JdBP> [ This is a pr cis of a message in the TAUCMD echo. ]

 SW> DESKARC LIST produces no output here.

LA>    Basic Structure:

LA>    21 bytes of unknown data, followed by a text label at offset 
  >    21d/15h, the label is: "Originally installed Archive", followed 
  >    by 53 bytes of nulls, a two byte numbering label at offset 
  >    102d/66h, followed by 246 bytes of nulls, followed by a 10 byte 
  >    string (":\Desktop" at offset 358d/166h, followed by 246 
  >    nulls, followed by a restart of the above sequence to a total of 
  >    four such entries.

LA>    The text string at 21d into each section *other* than the first, 
  >    is always: "Complete Archive", (There ain't much actual data 
  >    here...)

LA>    Note: All offsets are zero based.

LA>    Note: All of the "G:\Desktop" entries are followed by 246 nulls. 
  >    Add the 10 bytes of the string that preceeds those 246 nulls and 
  >    you get 256 bytes, or 16 paragraphs.

LA>    Note: The "numbering label" mentioned above is not sequential, 
  >    the four labels are: (in order of appearance) "0X", "02", "01", 
  >    and "03". On my maintenance partition, the order is: "0X", "01", 
  >    "03", and "02". My guess is this is more of a type label than a 
  >    sequence number, but what it indicates, I have no idea. Haven't 
  >    found any clues in my old DD kit, nor inside any executable on 
  >    disk, which seems reasonable considering they are compressed...

LA>    I'm guessing that the 21 byte section header contains a date and 
  >    time, and undoubtedly something else, but what?  I'll have a go 
  >    at decoding the date/time part tomorro.


   OK, following is a layout of the header portion (the bytes 
   immediately preceeding the "Original Archive"/"Complete Archive" 
   section)  of each record. This reflects an "old" existing 
   ARCHIVES.$$$ and a new (made today) ARCHIVES.$$$.

   As Will recollected, the entries rotate on a FIFO basis, the 
   oldest being overwritten and the number labels being rotated. But 
   notice how the 5th byte appears to be a record place number. The 
   date/time data appears therefore to begin at the 7th byte and 
   continue for 6 bytes from there.

   But I'll be danged if I can see how it's encoded. Anyone out 
   there familiar with OS/2's default D/T encoding scheme?  I'll 
   give a look through my header files and see what jumps out at me, 
   but based on past experience, I'm not optimistic.

   Anyone got any ideas?  BCD maybe?

   Here's the sequence:

                                           

   From ARCHIVE.NEW dated 11/23/99

NOTE: First line is decimal translation, second line is the original 
hex value. A "-" is a null.


********************************************
(Original Archive)

- - - - 88 - 21 56 43 37 2 2 207 7 255 255 2
0 0 0 0 58 0 15 38 2b 25 2 2  cf 7  ff  ff 2


(Entry label "3")

1 - - - 1 - 20 1 19 85 23 11 207 7 255 255 2
1 0 0 0 1 0 14 1 13 55 17 0b  cf 7  ff  ff 2


(Entry label "2")

1 - - - 2 - 10 37 18 51 15 9 207 7 255 255 3
1 0 0 0 2 0 0a 25 12 33 0f 9  cf 7  ff  ff 3


(Entry label "1")

1 - - - 3 - 12 9 27 63 11 9 207 7 255 255 6
1 0 0 0 3 0 0c 9 1b ef 0b 9  cf 7  ff  ff 6

****************************************

From ARCHIVE.OLD dated 9/15/99

(Original Entry)
.  [ Continued In Next Message... ]

___
 X SPEED 2.01 #2720 X Canadian DOS prompt: EH?\>


--- Maximus/2 3.01
270/101