Different offsets for rlock in comix vs. six3 
Author Message
 Different offsets for rlock in comix vs. six3

Hi

One testing application that i'm writing uses 2 rdd's: comix and six3, and
is written for clipper 5.2e, blinker 5.1, comix 3.09, six 3.02.-

When trying to set a record lock over the first record in a .dbf using comix
as rdd, i can see that the lock is placed at 03B9ACA01h, that is [1000000000
+ recno()] bytes ahead from start of file. It is the same offset that uses
dbfntx.
If i use six3 instead, it uses the offset 07FFFFFFDh, or something like
[2^31 - recno() - 3] bytes from start.
The offsets were seen under novell's monitor.

The thing is that i can't use both rdd in different applications over the
same set of dbfs at the same time, because the locking scheme is different
for both drivers, and that leads to corruption. In fact, from both
approachs, i think six3 is better.
Anyone using bro (broplus), can see this behaviour easily.

I will try linking in ntxlock2.obj, but i don't think this could be a
solution (still not tested, anyway).

Is there any solution or workaround?
--
Saluten
Claudio
Buenos Aires - Argentina
--
"Virus check complete. All viruses are working normally on this computer."



Mon, 15 Aug 2005 12:02:57 GMT  
 Different offsets for rlock in comix vs. six3

Quote:
> The thing is that i can't use both rdd in different applications over the
> same set of dbfs at the same time, because the locking scheme is different
> for both drivers, and that leads to corruption.

AFAIR, both offer an optional linkable object file, to make the lock scheme
compatible with the FoxPro lock scheme.

Ron



Mon, 15 Aug 2005 13:33:25 GMT  
 Different offsets for rlock in comix vs. six3


| > The thing is that i can't use both rdd in different applications over
the
| > same set of dbfs at the same time, because the locking scheme is
different
| > for both drivers, and that leads to corruption.
|
| AFAIR, both offer an optional linkable object file, to make the lock
scheme
| compatible with the FoxPro lock scheme.
|
| Ron
|

Hi Ron,
thank you for your answer!

I didn't look at the proper subdirectory, so I linked in cmxfox52.obj and
that's it.

I made a little experimentation, and here is my conclusion:

DBFNTX         1000000000+RECNO()
COMIX3         1000000000+RECNO()
Comix3+.obj    2^31 - recno() - 3
SIX3                2^31 - recno() - 3
DBFCDX         2^31 - recno() - 3

So it seems that comix without the cmxfox52.obj is compatible with dbfntx
but not with FOX, and comix + cmxfox52 seems to be compatible with six3,
dbfcdx and FOX, but not with dbfntx (for Clipper 5.2e), in relation to
record locking.
Six3 doesn't have any linkable .obj in its distribution disk to make it
backward compatible with dbfntx.
Its standard behaviour is to be compatible with fox.

HTH someone else...

--
Saluten
Claudio
Buenos Aires - Argentina
--
"Virus check complete. All viruses are working normally on this computer."



Mon, 15 Aug 2005 23:10:49 GMT  
 
 [ 3 post ] 

 Relevant Pages 

1. Comix - "v3.00.01 (Comix/5.2)"

2. clipper 5.2e + Comix 3.09 VS Delphi 4

3. RDD Comix vs. WIN98

4. RASQL/B Btrieve vs COMIX DBF/CDX

5. offset offset??

6. ptrs vs offsets

7. LEA vs OFFSET

8. LEA vs OFFSET

9. lea vs mov offset

10. Q: lea vs. offset

11. LEA vs. MOV OFFSET - x.x [1/1]

12. LEA vs. MOV OFFSET

 

 
Powered by phpBB® Forum Software