Author: Mario
Email: [EMAIL PROTECTED]
Message:
Hrmmmm. some debugging revealed a little information.
the tests are done on 2 seperate platforms
platform 1 is debian potatoe distribution, PIII, ext2
platform 2 is redhat 7.0 distro, PIII, reiserfs.
The Potatoe distro seems to be happy with the splitter, as I havent gotten a single
header skip yet with the same RAW data.
the Redhat Distro seems to be confused on where to put some header information (about
4 bytes worth after the header)
Here is the two hexdumps from a word(gananoque 0xA58f8000)
debian line 1 of hexdump:
0000000 0008 0000 0000 0000 87db a58f 0001 0000
redhat line 1 of hexdump: (DIFFERENT CACHE FILE)
0000000 0006 0000 9ff4 bffe 87db a58f 0001 0000
the only difference are:
-there are 2 more url in the debian (because I indexed more pages after to
preproduce error)
-there is this "9ff4 bffe"
"9ff4 bffe" seems to get the debian confused with the same messages I can produce from
the Redhat box, after I copied the file over from the redhat distro.
This narrows it down to the actual cache write function at least(?)
Reply: <http://www.mnogosearch.org/board/message.php?id=2504>
___________________________________________
If you want to unsubscribe send "unsubscribe general"
to [EMAIL PROTECTED]
