On Fri, 13 Oct 2000, Alfonso De Gregorio wrote:
> Usally only kernels, daemons and system utilities log through
> syslog. A large part of them use english since it's an ideal
> vehicle for such kind of messages.
Yes, but the messages may contain literal strings.
> However, the are exceptions to this behaviour.
> Potentially, all programs with i18n (and/or l12n) support can log
> messages with a different code set. They are many.
Hmm, I'm not sure explicit support (i.e. conversion etc) is really
necessary. Programs log in English not because the programmers are lazy
but because unified log messages help a lot in bug reports. Only some
local software may be logging in a different language, and the codepage
should be set up correctly for that.
> If we choose to support different character set we should also
> consider to adopt Unicode encoding.
Definitely, for literal strings.
> Unicode adoption present also drawbacks (ie. diffuclties for syslog
> message interpreters to parse datas)
Not really, if done right. UTF-8 makes things difficult, because you may
encounter really weird characters if someone sends chars >128.
> > Let's also be sensitive to the fact that there are syslog message
> > interpreters out there. If we do find additional characters in the
> > wild, can we confirm that they are acceptable to the better known
> > packages?
What about adding special support here, in the form of a code field that
these tools could recognize. Tools tend to adapt fast.
Simon
--
GPG public key available from http://phobos.fs.tum.de/pgp/Simon.Richter.asc
Fingerprint: A319 A60F 20F6 C8A4 3C86 54B4 99CD AC6E 79D1 B1E7
Hi! I'm a .signature virus! Copy me into your ~/.signature to help me spread!
