On 8/10/06, Horms <[EMAIL PROTECTED]> wrote:
On Thu, 10 Aug 2006 10:13:32 +0200, Andrew Beekhof wrote:
> [-- multipart/alternative, encoding 7bit, 1 lines --]
>
> [-- text/plain, encoding 7bit, charset: US-ASCII, 25 lines --]
>
> Can someone explain the purpose of
>
>
> case $SYSTYPE in
> *BSD)
> if $IFCONFIG $IFCONFIG_A_OPT | \
> grep "inet.*[: ]$OCF_RESKEY_ip " >/dev/null 2>&1; then
> continue;
> else
> return $OCF_SUCCESS
> fi;;
>
> in ip_stop()
>
>
> I must be missing something because I dont see the point. Isnt the
> prior call to find_interface() sufficient?
The handling of $SYSTYPE in IPaddr is, to be quite blunt, horrrible.
It is as if each person who added a line that calls $SYSTYPE
had a different idea about which lyer the abstraction should be made at.
For instance, find_interface() uses systype to decide weather to
call find_interface_solaris() or find_interface_generic().
But find_interface_generic als makes use of $SYSTYPE.
I've been scared to clean it up for fear of breaking stuff.
But I'll have a stab at it and post a patch. Well, perhaps a few
patches.
I'm in the middle of gutting the entire RA (hence the question)...
I'll post what I have shortly.
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
