Mailing List Archive

> The atomic operations used in debug.c and msg.c were not introduced
> until gcc-4.1.0;

Forgot to give specifics: the atomic operations as defined in
atomic.h. The change made to bring debug.c to 1.49 in particular.

Strike that. Even better if I just read the ML before posting... *sigh*

Long day, guys, sorry for the noise.

On Tue, Apr 8, 2008 at 5:53 PM, RB <aoz.syn at gmail.com> wrote:
> > The atomic operations used in debug.c and msg.c were not introduced
> > until gcc-4.1.0;
>
> Forgot to give specifics: the atomic operations as defined in
> atomic.h. The change made to bring debug.c to 1.49 in particular.
>

No problem at all, I think it's well-hidden in the ML. A good fix (via
autoconf is still not done, so I should probably add a bug tracker ;).
Please keep reporting.

All others please note that I will release updates for v3-stable and
beta soon. Its brewing ;)

Rainer

> -----Original Message-----
> From: rsyslog-bounces at lists.adiscon.com [mailto:rsyslog-
> bounces at lists.adiscon.com] On Behalf Of RB
> Sent: Wednesday, April 09, 2008 1:56 AM
> To: rsyslog at lists.adiscon.com
> Subject: Re: [rsyslog] Atomic operations in 3.15.0
>
> Strike that. Even better if I just read the ML before posting...
> *sigh*
>
> Long day, guys, sorry for the noise.
>
> On Tue, Apr 8, 2008 at 5:53 PM, RB <aoz.syn at gmail.com> wrote:
> > > The atomic operations used in debug.c and msg.c were not
> introduced
> > > until gcc-4.1.0;
> >
> > Forgot to give specifics: the atomic operations as defined in
> > atomic.h. The change made to bring debug.c to 1.49 in particular.
> >
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog

On 4/9/08, Rainer Gerhards <rgerhards at hq.adiscon.com> wrote:
> No problem at all, I think it's well-hidden in the ML. A good fix (via
> autoconf is still not done, so I should probably add a bug tracker ;).

I don't know about the remainder of the atomic changes you've made,
but a check for >=gcc-4.1.0 will suffice for the stuff that's in
atomic.h.

I must say - I can't express how excited I was to find this project.
It's been on my long-term plate for quite some time to fork syslog-ng
and add all the nice features the Pro license grants; now I don't have
to! I'm partial to the syslog-ng configuration syntax, but can most
definitely live with this for not having to re-implement all that
stuff. Tell me, are you also working on UDF-WORM? ;)

On Wed, 2008-04-09 at 09:08 -0600, RB wrote:
> On 4/9/08, Rainer Gerhards <rgerhards at hq.adiscon.com> wrote:
> > No problem at all, I think it's well-hidden in the ML. A good fix (via
> > autoconf is still not done, so I should probably add a bug tracker ;).
>
> I don't know about the remainder of the atomic changes you've made,
> but a check for >=gcc-4.1.0 will suffice for the stuff that's in
> atomic.h.

I just began, more outstanding... But it looks like that easy solution
is the right one - thanks!

> I must say - I can't express how excited I was to find this project.
> It's been on my long-term plate for quite some time to fork syslog-ng
> and add all the nice features the Pro license grants; now I don't have
> to!

Good to hear ;) And, believe it or not, I never ever installed
syslog-ng. But,of course, it's the primary competitor in this space.
IMHO we are now far ahead of it (I guess you already know that chart,
but...):

http://www.rsyslog.com/doc-rsyslog_ng_comparison.html

Bazsi seem currently to brew something in regard to the log store, which
is not (yet ;)) my priority. Also I don't think it helps to
cryptographically sign the store (at least for court), the message
itself must be signed (that's my route on that). So we will probably see
a few features in the *paid* edition of syslog-ng that rsyslog does not
yet have. Even with that, and even compared to the paid edition, I think
we are now far ahead.

> I'm partial to the syslog-ng configuration syntax, but can most
> definitely live with this for not having to re-implement all that
> stuff.

rsyslog's config file syntax is just plain ugly. No way around. But wait
a bit, I am working on a scripting engine:

http://rgerhards.blogspot.com/2008/02/introducing-rainerscript-and-some.html

http://www.rsyslog.com/doc-rscript_abnf.html

> Tell me, are you also working on UDF-WORM? ;)

Neither of them - just the boring stuff ;)

Rainer