Re: pmud and gkrellm on tcp???
On Tue, Oct 17, 2000 at 12:18:42PM +0100, Sergio Brandano wrote:
> ... I am reading the README of gkrellm's pmud plugin:
> > I also borrowed tcp code included with the snooze code.
> Why in hell one needs tcp code in the snooze code in the first place?
> Please, can the developers of the following programs get together
> and file a report on what is the need for using a tcp/ip service when
> managing the battery of a laptop:
> - pmud
> - gkrellm
> - pmud plugin
I'm not a author/maintainer, but how do you expect a daemon that is running
continually on the system to monitor battery/ac performance and user utilities
that run intermitantly and need to report these things to communicate?
The options other than a TCP socket are to:
- use a unix domain. This has problems with the safe creation of the unix
domain socket (since it would probably be created in /var/run or /tmp).
Also, some kernels don't support unix domain sockets by default. Conversely
it does give file permission like control to accessing the socket.
- use some sort of /proc interface, which would require pmud to actually be a
kernel module rather than a user space application. It would also require
the clients to poll the values.
Personally, I think that the current situation of having a service that is
listening on the loopback interface, and thus only accessable to the
localhost, is a good solution to the problem. The alternative of a unix
domain socket might be better in some ways, but I don't think its worth
On another note, you may wish to reconsider your choice of tone in future
emails. The authors of these programs have done so without pay, benefit or
training, and should be respected for that. Nobody will ever do a perfect
job, and if you think something is "not quite right", then you should research
the problem yourself and make suggestions or provide patches to the author.
If you can't or don't wish to do this, then you simply shouldn't use software
that you don't like.