Thanks Harris. You're right. The problem is because of exhausted stack. I modified the count_blocks function to use BFS algorithm, instead of the current DFS algorithm. With BFS logic, count_block is returned successfully. However, it still crashes in other places (where DFS logic is used) due to the same reason. Hopefully, future WinPcap/libpcap uses BFS in those functions to handle complex filter strings without depending much on the system stack.
--- Guy Harris <[EMAIL PROTECTED]> wrote: > On Mon, Mar 24, 2003 at 03:32:26PM -0800, Guy Harris > wrote: > > Infinite, or just "more than there's room for in > the stack"? A stack > > overflow does not necessarily imply infinite > recursion. > > I tried that same filter (after removing extra > blanks and "!"s that > somehow had gotten inserted into the file) with the > current CVS libpcap > (which isn't much different from WinPcap 3.0 beta - > especially not in > that part of the BPF code generator/optimizer) on > Solaris, and it didn't > crash. It might be that the stack limit was larger > on Solaris than on > the version of Windows on which it crashed. > > > ================================================================== > This is the WinPcap users list. It is archived at > > http://www.mail-archive.com/[EMAIL PROTECTED]/ > > To unsubscribe use > mailto: > [EMAIL PROTECTED] > ================================================================== __________________________________________________ Do you Yahoo!? Yahoo! Platinum - Watch CBS' NCAA March Madness, live on your desktop! http://platinum.yahoo.com ================================================================== This is the WinPcap users list. It is archived at http://www.mail-archive.com/[EMAIL PROTECTED]/ To unsubscribe use mailto: [EMAIL PROTECTED] ==================================================================
