Hi,

Vadim's mail raises an important question on compression:

***** Databases are incompatible if the compression code changes
even very slightly ****

-> The bugfix  I did  yesterday changed the  coding of  certain values
from 4 to 5 bits, making previously comrpessed db unreadable.

Wev'e  been thinking  of ways  to  convert db's  from one  compression
version to another, it's very important but we can live without it for
some time.

Iv'e included  a lot  of code  to look for  compression errors  and to
abort when  it encounters something strange... this  should avoid most
cases of  core dumps like Vadim's.  However due to the  very nature of
compression   (the  compressed  data   "looks"  randomized)   this  is
difficult.

The solution I'm  going to add is a compression  version on the header
of  the compressed data  to abort  if the  user is  mixing compression
versions...that will be better than a core dump :-)

Cheers,


Marcel.



    ----------------------

    Marcel Bosc

    ECILA
    100 av. du Gal Leclerc
    93500 Pantin  - France
    Tel:  33 1 56 96 10 85

------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED] 
You will receive a message to confirm this. 

Reply via email to