Virgin import of tcpdump.org libpcap v0.6.2

1 files changed, 38 insertions, 0 deletions

diff --git a/contrib/libpcap/TODO b/contrib/libpcap/TODO
new file mode 100644
index 0000000..e90b9d8
--- /dev/null
+++ b/contrib/libpcap/TODO
@@ -0,0 +1,38 @@
+ TODO list for libpcap
+=======================
+
+Important stuff (to be done before the next release)
+---------------
+
+General
+
+- configure should not be in the CVS. Most open source projects have an 
+  autogen.sh script to run autoconf etc. after checkout. I think we 
+  should stick to the standard. 
+
+- The source files should be better documented. There is no official 
+  design guideline what is done where. There should be a common coding
+  style (okay, you can guess that bye looking at the code) and a guideline
+  what needs to be documented.
+
+Linux kernel interface
+
+- Currently there is a race condition in that a socket is activated at the
+  same time when it is opened - before applying a filter. This has to 
+  be corrected so that capture starts when pcap_read is called for the 
+  first time. 
+
+Less urgent items
+-----------------
+
+- Better documentation and cleanup of the interface. I am seeing a few  
+  problems at the first glance which needs fixing:
+  + pcap_lookupnet makes little to no sense with protocols != IPv4
+  + not very suited for interactive programs (think ethereal). There should 
+    be a way for the application to get a file descriptor which it has to
+    monitor and a callback in pcap which has to be called on activity
+  + too many functions. There are a lot of functions for everything which
+    violates the KISS principle. Why do we need pcap_strerror, pcap_perror
+    and pcap_geterr?
+  + the manpage has a brief description of each function but where is the
+    big picture? Seems like you need to buy UNP for that...