much higher cluster count).

What language are you using? My suggestion is that you just compile the fat32
enabled version in protected mode. Don't bother to use XMS, just go straight
to 32 bit mode, it is much much easyer, faster and a xms version will never
be any better.

You can easely assume that a machine has enough RAM to fit the fat32 data. 
Usualy small 386s don't even use fat32 and if you gave a multy gigabyte 
drive it usualy is on a bigger cpu. ( this subject has already been discussed
at length, I just don't remember if privetly with Andreas or in the list)

Maybe even that the bigger HDs in fat16 won't fit in memory. The only
expected behaviour is a polite "not enough memory" message.
It happens sometimes with MS's scandisk.

> Moreover it lets you show  files and the accompanying defragmentation factor
> on the volume. As said it supports image files so my chkdsk is actualy the
> only utility that i know of that actually lets you "dir" inside an image
> file.

??

>  The reason why i wrote it and why it was done so quickly is because it
> contains for 90% the same code as defrag, which is to say it has been
> written with the library i wrote for defrag (the FAT Transformation Engine).
> So when testing the much easier algorithms in chkdsk (defrag has very
> sophisticated algorithms) i am in fact also the testing the code for defrag.

Can it be used to make just a simple surface scan function?
This is badly needed for FORMAT. Scanning is easy and the real problem
is probably what you call the FAT Transformation Engine ? :)
The surface scan is the (almost) only missing part for us to have a 
fully working FORMAT program.

> The reason why i say this now is because according to Alain, Andreas has
> already started on his version of chkdsk, while he said that he wasn't going
> to work on it before at least 2003 (if nobody else, like me, had already
> written it).

More or less. He has written most of it more then a year ago, and very 
little was tested (we discussed some of it then). You can count on me 
and most probably on Andreas to join you and work as a team ;-)

For ME (in the most selfish way) a scandisk/chkdsk/dosfsck is very
important, so I am most willing to do as many testing as I can. 
What is more: my "disk dammaging sw" may be usefull for any version
and maybe your library can help me in doing it.

Alain