Mailing List Archive

Please let me add that the symbol translation code inside imklog is
inherited from sysklogd. The BSD driver does not at all try to translate
the symbols (because the BSD syslogd did not do that).

I personally consider it even partly evil to do the symbol translation -
I supported it, because the original sysklogd did. From the designer's
point of view, translation code is not really a syslog functionality -
it belongs into the kernel or loader space. If a recent enough kernel
now supports the translation natively, I would even recommend to DROP
the code altogether (maybe after some period of being off by default).
That would also result in a good cleanup of MOST of the imklog code (and
code, as I said, that actually belongs to a different computing domain).

Feedback deeply appreciated.

Rainer

> -----Original Message-----
> From: rsyslog-bounces at lists.adiscon.com [mailto:rsyslog-
> bounces at lists.adiscon.com] On Behalf Of theinric at redhat.com
> Sent: Monday, June 09, 2008 4:52 PM
> To: rsyslog-users
> Subject: [rsyslog] imklog symbol translation
>
> Hi all,
>
> I want to propose that the symbol translation in imklog is disabled by
> default. The 2.6.x kernels do the translation automatically, so it
> seems unnecessary to enable it in imklog. Does anyone know a reason
> why it shouldn't be changed?
>
> Regards,
> Tomas
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog

> now supports the translation natively, I would even recommend to DROP
> the code altogether (maybe after some period of being off by default).
> That would also result in a good cleanup of MOST of the imklog code (and
> code, as I said, that actually belongs to a different computing domain).

I guess my question would be what valid reasons there could be for
retaining the functionality, turned off or not. Appropriately placed
or not, could anyone provide justification for leaving that bit in?
If the end result is getting rid of it, turning it off by default may
be a start but should come with a hard deadline of complete removal -
we all know how unused code rots.

I will now change the default starting in the beta (3.16.2). I'll also
schedule this feature for removal - maybe with 4.x.

Rainer

> -----Original Message-----
> From: rsyslog-bounces at lists.adiscon.com [mailto:rsyslog-
> bounces at lists.adiscon.com] On Behalf Of RB
> Sent: Tuesday, June 10, 2008 12:29 AM
> To: rsyslog-users
> Subject: Re: [rsyslog] imklog symbol translation
>
> > now supports the translation natively, I would even recommend to
> DROP
> > the code altogether (maybe after some period of being off by
> default).
> > That would also result in a good cleanup of MOST of the imklog code
> (and
> > code, as I said, that actually belongs to a different computing
> domain).
>
> I guess my question would be what valid reasons there could be for
> retaining the functionality, turned off or not. Appropriately placed
> or not, could anyone provide justification for leaving that bit in?
> If the end result is getting rid of it, turning it off by default may
> be a start but should come with a hard deadline of complete removal -
> we all know how unused code rots.
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog

> I will now change the default starting in the beta (3.16.2). I'll also
> schedule this feature for removal - maybe with 4.x.

Ummmm... of course, I meant 3.17.4 (3.16 is the current stable).

> Rainer
>
> > -----Original Message-----
> > From: rsyslog-bounces at lists.adiscon.com [mailto:rsyslog-
> > bounces at lists.adiscon.com] On Behalf Of RB
> > Sent: Tuesday, June 10, 2008 12:29 AM
> > To: rsyslog-users
> > Subject: Re: [rsyslog] imklog symbol translation
> >
> > > now supports the translation natively, I would even recommend to
> > DROP
> > > the code altogether (maybe after some period of being off by
> > default).
> > > That would also result in a good cleanup of MOST of the imklog
> code
> > (and
> > > code, as I said, that actually belongs to a different computing
> > domain).
> >
> > I guess my question would be what valid reasons there could be for
> > retaining the functionality, turned off or not. Appropriately
placed
> > or not, could anyone provide justification for leaving that bit in?
> > If the end result is getting rid of it, turning it off by default
may
> > be a start but should come with a hard deadline of complete removal
-
> > we all know how unused code rots.
> > _______________________________________________
> > rsyslog mailing list
> > http://lists.adiscon.net/mailman/listinfo/rsyslog
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog