> -----Original Message----- > From: [email protected] [mailto:rsyslog- > [email protected]] On Behalf Of [email protected] > Sent: Thursday, January 28, 2010 6:32 AM > To: rsyslog-users > Subject: Re: [rsyslog] Unicode & rsyslog - was: RE: PostgreSQL: > Problems with character encoding > > On Wed, 27 Jan 2010, Jonathan Bond-Caron wrote: > > I'm not sure if I understood but are you suggesting that all input to > > rsyslog is converted to UCS internally? > > That seems like a huge performance penalty to pay when most people > (?) log > > US-ascii or UTF-8 data. > > right now rsyslog doesn't do any unicode stuff, it treats everything as > a > string of bytes (with some code to escape specific characters). He is > saying that the path he has been planning to take would convert > everything > to UCS internally. you saw my argument against that.
I didn't yet respond to the original message because David's argument is a good one and I did not yet have time to think it over. Please note that there are many subtle issues, especially when combining it with the demands of the relevant RFCs (and if I implement it, I will definitely take a path that is standards-compliant). David's argument and proposed solutions sounds good to me, though I have some long-term concerns (eg. Can we really expect that Japanese/Chinese systems always use US-ASCII for the core logging information - I do not truly believe in that...). However, I simply have no time to implement Unicode right now, so what I most probably will do is copy over this valuable discussion and arguments into the design doc, so that I have them ready at hand when I can turn into that direction. But in general, I now tend to agree to David's argument and think that it can probably even speed up the process of a full Unicode implementation. Rainer _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com
