[Linux-ha-dev] Re: [Linux-ha-cvs] Linux-HA CVS: include by alan from 220.127.116.11
alanr at unix.sh
Wed Mar 3 08:59:45 MST 2004
Lars Marowsky-Bree wrote:
> On 2004-03-03T05:31:50,
> linux-ha-cvs at lists.community.tummy.com said:
> I'm not too happy with this one - tieing in the netstrings handling that
> deeply into the ha_msg handling makes it hard to access from port which
> doesn't use ha_msg (raw IPC etc).
> I think there should be a cl_netstrings.[ch] for the handling /
> conversion, and which is then used by ha_msg.[ch].
-rw-r--r-- 1 alanr 8496 2004-03-02 21:38 ./heartbeat/netstring.c
-rw-r--r-- 1 alanr 1397 2004-03-02 13:02 ./include/clplumbing/netstring.h
I want this moved into clplumbing. But, that will take some time, as it
would necessitate moving ha_malloc and ha_msg to clplumbing, which is more
complicated than it sounds.
I don't know of anyone who would bother with netstrings to send a single
field across the network. If that's the case, then you need some kind of
top level data structure.
It's helpful if we have some expectation of what the top level data object
being sent across is going to look like. For most people's purposes, that
would be a structure of some kind - which translates into a name/value pair
arrangement of some kind - aka an association. The only one implemented so
far is ha_msg.
Although I agree with you that having it separate is nice, I'm not sure if
I see a huge practical difference. Patches are being accepted.
I'd rather take the code in a working state, and get this change into the
next stable release which is coming in a few days, and work out any
internal rearrangements later. IMHO getting to the new on-the-wire format
is higher priority than internal organization. Having the new format is
essential before we freeze 1.2.x down to only bug fixes.
Alan Robertson <alanr at unix.sh>
"Openness is the foundation and preservative of friendship... Let me claim
from you at all times your undisguised opinions." - William Wilberforce
More information about the Linux-HA-Dev