[gccsdk] Syslog: DoggySoft vs. Unix

Graham Shaw gdshaw at sagitta.demon.co.uk
Sat May 26 13:24:00 PDT 2007


In article <5ce92fe94e.admin at snowstone.org.uk>,
   Adam <lists at snowstone.org.uk> wrote:
> In message <4ee914dfaegdshaw at sagitta.demon.co.uk>, Graham Shaw wrote:

> > but this gives numbers that are rather higher than would accord with
> > the Doggysoft documentation: notice really ought to be less than 125,
> > and warning no higher than 100.  Still it would be nice to have
> > something that was linear,

> Does it need to be linear?

To be clear, I'm not saying that it does - merely that a linear scheme
would be preferable on the grounds of simplicity if it can adequately
capture the semantics.

If something more complex is needed then fair enough.

> Here's a stab at a "non-linear" scheme:

> 0   <->             # Unix scheme doesn't account for armageddon!
> 10  <->             # ... or war
> 20  <-> Emergency
> 30  <-> Alert
> 40  <-> Critical    # Paraphrase: "critical should be less than 50"
> 50  <->
> 60  <-> Error
> 70  <->
> 80  <-> Warning
> 90  <->
> 100 <-> Notice      # This point seems to be fairly well defined
> 120 <->
> 130 <->
> 140 <->
> 150 <-> Information
> 160 <->
> 170 <->
> 180 <->
> 190 <->
> 200 <-> Debug
> 210 <->
> 220 <->
> 230 <->
> 240 <->
> 250 <->

That too would be reasonable.

The only point I would make is that informational messages are normally
logged on UNIX systems, which is why I was happy with their being placed
at 120.

(A counter-argument would be that UNIX systems tend in practice to filter
within the application rather than syslog, and for this reason, it is
normal to log debug messages too - but I very much doubt we would want
debug lower than 125 - and the same reasoning can arguably be applied to
informational messages too.)

Anyhow, regardless of the scheme, I think what matters most is to build a
consensus around it so that any other programs which need to perform the
conversion use the same method.

Before we take this further we should find out what the Select syslog
implementation does when logging remotely.

-- 
Graham Shaw (http://www.sagitta.demon.co.uk/graham/)
The RISC OS Packaging Project (http://www.riscpkg.org/)
The RISC OS Toolkit (http://rtk.riscos.org.uk/)




More information about the gcc mailing list