Tomasz Kojm wrote:
The problem here is not English. You're just ignoring my answers and
this cannot lead to a constructive discussion.
Sory if I did something wrong, I did not want it. But I am still
trying to understand you .
I really did not want to tell something bad for you, nor for your work,
please be cool.
The English are not problem but all people in programmers community are
comming from
different countries and usual we are writing as we are thinking in our
native language.
Maybe this has as result to unwanted misundertands. Maybe I am wrong
here but
maybe the same is for you.
I don't think there will be performance issues with thousands of
files you mentioned, unless you're running some really obsolete
hardware.
From a point of view you are right. But I believe that if you have
100-200 objects
or more with 5K size of each will be significant faster to scan them,
simultaneous,
in memory than in files.
I don't see a point why would you want to replace all
calls to open/read/seek in libclamav with your own functions.
In this case someone can use streams instead of files where streams can be
files or memory buffers or anything else, which can you open it, read
bytes from it
and seek to the desired position.
Iam thinking it only as possibility. I have a huge todo list to add more
thinks in it.
There's no simple answer to this question. Scan results of partial data
are not reliable.
OK. I had seen some implementations which use this implementation and
I wanted to hear a reliable answer.
Many thanks.
Really friendly,
Christos
P.S.
To Tomasz Kojm: My name in greeks means something like "angry". You
understand
which are the jokes from my friends about it. I hope that no-one of
them will read these mails ....
_______________________________________________
http://lists.clamav.net/cgi-bin/mailman/listinfo/clamav-devel
