From: =?iso-8859-1?B?R+Fib3IgTOlu4XJ0?= on
On Wed, Apr 07, 2010 at 07:52:56PM +1000, Jim Smith wrote:
[...]
> How can I make it so that Subject: =?utf-8?B?UX..... is displayed correctly
> in it's native language? (in this case it's Chinese but it could be another
> language)

As far as I know such a transformation is the job of a MUA for example, it's
not the MTA's task to do things like this, so maybe postfix is unable to do
that, and it's right that way. It's matter of taste but I would be upset to
see non-ASCII chars in syslog files anyway, but it's my problem :)

I think some syslog servers supports to redirect the log into the stdin of a
program, so you can have "on-line" conversion with that way, though ... I am
using this to have own custom logging format without the need to convert the
log file after the log rotate (which requires double of the disk space and
you can have it only after you have the log file to process, so otherwise I
wouldn't have information in my own format but I must wait a day ...)

From: mouss on
Jim Smith a �crit :
> Can it be done natively when writing to the maillog or do I need to copy
> the file elsewhere and then run a conversion tool across all the Subject
> fields to get the correct characters?
>
> My maillogs are very large and I don't want to use resources
> converting them on the mail server, hence I'd copy them somewhere else
> before converting the Subject: string
>

- logs are ASCII. This has always been the case.
- postfix logs text as it comes. This is what logging is for (write
"real" infos)

if you want to ease reading logs, then convert logs to whichever form
you prefer. a
tail -F $logfile |perl showmelogs.pl
will do...