----- Original Message ----- > Hello Jan, > > Is there any reason we shouldn't do this in trunk?
I don't see any reason. This patch was intended for trunk, but I don't have svn commit access, so I'm sending patches to this list :). It's also better that someone reviews my code, because I don't have so long experience with httpd development. Regards, Jan Kaluza > The patches and features seem generally correct to me with a cursory review. > > Thanks, > > Paul > > On Mon, May 27, 2013 at 3:23 AM, Jan Kaluža <jkal...@redhat.com> wrote: > > Hi, > > > > last week I was trying to write my own module to log error_log to > > systemd-journal [1] and I've found out that with the current error_log > > code, > > it's not possible to do that properly. > > > > I was able to use "error_log" hook, but there is no way to disable creation > > of "ErrorLog" file (One can set it to "/dev/null", but httpd will still > > write data to it without any reason). Syslog logger fixes that by > > hardcoding > > syslog methods in log.c/core.c, but I don't think that's the right thing to > > do with journald. > > > > Therefore, I've created following patches: > > http://people.apache.org/~jkaluza/patches/logging/ > > > > Their descriptions should be clear from their names, but I will describe > > them briefly here too. > > > > Patch 0001 declares ap_errorlog_provider which can be implemented by module > > providing error_log logger. Admin can later define "ErrorLog provider arg" > > to choose particular errorlog provider. Old syntax still works and the > > change is backward compatible. This patch also removes syslog logging from > > log.c (it is moved to newly created mod_syslog.c in next patch) > > > > Patch 0002 creates mod_syslog.c which uses the new API to implement syslog > > logging. It works the same way as the version in log.c I removed in > > previous > > patch, but it's in separate module. > > > > Patch 0003 shows how mod_journald.c can use the existing API. This module > > works well with systemd-journal, but unfortunately the performance of > > systemd-journal daemon is poor so far [2], but I presume it will be fixed > > and the module will be usable for general use in the future. There is > > probably no real benefit in accepting this last patch right now. It's here > > to only show why the previous two patches are useful. > > > > Note that this is my first bigger patch touching httpd core, so feel free > > to > > correct my possible mistakes... :) > > > > [1] http://0pointer.de/blog/projects/journalctl.html > > [2] https://bugzilla.redhat.com/show_bug.cgi?id=963620 > > [patches] http://people.apache.org/~jkaluza/patches/logging/ > > > > Regards, > > Jan Kaluza >
